The 1996 Ling-TeX Archives

5th January 1996: IPA fonts: Good in TeX?

From: kirk@jean.ling.umu.se (Kirk Sullivan)
Subject: IPA fonts: good in TeX?

Some colleagues and I are are writing a paper which 
considers the problems of phonetic typesetting.
We wish to ask you about your experiences.
A short set of questions follow which we would like you
to answer and e-mail to kirk@ling.umu.se 

Many thanks for your time!

1) Which computer do you use to write documents on?
 
2) Which word-processing package or equivalent do you use?
 
3) Which IPA-fonts do you use?
 
4) Do the fonts you use cover the entire 1993 IPA-chart?
 
4a) If they do cover the entire chart are there any limitations
    on the use of diacritics etc?
 
4b) If they do not cover the entire chart what is missing?
 
5) Are the fonts you use available in a variety of different
   styles?
 
6) Do you find that the fonts are easily included in documents?
 
6a) If not, what is the problem?
6b) Do you think that the problem is due to the fonts or the
    word-processor you are using?
 
7) Do you have an example of a particularly frustrating/negative
   experience? For example, a particularly complex transcription which you
   could not get to typeset in a pleasing manner.
 
8) If you have used or use other fonts can you briefly compare them.
 
9) Are you generally satisfied?
 
10) What would you like to see improved?
 
11) Would you recommend the system you are using to someone else?
 
12) Any other comments you would like to make.

20th January 1996: On the next version of TSIPA

Subject: On the next version of TSIPA
From: FUKUI Rei 

Hi. I am preparing a new version of TSIPA (version 1.2). This new
version will include the following enhancements over the previous
version.

 (1) complete LaTeX2e support
 (2) compatible with both OT1(cm) and T1(dc) encodings
 (3) slanted and bold extended font series/shapes
 (4) font sizes other than 10pt
 (5) increase the number of characters in a font (i.e. 128 -> 256)
 (6) change the selection of symbols and encoding completely
 (7) create slightly modified version which is compatible with Times
     Roman font in terms of font shapes and font dimensions when used
     with psnfss package

Changes from (1) through (3) have been almost done. Now I am working
on deciding a new encoding for the new TSIPA.

Previous versions of TSIPA (1.0, 1.1) have had the following
limitations, mainly due to the limited number of characters in a font.

 - Automatic kerning by ligtable was impossible because they needed
   to be used with symbols from other fonts (cmr and cmsy).

 - It was necessary to create pk fonts of cmr10 scaled to the same
   point size of tsipa fonts (for example, tsipa10 at 9pt should be
   accompanied by cmr10 at 9pt, not cmr9). This is cumbersome, and
   difficult to explain to those who are not familiar with Metafont.

 - In the cases of symbols which are made up by combining several
   glyphs, sometimes slight gaps between the glyphs were found
   according to the resolution of output devices.

 - Selection of symbols and their order in the font (encoding) were not
   systematic.

I hope all these limitations will be obliterated by the above changes
from (4) through (6).

Now the remaining problem is how to decide the new IPA encoding. As was
suggested in this mailing list a little while ago, it may be helpful to
assign character codes so that most frequently used symbols can be
directly represented by a single letter, not by a lengthy macro name in
an environment in which IPA font is selected(for example, '?' for
glottal stop, ':' for length mark, E for open e(epsilon) or schwa, O for
open o, S for esh, etc). I would like to hear your opinion on this
problem.

I would also like to hear from you if there are any symbols missing in
the previous versions of TSIPA.

Regards,

----------------------------------------------------------------------
FUKUI Rei, University of Tokyo
fkr@tooyoo.L.u-tokyo.ac.jp

20th January 1996: Re: On the next version of TSIPA

From: Don.Colton@cse.ogi.edu
Subject: Re: on the next version of TSIPA

   ... As was suggested in this mailing list a little while ago, it
   may be helpful to assign character codes so that most frequently
   used symbols can be directly represented by a single letter, not by
   a lengthy macro name in an environment in which IPA font is
   selected(for example, '?' for glottal stop, ':' for length mark, E
   for open e(epsilon) or schwa, O for open o, S for esh, etc). I
   would like to hear your opinion on this problem.

Here is a single-character mapping used by the DECtalk speech
synthesizer.  I think it is a good starting list.

  a  as in father
  @  as in bat
  ^  as in but
  o  as in bought
  W  as in bout
  x  schwa
  A  as in bite
  b  as in bet
  C  as in chin
  d  as in debt
  D  as in this
  E  as in bet
  L  as in bottle
  N  as in button
  e  as in bake
  f  as in fin
  g  as in guess
  h  as in head
  I  as in bit
  |  as in kiss_e_s
  i  as in Pete
  j  as in gin
  k  as in Ken
  l  as in let
  m  as in met
  n  as in net
  G  as in si_ng_
  o  as in boat
  O  as in boy
  p  as in pet
  r  as in red
  R  as in b_ir_d
  s  as in sit
  S  as in shin
  t  as in test
  T  as in thin
  U  as in book
  u  as in lute
  v  as in vest
  w  as in wet
  y  as in yet
  Y  as in cute
  z  as in zoo
  Z  as in leisure

i think i would largely agree with these.  especially D, S, Z, T.

20th January 1996: Re: On the next version of TSIPA

From: Don.Colton@cse.ogi.edu
Subject: Re: on the next version of TSIPA

   From: Don.Colton@CSE.OGI.EDU

     o  as in bought
     ...
     o  as in boat

oops, should be:

     c  as in b_ough_t
     ...
     o  as in boat

- don

20th January 1996: Re: On the next version of TSIPA

From: cthiele@ccs.carleton.ca (Christina Thiele)
Subject: Re: on the next version of TSIPA

What wonderful zeal and energy displayed by this announcement!

I wonder, though, if I might bring up a few suggestions and see how
others on this list feel about them.

a. if TSIPA is going to get a whole new encoding and be fully
   compatible with LaTeX2e, might it not be a good idea to also
   slighly modify the name `TSIPA', to reflect this major 
   upgrade? Actually, it's probably a necessary thing, to change
   the filename, since there will be more characters, more styles
   ... all kinds of goodies to make it an even more desirable
   IPA font to use. 

   For those who are exchanging files with colleagues and such,
   and for those who archive their files and would like to be
   able to access old files and print them just as they used to
   print (one of the chief aims and principles of TeX, after
   all), then it would make sense to ensure that no confusion
   -- or worse, incompatibilities -- could arise from using the
   pre-LaTeX2e fonts we know as TSIPA.

b. Perhaps a modified name such as ... oh ... `etsipa' or, 
   maybe a special name just for the beta version, in case you'd
   like some testing done before final release?

Christina

20th January 1996: Single character mapping on the next version of TSIPA

From: Don.Colton@cse.ogi.edu
Subject: single character mapping on the next version of TSIPA

   As for lowercase letters, they all have specific definitions in IPA
   so that I want to retain the original IPA usages. So the main
   candidates for this mapping would be uppercase letters, other symbols
   including punctuation marks, and, possibly, numerals.

i think this is right, and i agree with this list you sent:

     A  \cursa       Cursive A        a as in father
     B  \ipabeta     beta             voiced bilabial fricative
     D  as in this
     E  \ipaepsilon  Epsilon          open e
     F  \ipaphi      Phi              voiceless bilabial fricative
     G  \ipagamma    Gamma            voiced velar fricative
     H  \hooltoph    Hooktop H        voiced h
     I  as in bit
     N  \eng         Eng              ng as in sing
     O  \openo       Open O           open o
     R  \scr         Small capital R  uvular trill (German R)
     S  as in shin
     T  as in thin
     U  as in book
     X  \ipachi      Chi              voiceless uvular fricative
     Z  as in leisure

this leaves the following uppercase letters: C J K L M P Q V W Y

how about also using the following:

     @  as in bat

- don

21st January 1996: Re: On the next version of TSIPA

From: FUKUI Rei 
Subject: Re: on the next version of TSIPA 

Thank you for providing us an example of single-character mapping.
I think I can agree with the following cases:

  D  as in this
  I  as in bit
  S  as in shin
  T  as in thin
  U  as in book
  Z  as in leisure

As for lowercase letters, they all have specific definitions in IPA
so that I want to retain the original IPA usages. So the main
candidates for this mapping would be uppercase letters, other symbols
including punctuation marks, and, possibly, numerals.

Below is a list of what I have in mind for some of the rest of
uppercase letters.

     TSIPA-1.1    Symbol name      Example or meaning
  A  \cursa       Cursive A        a as in father
  B  \ipabeta     beta             voiced bilabial fricative
  E  \ipaepsilon  Epsilon          open e
  F  \ipaphi      Phi              voiceless bilabial fricative
  G  \ipagamma    Gamma            voiced velar fricative
  H  \hooltoph    Hooktop H        voiced h
  N  \eng         Eng              ng as in sing
  O  \openo       Open O           opoen o
  R  \scr         Small capital R  uvular trill (German R)
  X  \ipachi      Chi              voiceless uvular fricative

----------------------------------------------------------------
FUKUI Rei, University of Tokyo
fkr@tooyoo.L.u-tokyo.ac.jp

21st January 1996: Re: On the next version of TSIPA

From: KNAPPEN@VKPMZD.kph.Uni-Mainz.DE
Subject: Re: on the next version of TSIPA

Hello FUKUI Rei,

thanks for taking this project. You mayknow, I attempted to do this about 
1 and a half year ago, and even had a talk at EuroTeX94 (Gda\'nsk) on this
theme. However, I have given up the project, because I overtook the 
coordination of the dc/ec font development.

BTW, I like the nname TIPA very much. Good idea!

Please find included 1) Proposed font positions, 2) Proposed additional 
characters.

1) Font positions:

I will give the positions in order of general acceptance.

The first group is almost out of debatte, any known scheme adheres to it:

Lowercase a-z are just the known lowercase letters.
a ist the usual roman style a
g is ipa italic style g

The following assignments of capitals are generally used and accepted:

A italic style a
B ipa beta
D ipa edh (see also note below on iso-latin-1)
E open e (ipa epsilon)
F ipa phi
G ipa gamma 
I ipa iota/small cap I (no difference in meaning)
L turned y
M `emge'
N eng
O open o (turned c)
S esh (like integral sign)
T ipa theta
U closed omega/`mho' (turned small cap omega)/small cap U 
  (no difference in meaning)
V italic style v
W turned m
X ipa chi
Z ezh

The following SYMBOLS are generally used and accepted

@ schwa
: colon (length mark)
; half length mark
' primary stress mark
, secondary stress mark
| pipe (vertical bar)


The second group are somewhat less accepted, and the may be some dispute 
over it. You -- the designer -- are the last instance to decide any 
dispute.

C, J: In the spirit of ipa1990, the IMHO best assignment is
      c with curl and j with curl.
      (More traditional schemes assign c with cedilla to C, but for c with 
      cedilla see below)) 
R: either fishhook r (flap) or small capital R
Y: either turned h or small capital Y
  (In both cases, the small capital would be my personal preference)
H: hooktop h 
K: turned k or hooktop k (both rare)

The following ipa charaters have iso-latin-1 values, and should be placed 
on that positions (identical to their position in the dc fonts):

c with cedilla  (oct 347)
ae ligature     (oct 346)
o with stroke   (oct 370)

(incidentally, edh has iso-latin-1 position oct 360, however it is better 
placed on capital D. There also no concurring assignment for capital D I 
can think of).

The following two aren't stricktly in latin-1 but have tentative positions:

enj --> oct 361 (is spanish n with tilde)
oe ligature --> oct 367 (is $\div$, but oe in DEC international character 
   set). oe could alternatively place on the capital o with stroke on 
   position oct 330

Now comes the most controversial part:

We still have the capital letter P and Q free of assignments, and there is 
missing a place for the glottal stop. IMHO, the question mark (?) should 
just print a question mark, it is a valuable sign to denote unidentified 
sounds or pronounciations of questionable acceptability.

The shape of P is nicely hinting the glottal stop sign, this would make a 
nice assignment. On the other hand, Q is used in some tradition. However, Q 
made a nice mirror-P, thus the mirror-glottal-stop could nicely be assigned 
to Q. 


2) I propose the following addition to tipa, which aren't currently in 
tsipa:

hv ligature (gothic, in wsuipa)
z with comma-tail (old high german, middle high german)

cyrillic soft sign
cyrillic hard sign

mirror-ezh

double grave accent (used as a tonal mark in slavic languages)

If there is really too much space left, some more accent combinations can 
be added:

./  (dot and acute accent, placed horizontally together)
\.  (grave and dot accent, placed horizontally together)
^  
.   circumflex over dot accent
~
.   tilde over dot accent

Yours, J"org Knappen.

21st January 1996: Re: On the next version of TSIPA

From: FUKUI Rei 
Subject: Re: on the next version of TSIPA 

Thank you for your comments. I understand all the difficulties which
might arise if I cotinue the work of revision with the same name.

As for changing the name, I myself have a wish for that from other
reasons. If I add a new font series like *sl10, *bx10, it will be
convenient to keep the prefixing part of the name within 4 letters so
as to confine the total file name length within 8 letters.

How about `tipa', which can mean Tokyo IPA, TeX IPA or anything
else (for that I don't care)?

----------------------------------------------------------------
FUKUI Rei, University of Tokyo
fkr@tooyoo.L.u-tokyo.ac.jp

21st January 1996: Re: Single character mapping on the next version of TSIPA

From: FUKUI Rei 
Subject: Re: Single character mapping on the next version of TSIPA 

> how about also using the following:
> 
>      @  as in bat

It can be, but in this case TeX's own name \ae is short and easy to
remember so I think @ can be reserved for schwa.

Rei

22nd January 1996: Re: On the next version of TSIPA

From: FUKUI Rei 
Subject: Re: on the next version of TSIPA 

Many thanks for your detailed explanation on IPA encoding!! I think I
can agree with you on most of the points. I will make a draft of TIPA
encodig in a few days.

----------------------------------------------------------------
FUKUI Rei, University of Tokyo
fkr@tooyoo.L.u-tokyo.ac.jp

22nd January 1996: Re: On the next version of TSIPA

From: cthiele@ccs.carleton.ca (Christina Thiele)
Subject: Re: on the next version of TSIPA

KNAPPEN@VKPMZD.KPH.UNI-MAINZ.DE writes:
> 
> Hello FUKUI Rei,
> 
> thanks for taking this project. You mayknow, I attempted to do this about 
> 1 and a half year ago, and even had a talk at EuroTeX94 (Gda\'nsk) on this
> theme. However, I have given up the project, because I overtook the 
> coordination of the dc/ec font development.

For those who have a copy of the proceedings from Gdansk, Jorg's
article appears on pp. 108--09 (``Towards a 256-character IPA font''). 

 
> BTW, I like the nname TIPA very much. Good idea!
>
> ...

I agree. I think a 4-letter name will certainly allow for more
flexibility in font naming -- and `TIPA' sounds pretty good too!

Ch.

23rd January 1996: Re: On the next version of TSIPA

From: Martin.Haase@CL-KI.Uni-Osnabrueck.DE (Martin Haase)
Subject: Re: on the next version of TSIPA

Here are some comments on Joerg's ipa encodings for an
ipa-environment. First of all, I agree with Joerg on almost all
points. Here are the divergences...

>1) Font positions:

>I ipa iota/small cap I (no difference in meaning)
well, small cap I is preferable

>V italic style v
I would plead for inverse v here instead of italic style v, this is
standard transliteration practice in Amerindian studies. Inverse v
represents a rather common vowel, I haven't seen italic style v very
often (as a phoneme at least). Of course, this depends on the
languages you're working on and is no valid argument here.

>: colon (length mark)
the length mark is not a colon of course, but that is pure terminology

>C, J: In the spirit of ipa1990, the IMHO best assignment is
>      c with curl and j with curl.
j with a curl is again less commonly found than inverse f, I would
prefer the latter for J -- C = curly c is a very good placement

>      (More traditional schemes assign c with cedilla to C, but for c with 
>      cedilla see below))
ok
 
>R: either fishhook r (flap) or small capital R
I plead for the flap which is quite common

>Y: either turned h or small capital Y
>  (In both cases, the small capital would be my personal preference)
Joerg's preference is understandable from a German point of view; as a
Romanist I would prefer turned h, if there is no general solution for
inversed signs. Or put inv h on 4, at least there is some resemblance.

>H: hooktop h 
>K: turned k or hooktop k (both rare)
hooktop k is not rare in indo-european studies. If I'm not mistaken
turned k is only used in Germany, and has no official status


>IMHO, the question mark (?) should 
>just print a question mark, it is a valuable sign to denote unidentified 
>sounds or pronounciations of questionable acceptability.
quite so, I totally agree

>The shape of P is nicely hinting the glottal stop sign, this would make a 
>nice assignment. On the other hand, Q is used in some tradition. However, Q 
>made a nice mirror-P, thus the mirror-glottal-stop could nicely be assigned 
>to Q.
This is an excellent idea.  

Additions:

>cyrillic soft sign
>cyrillic hard sign
good idea, though not part of the official set people use them

Of course, lots of suprasegmentals and diacritics are necessary.
Please consult: Journal of the International Phonetic Association 23
(1), 1993 (center pages).

And here is another good book for detailed explanation of phonetic
encodings:
Heepe, Martin (1983): _Lautzeichen und ihre Anwendung in verschiedenen
Sprachgebieten._ -- Hamburg: Buske.

Here are some ideas of how to use the number keys. Of course, these
will be controversial:
0 bilabial click (bull's eye)
2 ejective (similar to glottol stop, but smaller -- not really
official, I suppose)
3 reversed ipa epsilon
4 inverse h
7 belted l
6 tophooked b

It is desireable to have an easy convention for upside down letters,
since there are some, for retroflex sounds and so on. The more I think
about it, the more difficulties I see.

Good luck,
Martin
-----------------------------------------------------------
Martin Haase, Universitaet FB 7, D-49069 Osnabrueck
Tel.:+49-541-969-4340, FAX:+49-541-969-4256
priv.: Hacklaenderstr. 7, D-49074 Osnabrueck, +49-541-29855
http://hal.cl-ki.Uni-Osnabrueck.de/~haase

31st January 1996: Re: On the next version of TSIPA

From: FUKUI Rei 
Subject: Re: on the next version of TSIPA 

Sorry for being late in preparing a new encoding for tipa.
What follows is my opinion on the most controversial points.

   Knappen                   Haase          Fukui
I  ipa iota/small cap I      small cap I    small cap I
V  italic style v            turned v       turned v
J  j with a curl             inverse f      j with a curl (?)

I agree with Haase in that inverse f is more common, but j with a curl
is now an official IPA symbol and make a good pair with c with a curl.

R  fishhook r / small cap. R fishhook r     fishhook r
Y  turned h / small cap. Y   turned h       turned h
K  turned k / hooktop k      ?              reversed small cap. R

There is no good symbol for the place K. I'd like to propose reversed
small capital R (as in French r) which somehow looks like K.
(I'm not sure whether this is MY proposal or I've seen this somewhere
else or in the opinion of someone else.)


(J"org Knappen)

>' primary stress mark
>, secondary stress mark

I prefer " for primary stress, because ' is used for ejectives.
And , can be preserved for for its normal use.

>hv ligature (gothic, in wsuipa)

This is already included in tsipa-1.1.

>z with comma-tail (old high german, middle high german)

OK.

>cyrillic soft sign
>cyrillic hard sign

As they are in a cyrillic font, I'm not sure about the necessity for
these two in a ipa font.

>mirror-ezh

OK.

>double grave accent (used as a tonal mark in slavic languages)

This is already included in tsipa.

>./  (dot and acute accent, placed horizontally together)
>\.  (grave and dot accent, placed horizontally together)
>^  
>.   circumflex over dot accent
>~
>.   tilde over dot accent

These accents can be made by macros. 


(Martin Haase)

>0 bilabial click (bull's eye)
>2 ejective (similar to glottol stop, but smaller -- not really
>official, I suppose)
>3 reversed ipa epsilon
>4 inverse h
>7 belted l
>6 tophooked b

I'm inclined to use numerals as they are, because in the study of
suprasegmentals, numbers are sometimes used to indicate tones,
intonation, etc. But still I'm not sure.

----------------------------------------------------------------
FUKUI Rei, University of Tokyo
fkr@tooyoo.L.u-tokyo.ac.jp

1st February 1996: Re: On the next version of TSIPA

From: KNAPPEN@VKPMZD.kph.Uni-Mainz.DE
Subject: Re: on the next version of TSIPA

Fukui,

first I want to mention, that it is easy (within the context of LaTeX2e)
to define a generic macro \* to turn letters around by adopting techniques 
from the T$n$enc.def files. The `turning' is achieved by table look-up.

Thus, \*v, \*h, \*f etc. are possible for the input of turned v, h and f 
respectively.

To the details:

I: small cap I   is perfectly fine with me
Y: small cap Y still preferred, because it make a nice symmetrie
   i:I, u:U, and y:Y. For turned h, I propose \*h
turned v: In usual ASCII ipa schemes, turned v is assigned to some symbol,
   either & or ^. However those symbols have special meaning within TeX and
   generally it is a bad idea playing around with them. If \*v is good 
   enough for turned v, I plea for 
V: italic style v.

Your thoughts on making K: small cap R are interesting. Since there are too 
many different r's around, it is a good idea, I think.

Yours,
J"org Knappen.

8th February 1996: Conference announcements

From: Csg0070@Queens-Belfast.AC.UK
Subject: Conference announcements

Would it not be an idea to create a separate e-mail list for TeX-related
conference announcements, and thus avoid multiple copies of (long!)
conference announcement for those of us who are subscribed to multiple
TeX lists?

Ciar\'an \'O Duibh\'in.

9th February 1996: Re: Conference announcements

From: Piet van Oostrum 
Subject: Re: Conference announcements

>>>>> Csg0070@Queens-Belfast.AC.UK (C) writes:

C> Would it not be an idea to create a separate e-mail list for TeX-related
C> conference announcements, and thus avoid multiple copies of (long!)
C> conference announcement for those of us who are subscribed to multiple
C> TeX lists?

How many of these announcements are there? Four per year or so? I don't
think the volume warrants a separate list. I don't think a separate list
will stop people posting announcements on these lists.
-- 
Piet van Oostrum 
http://www.cs.ruu.nl/~piet

11th February 1996: Alpha-test version of tipa

From: FUKUI Rei 
Subject: alpha-test version of tipa

I have just put files for tipa alpha-test version on:

  ftp://tooyoo.L.u-tokyo.ac.jp/pub/TeX/tipa

This version consists of only Metafont sources and a few TeX files for 
viewing character table of tipa. No style files are included.

The encoding of tipa is not yet decided. The purpose of this test
version is to show a tentative encoding and all the symbols made so
far. I am expecting feedbacks from many people.

A few notes on the contents of this test version:

  - there are seven empty slots.
  - symbols found in the area '200--'223 are intended to be parts of
    `tone letters'.
  - basic IPA symbols that cannot be allocated in the area of '000--'177
    are located in the area '340--'377.
  - less frequent symbols than the ones in '340--'377 are located in
    the area '300--'337.
  - old IPA, non IPA symbols are located in '240--277.

FUKUI Rei

13th February 1996: Journals on Linguistics (fwd)

From: cthiele@ccs.carleton.ca (Christina Thiele)
Subject: Journals on Linguistics (fwd)

I received the following query from Gabriel Valiente Feruglio
(valiente@lsi.upc.es) and suggested that this list might be a good
source of information. 

What he's looking for is linguistics journals accepting (La)TeX
submissions. I can only speak for the _Canadian Journal of
Linguistics_, where TeX is an in-house tool. There are no style files
or macros or guides or anything like that for authors -- I'll take
(almost) anything in electronic form and then code it.

Same went for the _Papers of the Algonquian Conference_ until 1994
when editorship and production moved to Manitoba.

Actually ... this is a query that's been needed for a long time ;-)
The ling-tex list is for the intersection of TeX and linguistics ...
and the results of that intersection are often intended for
publication somewhere ... eventually ... ;-) But what do we know about
TeX and linguistics journals?

So, perhaps members of the list could respond here -- for our own
edification -- and then I'll bundle all the info up for Gabriel. 

Note that he's looking for specific details so if you do have info,
try to be as complete as possible ... and if there's a gap here and
there, someone else on the list can probably provide the missing info. 

Actually ... let me post another message with just the headings of
stuff he's looking for ... and maybe at the same time, I'll add a few
other headings, of interest to us here ... 

Christina


=====================================
Gabriel Valiente Feruglio writes:
> ...
>
> Hi, Christina. I'm compiling a list of printed and/or electronic journals
> that accepts LaTeX submissions and that have some WWW page, and I thought
> you would be the right person to ask for such information within the field
> of linguistics. Of course, if you have such information for other fields
> as well, any help is welcome. The data I look for is the name of the
> journal, the field, ISSN number for the printed edition and/or the electronic
> edition, and the WWW address for its home page on the Internet, although
> I could find out the data by myself if I'm given some WWW addresses.
> Thanks a lot in advance. Best regards,
> 
> Gabriel
> 

=======================================

13th February 1996: Linguistic journals and (La)TeX

From: cthiele@ccs.carleton.ca (Christina Thiele)
Subject: Linguistic journals and (La)TeX

                     Linguistics Journals and TeX
                     ============================

Name of Journal (and usual acronym): 

Publisher:

ISSN number 
   a. printed edition:
   b. electronic edition:

WWW address:

============

Issues per year:

Pages per issue (on average):

Electronic submissions in 
   a. any format
   b. specific format(s)

TeX or LaTeX use: 
   a. in-house only
   b. submissions from authors

Style guide available in 
   a. printed edition
   b. e-form (include address)

TeX materials used for production 
   a. style files
   b. fonts (text font, which IPA)

Production environment
   a. computer
   b. printer

========

When did (La)TeX use begin (specify which is used)?

Contact person and e-mail/snail address:

=============================================================

14th February 1996: Re: Linguistic journals and (La)TeX

From: cthiele@ccs.carleton.ca (Christina Thiele)
Subject: Re: Linguistic journals and (La)TeX

.. might as well start things off with this ... 

> 
> 
>                      Linguistics Journals and TeX
>                      ============================
> 
> Name of Journal (and usual acronym): 

   Canadian Journal of Linguistics/Revue canadienne de linguistique
      CJL/RCL

> Publisher: 

   Canadian Linguistic Association/Association canadienne de linguistique

> ISSN number 
>    a. printed edition:

   0008-4131      

>    b. electronic edition:

   None

> WWW address:

   None

> ============
> 
> Issues per year:

   Four

> Pages per issue (on average):

   96 pages

> Electronic submissions in 
>    a. any format
>    b. specific format(s)

   Any format -- any machine (Mac, DOS, Windows, UNIX, via e-mail ... )

> TeX or LaTeX use: 
>    a. in-house only
>    b. submissions from authors

   In-house only; rarely are .tex files received (I can only recall 2
     articles for certain since 1984, and perhaps 3-5 reviews in that
     same time period). 

> Style guide available in 
>    a. printed edition
>    b. e-form (include address)

   Printed edition only -- latest version appeared in CJL 38.4 (1993)
   pp.475-486. A new version is scheduled to appear in 40.4 (1995).

> TeX materials used for production 
>    a. style files
>    b. fonts (text font, which IPA)

   From 1984 until 1993, in-house plain TeX macros were used.
   Initially, an in-house bitmapped IPA was used (PH10), but in
   1986/87, the WSUIPA fonts were introduced. 

   In 1994, the journal was switched over to LaTeX, mainly to take
   advantage of all the wonderful style files that were coming to
   light via the inventory started on the ling-tex list. The
   style files currently in use include:

   a. cm-lingmacros.sty (Christopher Manning)
   b. tree-dvips.sty (Emma Pease)

   Fonts: 1984 -- 1988: AMR fonts (+ PH10 till 1986/7, then WSUIPA)
          1988 -- 1994: CMR fonts (+ WSUIPA)
          1995 --     : Adobe Times Roman (+ WSUIPA)

> Production environment
>    a. computer
>    b. printer

   1984 -- 1991: Honeywell mainframe running CP-6
   1991 --     : UNIX workstation

   1984 -- 1992/3: ImageStation (300dpi)
   1993 --       : LaserJet 4M  (600dpi)

> ========
> 
> When did (La)TeX use begin (specify which is used)?

   See above.

> Contact person and e-mail/snail address:

Editor: Mme. Anne Rochette
        Univ. du Quebec a Montreal
        cjlrcl@er.uqam.ca

Or myself: Christina Thiele
           cthiele@ccs.carleton.ca
> =============================================================

14th February 1996: Re: Linguistic journals and (La)TeX --- JNLP

From: bond@kssalt.cipl.cae.ntt.jp (Francis Bond)
Subject: Re: Linguistic journals and (La)TeX --- JNLP

                     Linguistics Journals and TeX
                     ============================

Name of Journal (and usual acronym): 
	Journal of Natural Language Processing (JNLP)

Publisher:
	Tokyo Shoseki Insatsu

ISSN number 
   a. printed edition:		1340-7619
   b. electronic edition:

WWW address:			http://www.kyutech.ac.jp/nlp/index.html	

============

Issues per year:			4

Pages per issue (on average):		90

Electronic submissions in		(b) jLaTeX
   a. any format
   b. specific format(s)		

TeX or LaTeX use:			(b) submit with style file
   a. in-house only
   b. submissions from authors

Style guide available in		
   a. printed edition			(a) back of journal
   b. e-form (include address)		
		(b) ftp.kyutech.ac.jp pub/society/NLP/nlpsty.tar.gz

TeX materials used for production 
   a. style files			
	(a) Japanese paper:
		jnlp_j.sty,  jnlpbbl.sty, jnlpbbl.bst, nlpbbl.sty
	    English paper:
		jnlp_e.sty, nlpbbl.sty, nlpbbl.bst
	    Non-papers
		jnlp_o.sty

   b. fonts (text font, which IPA)
		NOTE: Uses jLaTeX based on jTeX, an TeX version that allows
	use of 2-byte Japanese characters.

Production environment
   a. computer
   b. printer

No idea, most of the checking and final formatting is done at the
Kyushu Institute of technology.

========

When did (La)TeX use begin (specify which is used)?

It has been in jLaTeX right from the beginning.

Contact person and e-mail/snail address:

Jun'ichi Nakamura

fax: 0948-29-7601, e-mail: nakamura@ai.kyutech.ac.jp

=============================================================

21st February 1996: Beta test version of tipa

From: FUKUI Rei 
Subject: beta test version of tipa

A beta test version of TIPA is now available from:

  ftp://tooyoo.L.u-tokyo.ac.jp/pub/TeX/tipa

Changes from the alpha version are as follows:

  - sans serif style (tipass) is created.
  - all the size varieties are prepared.
  - a tentative style file and font description files are made.
  - two documents `codelist' and `symblist' are made. The latter is an
    annotated list of TIPA symbols. These two documents are also
    available in the form of postscript file (made by dvips).

FUKUI Rei

6th March 1996: Prosody and Metrics Symbols for TeX

From: Egisto Sani 
Subject: Prosody and Metrics Symbols for TeX

Dear Sir,
my wife is working at Pisa University, Department of Greek Philology, where she
teaches Greek Metrics.
In the past, she edited her papers using TeX and some beautiful Greek fonts I
down-loaded from the net.
Now she needs some special symbols suitable to describe the prosody and metrics
of the old Greek and Latin poems, but, unfortunately, I was not able to find
any character set suitable for this purpouse.

Could you help my in this search, putting me through with someone who's had and
solved (I hope) this problem ?

Thank you

Egisto Sani


PS Technical specification:
- we are running our TeX Version 3.14159 (C version 6.1), and related programs
on a SGI Indy (o.s. IRIX 5.2) or under LINUX on a PC-IBM. We can print our
papers by using a Laserjet III printer or a Digital Postscript printer.
- The LaTeX2e patch level is the N. 3.
- We get our programs and the annexed font and macro set from the teTeX
distribution.
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
EGISTO SANI
c/o ENEL - CRT
Via Andrea Pisano 120, 56122 Pisa, ITALY
e-mail sani@crt.enel.it
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

11th March 1996: Seeking details on `ALLC/ACH' conference

From: cthiele@ccs.carleton.ca (Christina Thiele)
Subject: seeking details on `ALLC/ACH' conference

I've been asked to check into a conference listing for more details.
The _fonetiks_ newsletter for February, which just came i,n doesn't
include it but perhaps someone reading this list might have some
suggestions.

>     25-29 June 6 1996. Joint International Conference ALLC/ACH'99
>     (Association for Literary and Linguistic Computing, Association 
>     for Computers and the Humanities), University of Bergen, Norway.
>     Contact: Espen S. Ore


Does anyone have any leads on this? (Oh, we're not sure if that
``ACH'99'' is correct or a typo.) 

Ch.

11th March 1996: ALLC/ACH found -- thanks to all who replied

From: cthiele@ccs.carleton.ca (Christina Thiele)
Subject: ALLC/ACH found -- thanks to all who replied

What a list! 

Thanks to Greg Lessard, Evan Antworth, and Chris Manning, I've now got
all the gory details of this joint conference in June of this year ...
and yes, that typo of `ACH'99' was in the original snippet of info we
had.

Uh ... if anyone does want more details, I now possess them in all
their glory.

Ch.

12th March 1996: A new beta version of tipa

From: FUKUI Rei 
Subject: A new beta version of tipa

A new beta version of tipa is now available from:

  ftp://tooyoo.L.u-tokyo.ac.jp/pub/TeX/tipa/beta0312

as before.

This time the postscript file of the manual is placed in:

  /pub/TeX/tipa/tipaman.ps

not in the beta0312 subdirectory.

Regards,

FUKUI Rei

15th March 1996: More on tipa encoding

From: FUKUI Rei 
Subject: more on tipa encoding

Hi all.

Tipa is now on its final stage of development toward the official
release. The last beta version (beta0312), which I announced a few
days ago, includes substantially almost all the things I had in mind
when I announced the idea of developing a new version of tsipa except
for one thing, that is, to create a slightly modified version of fonts
that make a good harmony with Times roman fonts. Now it seems
difficult to find free time to do so at this time so that I'd like to
release the first version of tipa without Times roman compatible
fonts.

For the past few weeks, I have been busy with struggling with font
designing, making TeX macros and preparing documentation, forgetting
for a while the problem of encoding. But now, at this stage, several
questions arise as to the selection and exact shapes of symbols. And
it is not difficult to delete/add some symbols now.

Before expressing my questions, it is necessary to explain on thing which
I myself had been forgetting until quite recently.

The selection of *tsipa* symbols was based on an article `Computer
coding of the IPA: supplementary Report' by John Esling, which
appeared in the Journal of the International Phonetic Association,
vol. 20, no. 1, 1990. Tsipa included all the symbols that appeared in
this article. And not only the selection of symbols but the naming and
shapes of symbols were also based on it.

The questions I have now are as follows:

(1) tsipa (and the beta version of tipa) includes the following two
accents.

  Acute accent (under)
  Grave accent (under)

These accents are placed under a symbol and, moreover, their shapes
are not just the ordinary acute and grave accents. They are
inverted. But I found no sources in which such accents are recorded,
except for the above mentioned article. In the 49 and 79 versions of
IPA, acute and grave accents placed under a symbol are included but
their shapes are not inverted.

So I'd like you ask you whether there are any other sources in which
such inverted accents are recorded. If not, placing acute and grave
accents under a symbol can be made by a macro, and this saves two
slots in the tipa encoding.

(2) I have some doubt on the shapes of the following two symbols.

  Stretched C
  T-S Digraph (in tipa, I renamed this `T-S ligature')

In the early literature on phonetics, the height of stretched c equals
to x_height and stretching was only made toward the descender
region. But in the above article (and also in the ``Phonetic Symbol
Guide'') stretching is also made toward the ascender region, and the
entire shape looks like a thick staple rather than a stretched c.
I am now tempted to reshape this symbol to its original. Does anyone
have a definite idea on the exact shape of this symbol?

In the case of T-S Digraph, the shape appeared in the above article
differs from the ones found in the 49 and 79 versions of IPA, in that
in the original IPA publications, t and s are more closely linked and
s is not as tall as an ordinary s. I'm not sure which shape is
preferable.

(3) The last beta version of tipa has three empty slots. And I newly
created two symbols which might be the candidates for these
slots. These are:

  Reversed Esh with Top Loop
  Bent-tail Yogh

These two symbols appeared in 49 version of IPA but the approval was
withdrawn later. So I didn't included these symbols in the beta
version, keeping them in a separate file `extra.mf'.

If I were creating a museum of phonetics symbols, fossils are precious 
things but I'd like to make tipa a more useful tool for those actually 
working in many fields of phonetics. So I am seeking better candidates.

(Oh, too much babble on tiny problems. But there are so many minute
and often trifling things to consider. Sigh.)

----------------------------------------------------------------
FUKUI Rei, University of Tokyo
fkr@tooyoo.L.u-tokyo.ac.jp

4th April 1996: Ancient Greek hyphenation

From: Yannis Haralambous 
Subject: Ancient Greek hyphenation

I finally got access to the Thesaurus Linguae Grecae, and would like
to use texts from there to test my Ancient Greek hyphenation patterns.
In your opinion, what would be the best texts to start with? is there
any volunteer to check the hyphenated text and spot wrong hyphens?

Thanks in advance
Yannis

 +-----------------------------------------------------------------------+
 | Yannis Haralambous, Ph.D.           Yannis.Haralambous@univ-lille1.fr |
 +-----------------------------------------------------------------------+
 | 187, rue Nationale                              fax: (33) 20.40.28.64 |
 | 59800 Lille, France                    ISDN/Numeris: (33) 20.15.81.77 |
 +-----------------------------------------------------------------------+
                            ...the ball I threw while playing in the park
                                             has never reached the ground

23rd April 1996: TLG

From: Yannis Haralambous 
Subject: TLG

Does anyone know where I can find a complete description of the TLG syntax?

thanks in advance

Yannis

29th April 1996: A new beta version of tipa

From: FUKUI Rei 
Subject: a new beta version of tipa

Hi. I've just placed a new beta version of tipa in:

  ftp://tooyoo.L.u-tokyo.ac.jp/pub/TeX/tipa  (130.69.116.39)

   beta0429.tar.gz ... tar and gzipped package
   beta0429/       ... same thing, extracted
   tipaman.ps.gz   ... new manual produced by dvips (600dpi)
   tipaman.ps      ... same thing, ungzipped

Everyone seems busy these days, and I'm busy too. So the development
of tipa is very slow, but the final version will be coming soon.

This time I attached a separate package vowel.sty which is useful for
drawing vowel diagrams. This package is not yet fully documented.

----------------------------------------------------------------
FUKUI Rei, University of Tokyo
fkr@tooyoo.L.u-tokyo.ac.jp

13th June 1996: Re: HELP NEEDED

From: FUKUI Rei 
Subject: Re: HELP NEEDED 

I'll look at your work on IPA font this weekend. Right now I'm busy
with so many things to do. Please wait for a couple of days.

----------------------------------------------------------------
FUKUI Rei, University of Tokyo
fkr@tooyoo.L.u-tokyo.ac.jp

24th June 1996: New beta version of tipa

From: FUKUI Rei 
Subject: new beta version of tipa

Hi. I have just uploaded the new beta version of tipa.

  ftp://tooyoo.l.u-tokyo.ac.jp/pub/TeX/tipa/beta0624.tar.gz

There have been so many chnages since I announced the previous beta
version. 

 - The encoding of tipa is now called the T3 encoding.
 - Macros heavily modified accordingly.
 - A set of new fonts (xipa*) that goes well when used with times.sty
   added. etc.

This directory is now mirrored to ftp.dante.de (under the directory
fonts/tipa) so that it will be also available from a CTAN archive.

Regards,
----------------------------------------------------------------
FUKUI Rei, University of Tokyo
fkr@tooyoo.L.u-tokyo.ac.jp

26th July 1996: New beta version of tipa

Dear ling-tex readers,
I am forwarding my discussion (slightly shortened) with Fukui Rei
(with his permission) on tipa encodings in what follows (all quoted
is mine, all unquoted is Fukui Rei's). Perhaps, somebody would like
to join the discussion.
Best wishes,
Martin

==================BEGIN FORWARDED MESSAGE==================
mh> This reminds me that I owe you a comment:
mh> I looked at your manual from April 29. Congratulations! I would
have
mh> immediately replaced wsuipa, if you hadn't changed some of the
macro
mh> names. For sure, your names are more coherent, but isn't it
possible
mh> to keep on using the old names along with the new ones, just to
mh> ensure compatiblity (e.g. \niepsilon along with \Iepsilon), or at
mh> least produce a compatiblity style-file.
mh> Thank you for considering this problem!

It is good to hear from you again.

Unfortunately, the situation became even worse since I started to
talk
with the LaTeX2e people. Under the general convention of LaTeX2e,
symbols names that can be used in the text mode are supposed to have
a
prefix \text. So \schwa is now called \textschwa, etc. In the case of
\niepsilon and \Iepsilon, although both names are in fact not so
good,
it is now called \textepsilon.

It is possible to make a slightly different style file that contains
symbol names like \niepsilon, etc. but I don't think it is good to
stabilize old symbol names.

Fukui Rei

mh>> prefix \text. So \schwa is now called \textschwa, etc. In the
case of
mh> Well, IMHO this seems a bit exaggerated. Imagine how
mh> long the names become: \textIepsilon etc.
mh> Any sensible user will create abbreviations.

I was INDEED persuaded by Frank Mittelbach to attach the prefix
\text to every text commands. In the recent version of LaTeX2e a few
symbol names were added, such as \dh \dj \ng \th, but in his opnion,
he said, these were `mistakes'.

I can understand what you have said very well since I myself am a
linguist and I usually uses very short `abbreviations' in my papers.
And TIPA now defines various forms of IPA environments in which most
of the frequently used symbols can be represented by a single letter
so that I personally won't care about the long official(?) names.

I wrote, in a reply to Frank, that I don't like extremely long names
like \textepsilon, \textfishhookr, etc. but he wouldn't understand.
The LaTeX2e people are concerned about the official symbol names that
are defined in font encoding files like T1enc.def, T3enc.def etc. but
this doesn't mean that they are against making private or less
official files that contain more familiar names. And, please
understand, I've just started to do the official part of the job and
I
am still unable to found enough free time to do further things.

Frankly speaking, I'm now very much confused because I am doing the
job of making TIPA from a linguist's vewpoint but at the same time I
have to talk with LaTeX2e people who don't pay so much attention to
average users of linguistics.

As for the compatibility with earlier versions, I am not against your
opinion, in principle. But in this case the only problem is that
I haven't been using WSUIPA.


===================END FORWARDED MESSAGE===================

-----------------------------------------------------------
Dr. Martin Haase, Universitaet FB 7, D-49069 Osnabrueck
Tel.:+49-541-969-4340, FAX:+49-541-969-4256
http://www.cl-ki.Uni-Osnabrueck.de/~haase/

25th July 1996: Re: New beta version of tipa

If I understand the arguments of the LaTeX3 team right, they do not 
propose, that the end user should be forced to use the long \text...
names, but that there should be three layers instead of two.

The first layer is a pair (number,font) which characterises a glyph inside 
a given font,

the second (and new) layer are standard LaTeX names beginning with 
\text...,

and the third layer are user names, which are defined in terms of the long 
\text... names.

This approach cost about 260 controll sequences and a lot of characters 
therein in term of TeX memory. The gain of the new approach is an interface 
to fonts of different encoding and different user styles, which makes 
switching e.g. between wsuipa, tsipa and tipa easier.

--J"org Knappen.

25th July 1996: Re: Tipa

Fukuisan,
since you proposed to continue the discussion of tipa encodings in
public, I send a copy of this to ling-tex. Unfortunately, I haven't
kept a copy of our previous correspondence.
If I have understood right, the latex3 people prefer long macronames
for the special characters beginning uniformely with the sequence
\text, like the following:
\textcloserevepsilon
After testing, I come to the following conclusions:
1. At least in the IPA-environment the name without \text should
suffice, because there aren't abbreviations for all signs, and esp.
the more exotic ones don't have an abbreviation nor are their names
short.
2. In tables, one wouldn't like to switch to the ipa-environment,
esp. if there are only individual signs in a row. But of course, in a
table, long names are a nuisance too. 
3. It would even be more coherent to have ipa prefixed to all
ipa-macros. I *do* hope that this does not mean that the macro names
will end like this:
\textipacloserevepsilon :-)
4. In order to make tipa usable for those of us who (like me) have
used WSUipa (and are not at all satisfied with it), a compatibilty
sty has to be added: wsu2tipa.sty or so. If I had an idea how it
should be best designed (should it contain lines like
\let\revepsilon=\textcloserevepsilon, or \def-commands or what?), I
would (time permitting) start working on it, because I have the
strong desire to switch definitely to tipa, but still I have all
these files containing wsuipa-macros (ipamacs.tex).
Hope to hear from you (all)!
Best wishes,
Martin
-----------------------------------------------------------
Dr. Martin Haase, Universitaet FB 7, D-49069 Osnabrueck
Tel.:+49-541-969-4340, FAX:+49-541-969-4256
http://www.cl-ki.Uni-Osnabrueck.de/~haase/

26th July 1996: Re: Tipa

On Fri, 26 Jul 1996 14:57:12 +0900, FUKUI Rei wrote:
>On Thu, 25 Jul 96 15:45:55 +0200 Martin Haase wrote:
mh>> 4. In order to make tipa usable for those of us who (like me) have
mh>> used WSUipa (and are not at all satisfied with it), a compatibilty
mh>> sty has to be added: wsu2tipa.sty or so. If I had an idea how it
mh>> should be best designed (should it contain lines like
mh>> \let\revepsilon=\textcloserevepsilon, or \def-commands or what?), I
mh>> would (time permitting) start working on it, because I have the
mh>> strong desire to switch definitely to tipa, but still I have all
mh>> these files containing wsuipa-macros (ipamacs.tex).

fr>I'd be grateful if you would do the job. As a related thing, I was
fr>requested by Frank Mittelbach to make `OT3enc.def' for wsuipa but I
fr>still don't have time to do so.
fr>
fr>Fukui Rei

Although I'm extremely under timepressure myself, I could look after
it (at least start to), but I haven't got the faintest idea what it
should look
like...
Best wishes,
Martin
PS.: I will be away for 10 days from tomorrow.

-----------------------------------------------------------
Dr. Martin Haase, Universitaet FB 7, D-49069 Osnabrueck
Tel.:+49-541-969-4340, FAX:+49-541-969-4256
http://www.cl-ki.Uni-Osnabrueck.de/~haase/

26th July 96: Re: Tipa

> 1. At least in the IPA-environment the name without \text should
> suffice, because there aren't abbreviations for all signs, and esp.
> the more exotic ones don't have an abbreviation nor are their names
> short.

I think it is possible for define names without \text within the ipa
environment.

> 2. In tables, one wouldn't like to switch to the ipa-environment,
> esp. if there are only individual signs in a row. But of course, in a
> table, long names are a nuisance too. 

But as is explained in the manual, one wouldn't get a proper
inter-word kerning outside the ipa environment. For example in a
string like `a\textschwa' the letter `a' is selected from the default
text font (OT1 or T1) and \textschwa from the tipa font (T3), whereas
in the ipa environment both \textipa{a@} and \textipa{a\textschwa}
give a proper inter-word kerning. So I'd like to recommend to use the
ipa environment wherever possible. I personally use an abbreviated
form of the \textipa command by defining \let\ipa\textipa, and writing
something like

  \ipa{...} & \ipa{...} & \ipa{...} & \ipa{...}

in a table doesn't seem to be so difficult.

> 3. It would even be more coherent to have ipa prefixed to all
> ipa-macros. I *do* hope that this does not mean that the macro names
> will end like this:
> \textipacloserevepsilon :-)

I think you are right. An ipa beta and a greek beta are different
symbols that can coexist in the same encoding.

> 4. In order to make tipa usable for those of us who (like me) have
> used WSUipa (and are not at all satisfied with it), a compatibilty
> sty has to be added: wsu2tipa.sty or so. If I had an idea how it
> should be best designed (should it contain lines like
> \let\revepsilon=\textcloserevepsilon, or \def-commands or what?), I
> would (time permitting) start working on it, because I have the
> strong desire to switch definitely to tipa, but still I have all
> these files containing wsuipa-macros (ipamacs.tex).

I'd be grateful if you would do the job. As a related thing, I was
requested by Frank Mittelbach to make `OT3enc.def' for wsuipa but I
still don't have time to do so.

Fukui Rei

16th August 1996: Re: LaTeX Macros for Linguists

Michael Covington writes:
> 
> I've just released a new version of "covington.sty" (note the new spelling,
> with all 9 letters of my name).
> 
> This is a package of LaTeX macros for such things as sentences with word-by-
> word glosses, phrase-structure rules, feature structures, and Discourse
> Representation Theory.
> 
> The only significant change since the 1992 version is that the new version
> is fully compatible with NFSS and LaTeX2e.
> 
> The new files are located on ftp://ai.uga.edu/pub/tex/covington.sty
> and covington.tex and are accessible through my web page.
> 
> -- 
> Michael A. Covington                http://www.ai.uga.edu/faculty/covington/
> Artificial Intelligence Center                                           <><
> The University of Georgia                Unless specifically indicated, I am
> Athens, GA 30602-7415 U.S.A.                not speaking for the University.
> 



Michael, many thanks for the note about covington.sty -- and for all
the work that went into it!

===

Now ... it's been a while since we've posted the ling-mac.tex file,
which lists all the style files we seem to know about.

Is there anyone else who's done an update of their LaTeX macros for 2e
and/or NFSS? 

If so, why don't you post a note here, and I'll update the style file
inventory for, say ... mid-September or perhaps end of Sept? If you're
thinking of doing an update to your existing style file(s), drop us a
line as well, so we know it's in the works. 

And of course, if there are any new style files or fonts or whatevers
floating around, and you've used them, give us all a shout here on
ling-tex ... which is enjoying a very quiet summer so far ... ;-) 

Thanks, everyone. See you in September.

Christina

P.S. To see the current inventory, check CTAN in /tex-archive/info for
     the file ling-mac.tex

2nd September 1996: Multilingual hypenation

I have a very stupid problem: 
1. I have succeeded in producing a fmt-file with hypenation patterns
for English, German and Italian, but I do not know how to switch from
one hypenation pattern to the other within my document. Actually, I
need hyphenation patterns other than German only in the
quote-environment.
The solution should be rather easy, something like \language=1 or so
within the environment.
2. A perhaps less stupid problem: The French hyphenation file I found
on CTAN seems to be only a compatiblity file, so where's the real
thing?
Thank you for your help
Martin
-----------------------------------------------------------
Dr. Martin Haase, Universitaet FB 7, D-49069 Osnabrueck
Tel.:+49-541-969-4340, FAX:+49-541-969-4256
http://www.cl-ki.Uni-Osnabrueck.de/~haase/

2nd September 1996: Re: Multilingual hypenation

From: Bernard GAULLE 

>>>>> On Mon, 02 Sep 96 13:35:03 +0200,
>>>>>    "Martin Haase"  write
about "multilingual hypenation":
MH> I have a very stupid problem: 
MH> 1. I have succeeded in producing a fmt-file with hypenation patterns
MH> for English, German and Italian, but I do not know how to switch from
MH> one hypenation pattern to the other within my document. 

Various solutions exist. The most commonly used is via the "babel"
package which include "styles" for dozens of languages, including
English, German, Italian and french (at least two french styles).

French people often use my "french" package instead of babel, because
it includes also a switch mechanism between languages. The french
package is now running also together with babel.

MH> Actually, I
MH> need hyphenation patterns other than German only in the
MH> quote-environment.

Be carefull, TeX has a terrible drawback: it can use various languages
but only one (the last one) in a paragraph.

MH> The solution should be rather easy, something like \language=1 or so
MH> within the environment.

oh! it's really ugly to switch that way via an absolute value that
could be changed at any format creation.

Babel mechanism let you use \selectlanguage{french} yet
french mechanism is simpler, just say \french.

MH> 2. A perhaps less stupid problem: The French hyphenation file I found
MH> on CTAN seems to be only a compatiblity file, [...]

??? 
The French hyphenation files you find in the french package or in
the "languages" directory on CTAN are really the running file (but notice
you need at least 2 files, frhyph.tex and f7hyph.tex or f8hyph.tex).

Hope it helps,

  --bg

2nd September 1996: Re: Multilingual hypenation

The canonical way to do multilingual hyphenation is using the LaTeX babel 
package. It comes with a special file lthyphen.cfg, where you declare the 
languages and their hyphenation patterns. There are some pitfalls (e.g. the 
idetifier for french is `francais'). You may want to consult the file 
germdoc.tex (documentation of german.sty by Bernd Raichle) first, it
describes the procedure accurately.

For french, I currently use the pattern fr8hyph.tex (8bit patterns to be 
used best with the fontencoding T1 (dc-fonts)).

AFAIK, multilingual support for plain TeX was never implemented, except 
rudiments.

--J"org Knappen.

4th September 1996: Re: Multilingual hypenation

From: Bernard GAULLE 

>>>>>    KNAPPEN@VKPMZD.KPH.UNI-MAINZ.DE write about "Re: multilingual hypenation":
JK> The canonical way to do multilingual hyphenation is using the LaTeX babel 
JK> package. It comes with a special file lthyphen.cfg, where you declare the 
JK> languages and their hyphenation patterns. 

"canonical" isn't the good word...
I also must correct few things : with babel, as well as with french, 
languages and hyphenation files are declared in "language.dat" file.

JK> There are some pitfalls (e.g. the 
JK> idetifier for french is `francais'). 

"francais" is the first try and was made by J.Braams.
"french" was an alias for "francais", until my French package do the job.
"frenchb" will be the third try, by D.Flipo, in the incoming version.
"french" will remain an alias, for frenchb,   "  "   "        "
I don't know the future status of francais but i guess it will be
removed.
  
JK> For french, I currently use the pattern fr8hyph.tex (8bit patterns to be 
JK> used best with the fontencoding T1 (dc-fonts)).

but you also may use 7bit patterns if you are using MlTeX.

JK> AFAIK, multilingual support for plain TeX was never implemented, except 
JK> rudiments.

As far as i know, babel was runnig for plain and i'm sure that
my French package has multilingual support too. And i've read 
that there are others but not widely spread.

  --bg

9th September 1996: Chinese and TeX

Is it possible to write chinese characters using tex? How does it work?
I would appreciate very much any helpful comment.

Thanks very much for your effort
Markus Klein

9th September 1996: Fwd: Bibliographical references

I forward this with no comment. If I find the time, I shall reply
directly to Oesten. And perhaps others more competent than myself
should do so too. By the way, is there a bibstyle usable for
linguists? As you know, plain-bibstyle-like reference by numbers is
not used in linguistics (and IMHO not very practical).
Best wishes,
Martin

==================BEGIN FORWARDED MESSAGE==================
Date:  Fri, 06 Sep 1996 16:24:19 BST
>From:  oesten@ling.su.se (Oesten Dahl)
Subject:  Standards for bibliographical references

Is there anyone except me who feels the need for a unified way of
writing bibliographical references? You may not believe me, but the
style sheets of two major publishers of linguistics literature,
Benjamins and Mouton, differ on EIGHT points in how a totally ordinary
journal article should be referenced. (After having looked into a few
more publications from these publishers, I have to modify this
statement. Actually, each publisher seems to have several style sheets
which differ between each other no less than the ones I mentioned.)
Every time you submit a paper, you have to spend several hours
studying intriguing details such as single vs. double quotes, the
order of first and last names, full stops vs. commas, italics vs. no
italics etc. And still you don't get it right...

I think we need a standard for these things that has the following
properties:
* It should be easy to learn and use.
* It should allow import and export of records from and to bibliographical
databases using any standard database program.
* It should be accepted by everyone, in such a way that you can
always recycle a reference that you have used in an earlier publication.

It is possible that someone has thought of such a standard already. In
that case, it should be made known to all of us and we should try to
get it accepted as soon as possible. If no such standard exists, we
should decide upon one. I have some ideas what it should look like,
but first I would like to hear what other people think about this. The
time we invest in this project will be well spent; we'll save many
hours in the future that can be used for more productive purposes. If
you write to me what you think, I will be back with a summary and a
proposal.

Oesten Dahl
oesten@ling.su.se


===================END FORWARDED MESSAGE===================

-----------------------------------------------------------
Dr. Martin Haase, Universitaet FB 7, D-49069 Osnabrueck
Tel.:+49-541-969-4340, FAX:+49-541-969-4256
http://www.cl-ki.Uni-Osnabrueck.de/~haase/

9th September 1996: Re: Bibliographical references

> Date:  Fri, 06 Sep 1996 16:24:19 BST
> >From:  oesten@ling.su.se (Oesten Dahl)
> Subject:  Standards for bibliographical references
> 
> Is there anyone except me who feels the need for a unified way of writing
> bibliographical references? You may not believe me, but the style sheets of
> two major publishers of linguistics literature, Benjamins and Mouton, differ
> on EIGHT points in how a totally ordinary journal article should be
> referenced. (After having looked into a few more publications from these
> publishers, I have to modify this statement. Actually, each publisher seems
> to have several style sheets which differ between each other no less than
> the ones I mentioned.) Every time you submit a paper, you have to spend
> several hours studying intriguing details such as single vs. double
> quotes,  [etc...]

Yes!  This variation has annoyed me.  The Linguistic Society of America
has a standard style which they've been promoting for a long time, but
no one else uses it.  The scientific styles of the U. of Chicago Press
and Cambridge U. Press are very similar to it, and we might be able to
make a compromise out of those.

I think it is important to cite by name and date, not by bracketed
numbers.  Chomsky 1957 is Chomsky 1957, and is recognizable as such
in any paper, but if he is [4] in one paper and [23] and another, you'll
hardly recognize him.

-- 
Michael A. Covington                http://www.ai.uga.edu/faculty/covington/
Artificial Intelligence Center                                           <><
The University of Georgia                Unless specifically indicated, I am
Athens, GA 30602-7415 U.S.A.                not speaking for the University.

9th September 1996: Re: Bibliographical references

Covington writes:
> > Date:  Fri, 06 Sep 1996 16:24:19 BST
> > >From:  oesten@ling.su.se (Oesten Dahl)
> > Subject:  Standards for bibliographical references
> > 
> > Is there anyone except me who feels the need for a unified way of
> > writing
> > bibliographical references? You may not believe me, but the style
> > sheets of
> > two major publishers of linguistics literature, Benjamins and Mouton,
> > differ
> > on EIGHT points in how a totally ordinary journal article should be
> > referenced. (After having looked into a few more publications from
> > these
> > publishers, I have to modify this statement. Actually, each publisher
> > seems
> > to have several style sheets which differ between each other no less
> > than
> > the ones I mentioned.) Every time you submit a paper, you have to
> > spend
> > several hours studying intriguing details such as single vs. double
> > quotes,  [etc...]
> 
> Yes!  This variation has annoyed me.  The Linguistic Society of America
> has a standard style which they've been promoting for a long time, but
> no one else uses it.  The scientific styles of the U. of Chicago Press
> and Cambridge U. Press are very similar to it, and we might be able to
> make a compromise out of those.
> 
> I think it is important to cite by name and date, not by bracketed
> numbers.  Chomsky 1957 is Chomsky 1957, and is recognizable as such
> in any paper, but if he is [4] in one paper and [23] and another, you'll
> hardly recognize him.

Agreed also with two caveats. Since Chomsky probably wrote more than
two things in 1957, Chomsky 1957 in two different articles could refer
to different Chomsky works.  The author date system also has problems
when more than one person in the field has the same last name or even
worst same last and first names.

In answer to Dr. Haase's question:

>By the way, is there a bibstyle usable for
>linguists? As you know, plain-bibstyle-like reference by numbers is
>not used in linguistics (and IMHO not very practical).

There are several author date bibstyle files for bibtex.  All have to
be used in combo with a regular latex sty file

newapa.bst and apalike.bst (the granddaddies of the following)
chicago.bst (used in combo with chicago.sty)
mla.bst (variant of chicago, used with mla.sty)
cslibib.bst (used with cslibib.sty, my homegrown variety not recently
	overhauled)

There are others but starting with these you can probably tailor
something to fit your wishes.

Most use at least two variants of \cite.  The most common are

\cite -> for citations of the form (Chomsky 1957)
\shortcite -> for citations of the form (1957) {for when the author's
	name is in the actual text}

chicago at least has a whole slew of other cite variants.

I haven't looked closely recently but I suspect there are many others
out there.

Emma Pease
who will leave it to the linguists to get the journals to agree on a
standard format

9th September 1996: Re: Bibliographical references

This discussion about a uniform style of citation has an obvious
appeal, but seems to me absolutely futile. For publishers of
linguistic work, even those with some serious commitment (like
Cambridge University Press), ours is a tiny and marginal field. They
come up with their style sheets in the context of their overall
publishing program, and their editors tend to be quite obsessive about
maintaining a uniform "look and feel" across a publisher's offerings.
Trying to impose "our" standard on publishers would only be likely to
increase their reluctance to do linguistics. A similar point was made
many years ago by Bloomfield, who encouraged linguists to avoid
diacritics and special symbols in phonetic transcription so as to make
publication more practical, given the realities of the business.

So no matter how sensible and widely accepted a standard for
Linguistics might be, along the lines Oesten suggests, I see virtually
no chance of getting editors and publishers to accept it. Not to
mention the difficulty of overcoming sheer inertia. Or the problem of
getting any group of half a dozen or more linguists to agree on much,
especially in matters of sytle (about which we each have opinions
galor, all very strongly held and all no doubt totally correct).

So my (defeatist) point is: there's not much sense in spending time
and energy coming up with a standard of this sort. It IS a pain to
have to attend to all of these things that differ from one journal or
other publication venue to another, but it's one we'll have to live
with for the forseeable future. 

Less defeatist: If you want to spend your time on something that will
make linguists' lives easier in this regard, working on BibTeX .bst
files that conform to existing standards is probably much more useful.
That way, of course, authors can maintain a single .bib file and let
BibTeX worry about how to format stuff.

I have a couple of these, one for Language, one for The Linguistic
Review, etc., based on other people's work and intended for use with
the harvard bibliography style. None work quite well enough for me
to want to publish them, but if there's someone else out there who's
done some BibTeX hacking before, they might be a useful starting
point.

--Steve Anderson

9th September 1996: Re: Bibliographical references

Stephen Anderson has a point, and behind it is a bigger point: Linguistics
does not need a citation format all its own.  The needs of linguists are the
same as those of other scientists.  I think we should throw our support
behind one of the major scientific styles, perhaps that of the American
Psychological Association or that of Chicago and Cambridge (all three of these
being nearly the same).
-- 
Michael A. Covington                http://www.ai.uga.edu/faculty/covington/
Artificial Intelligence Center                                           <><
The University of Georgia                Unless specifically indicated, I am
Athens, GA 30602-7415 U.S.A.                not speaking for the University.

10th September 1996: Re: Bibliographical references

I agree with the major lines of Michael and Steve's argumentation:
0. The present situation is undesirable (for costs and loss of
energy)
1. There is no point in a special solution for linguistics (although
I wouldn't choose psychology as a frame that embraces linguistics,
pace Chomsky and Covington). Except for numbered references, I'm
quite happy with plain bibtex style, and Chacago and Cambridge would
do as well.
2. The bibtex philosophy is much better than one Procrustean
solution, esp. for people like myself who publish in different
languages and across traditional academic fields (and who have
political/moral objections against enforced uniformism); that is to
solve the problem through the automatic adaption to different style
sheets with the help of bibstyle macros. 
Best wishes and thank you for the helpful information (also to Emma
for pointing out bibstyle sources),
Martin
-----------------------------------------------------------
Dr. Martin Haase, Universitaet FB 7, D-49069 Osnabrueck
Tel.:+49-541-969-4340, FAX:+49-541-969-4256
http://www.cl-ki.Uni-Osnabrueck.de/~haase/

10th September 1996: Biblio ref styles

I have to agree with Steve Anderson (great e-mail address, Steve!)
that it would be more useful to attack this vexatious problem via the
source file -- BibTeX and .bst style files -- rather than attempt to
impose a single monolithic biblio style on all linguistics journals
.. although I'd be more inclined to say it's a pragmatic rather than
defeatist approach to consider ;-)

And while I commend Oesten Dahl's implied diligence when he says that
it's a pain to have to study each journal's style in order to submit
.. as typesetter for a couple of journals and books in
non-math/science fields, authors who submit articles with their
biblios actually done in the style of the journal they're submitting
to ... well, it's probably something like 5%! 

Well, ok ... probably better than that, but still, it's pretty low --
it seems biblios are not where scholarship shines, I regret to report
.. ;-)

However -- unless you're an incredible nit-picker for details that
seem completely by the by, as far as the actual contents of the
article goes, biblios really are a pain in the a** to do, to do
properly and to do completely. Editors know that; authors know that.

Enter BibTeX. I mean, that's what this thing was supposed to do for us
-- both as authors and as typesetters of material from authors -- make
biblio work a lot easier and more consistent. It allows you to set up
your own little database of frequently-cited sources, and then from
there, it allows you -- and more particularly, publishers -- to then
shuffle all that tagged stuff around, according to the instructions in
the .bst style file -- at least, that's how I understand it. Please --
active BibTeXers, comment and correct before errors go any further!

\end{harangue} 

[Yeah, I know ... but I'm not sure where the {harangue} actually
 started ... ;-)]

===

So -- it seems to me that an eminently achievable task is for us to
look at BibTeX, consider whether it's indeed a realistic route for us
to take, and then go from there.

There are people on this list who are very familiar with .bst files;
there are people who are very familiar with various journal styles.
I'm sure that if we decide that it's worth spending time getting some
of the .bst files Emma so kindly listed for us out into more general
circulatiion, along with the additional draft versions which others
have worked out for their own purposes, that we could very quickly see
authors starting to use .bst files, and finding that it in fact
becomes a pleasure to be able to re-do an article in this or that
journal's biblio style, just by changing the .bst file. Imagine how
the editors would feel too -- getting an article with the biblio
actually done up in their journal's style ... wow ... what a treat!


Damn! This forces me to ask the question: so, if the hardcopy comes
out right, how do I get the stuff in ASCII form to the editor in the
same order and shape ... ;-) ... well ... let's not throw spanners
into the works before the wheel even begins to move. As I said, I know
the theory, but don't actually have any BibTeX experience ... yet ...

I guess the first thing then is to ask: do readers of this list feel
that it's worth discussing this BibTeX route? 

Ch.

P.S. There is actually a TWG (Technical Working Group) for linguistics
     and TeX, and I'm supposed to be chair. If there's a consensus on
     this subject, perhaps the TWG can help manage/coordinate things
     and maybe we'll end up with some nice solutions to this biblio
     problem (which plagues all subject areas, not just ours!).

P.P.S. Martin, if Oesten isn't on ling-tex, please forward this
       message and any others you think he'd be interested in ...
       Or better yet -- get him onto the list!

11th September 1996: Keeping track of .bst files

I like Christina Thiele's suggestion to systematize .bst files
as a way of encouraging authors and editors in the direction 
of TeX.  And let's not forget the honorable exceptions: journals,
publishers, and conf. proceedings that now accept LaTeX input.  

Journal of Logic, Language and Information (Kluwer)
CSLI (Chicago)
Computational Linguistics (MIT Press)
ACL Proceedings

(There must be others: the list above comes out of my own experience.)
A useful task for someone on this list would be to keep track of 
what exists and offer a web site that linguistics organizations
could point too.  The same list of references done in 5 or so 
different ways, a reference to a LaTeX book, and downloadable
.bst files ought to do it.  I'd offer but I'm just too swamped.

--John Nerbonne

Christina Thiele writes:
> There are people on this list who are very familiar with .bst files;
> there are people who are very familiar with various journal styles.
> I'm sure that if we decide that it's worth spending time getting some
> of the .bst files Emma so kindly listed for us out into more general
> circulatiion, along with the additional draft versions which others
> have worked out for their own purposes, that we could very quickly see
> authors starting to use .bst files, and finding that it in fact
> becomes a pleasure to be able to re-do an article in this or that
> journal's biblio style, just by changing the .bst file. Imagine how
> the editors would feel too -- getting an article with the biblio
> actually done up in their journal's style ... wow ... what a treat!

11th September 1996: Keeping track of .bst files

John Nerbonne writes:
> 
> ...
> A useful task for someone on this list would be to keep track of 
> what exists and offer a web site that linguistics organizations
> could point too.  The same list of references done in 5 or so 
> different ways, a reference to a LaTeX book, and downloadable
> .bst files ought to do it.  I'd offer but I'm just too swamped.
> 
> --John Nerbonne
> 
> ...

Just as a matter of interest ... the Canadian Linguistic Association
has some people who are getting a CLA web page together (I'm supposed
to provide some stuff ... haven't done it yet ...  `swamped' is a good
enough description). So that's at least one ling org that's a
potential site for links. 

Anyways ... I would certainly suggest to them that anything related to
linguistics and TeX might be a nice link ... such as this very list,
for example ... ;-) 

Anyone know of other linguistics organisations/societies that perhaps
already have web pages that it might be nice to list here? 

Oh ... about a month ago, I'd posted a message saying that an updated
version of the ling-mac.tex file (the inventory of TeX style files and
such) would be in the works. New items and updates to current listings
should be posted here by the end of September ... mid-October ... 
How about we also include a section on web sites for linguistics
organisations? 

Well ... gotta get back to typesetting some of that linguistics stuff
again ... I think I hear the client starting to make noises ... ;-)

I'll keep tabs on comments and suggestions on the issue of .bst style
files, as well as items related to the update of the ling-mac.tex
file. 

Ch.

13th September 1996: Re: Biblio ref styles

Oops!  I blew it!  Sorry for resending Christina's entire message to the
list.  I'll try not to do it again.

Christina wrote:  
>Damn! This forces me to ask the question: so, if the hardcopy comes
>out right, how do I get the stuff in ASCII form to the editor in the
>same order and shape ...

I've had success doing this using dvi2tty.  The result is (or can be)
an ASCII file with all the basic formatting of the good-looking
version, but without any italics, bolding, etc.  The order of elements
in the bibliography is fine, though.

	Paul
========
Paul Neubauer                                00prneubauer@bsuvc.bsu.edu

13th September 1996: Re: Biblio ref styles

00prneubauer@bsuvc.bsu.edu writes:
> 
> Oops!  I blew it!  Sorry for resending Christina's entire message to the
> list.  I'll try not to do it again.
> 
> Christina wrote:  
> >Damn! This forces me to ask the question: so, if the hardcopy comes
> >out right, how do I get the stuff in ASCII form to the editor in the
> >same order and shape ...
> 
> I've had success doing this using dvi2tty.  The result is (or can be)
> an ASCII file with all the basic formatting of the good-looking
> version, but without any italics, bolding, etc.  The order of elements
> in the bibliography is fine, though.
> 
> 	Paul
> ========
> Paul Neubauer                                00prneubauer@bsuvc.bsu.edu
> 


Neat! Thanks for the tip. If anyone else wants to try this out, you
can probably get dvi2tty from your nearest CTAN site -- and post your
test results. Aha -- just found it on CTAN -- look in:

   tex-archive/dviware/

As for bibtex itself, there's a whole subdirectory of stuff on CTAN.
Look at:

   tex-archive/biblio/bibtex/...

15th September 1996: Re: Chinese and TeX

I have just come back online after being away a couple of weeks in
Bangladesh and saw this note...

>Is it possible to write chinese characters using tex? How does it work?


The Taiwan company that produces ETEN a Chinese add-on for DOS has a version
of TeX that works much like the original and can handle English and Chinese
together in the same document (We have used it to typeset several large
publications and have been pleased with it, especially in the way it typeset
the recent Indonesian-Chinese Diglot of the Bible).

The unusual features of this system is that the operating system ETEN uses
Big 5 encoding for tradition scripts and Gaubao for simplified script.
Both of these encoding are a mixture of both 8 and 16 bit encoding. However
the ETEN version of TeX can handle both system this as well as user-defined
characters. It also knows about intercharacter spacing between Chinese
letters and the need to prevent line breaks before or after certain
punctuation. We did have to play a few tricks to get the system to typeset
the wavy-lined ruled Chinese phrases without disturbing the intercharacter
skip but we got it work well.

I am told that MicroPress the authors of VTeX are nearing completion of
their Chinese version that will support Unicode and True Type fonts. Michael
Valis can supply you more information on this if you are interested.
===============================================================
 BOB BATZINGER                         UBS CAP OFFICE
 EMail: bob@bible.org.sg               7 Armenian St, #04-09
 Fax: (65) - 337-3036                  Singapore 179932

20th September 96: Caveat BibTeX!

From: Thomas Martin Widmann 
Subject: Caveat BibTeX!

I have been following the discussion of BibTeX with great interest
(unfortunately I have not been able to join the discussion before now,
but I am studying Georgian (and some linguistics) in Tbilisi, Georgia,
and the computer revolution has yet to cross the Caucasus!)...  It seems
to me that BibTeX is a bit like candyfloss:  It looks great and it is OK
in the beginning, but after a short time you get problems and at some
point you get stuck. :-)

Honestly, the problem with BibTeX is primarily that it is too
linguacentric, it assumes that you write in English and that the books
in your bibliography are written in English, too (yes, I know that there
exist adaptions to some other languages, but if your language is not
there, you'll have to do it yourself which may be very difficult if your
language happens to  differ a lot from English).  Even if you write only
in English but want to share your .bib files with others, you may get
into trouble.  Let's take the "Month" field as an example. The names of
the months are defined in the .bst file, so that the publisher can
decide whether he wants "Oct.", "October" or (e.g., in a Danish .bst
file) "oktober".  However, it is common to find month fields that say
"10 " # nov or nov # " 10" (# is BibTeX for string concatenation), but
both of these refer--of course--to the same date and should be formatted
the same way (and as "10.~november" or "10.~nov." in Danish).

The "Month" field is not a big problem, it could easily be solved by
an addition to BibTeX syntax, but it is there today and there is no
easy work-around--you have to work through all the .bib entries (which
is a BAD THING, because one of the supposed avantages of BibTeX is that
it is claimed to let people share their bibliographies!)...

Then, of course, there is the problem of sorting.  Maybe it is possible
to specify a different alphabetic ordering in the .bst file, but I don't
know about it then.  Furthermore, BibTeX can only handle 7-bit ASCII
input which means that you have to type a name like "¥stergård" as
"{\O}sterg{\aa}rd"--not very confortable if such characters are common. 
I have seen an extension to BibTeX that knows about latin1 and even
handles the problem of different languages sorting differently (e.g.,
six authors called Åen, Andersen, Llama, Lund, Östervik and Uhl should
be sorted as Andersen, Llama, Lund, Uhl, Östervik, Åen in Danish, as
Andersen, Llama, Lund, Uhl, Åen, Östervik in Swedish, and as Åen,
Andersen, Lund, Llama, Östervik, Uhl in Spanish), but if you happen to
cite a lot of Russian authors (and don't want to add their writings in
transliteration), BibTeX may not be for you.

In fact, it should not be you but rather the publisher who should
decide whether Russian literature should be transliterated (and
what transliteration scheme to use)...!  Let's take an example
(adapted from a posting I made three months ago on the Omega
mailing list--Omega is a 16 bit, Unicode-aware TeX successor which
is going to be great news for linguists):

Let's say I write a paper where I cite four (non-existent) books:

1) (in Danish):   Hjelmslev: "Det danske Sprog."
2) (in Russian):  Klimov: "Datskiy Yazyk."
3) (in Georgian): Gamqrelije: "Daniuri Ena."
4) (in Japanese): Miraa: "Demmaakugo."

Now, Hjelmslev is a Danish name written in Danish, Klimov is a Russian
name written in Russian, Gamqrelije is a Georgian name written in
Georgian (but he actually prefers to latinize his name as Gamkrelidze),
and Miraa is an American name (Miller) written in Japanese.

There are several approaches:
1)  We could write all the books in transliteration, using preferred
    forms of the names, i.e.:
	Gamkrelidze:...
	Hjelmslev:...
	Klimov:...
	Miller:...
2)  We could write all the books in the original form, sorting the alphabets
    as (e.g.) Latin, Cyrillic, Mxedruli, Japanese (like the way I presented
    the books above).
3)  A normal approach could be to assume that the reader knows Latin and
    Cyrillic, but no other alphabets:
	Gamkrelidze:...
	Hjelmslev:...
	Miller:...
	[in Russian] Klimov:...

And so on.  This is very complicated.  If all of this should be generated
from the same .bib file, the following information should be contained
in it:

* The language and encoding the entry is written in.
* Preferred transliteration (optional).
* Enough information for automatic transliteration.

The last point is drastic.  For some languages (e.g., Georgian), it is
very easy.  For some, e.g., Russian, it is just a small addition (for
example, Gorbachev is written as Gorbatjov in Danish.  In order to
transliterate 'jo' and not 'je', one has to know that the Cyrillic
letter is ë and not e, although this difference is not often made in
writing.). For others, e.g., Arabic, it is drastic:  Normally, Arabic
texts are not vocalized, and so the .obib file should be vocalized by
hand (with a note to Omega that these vowel signs should not be used
when writing Arabic).  The same is the problem with Japanese:  Every
Kanji (Chinese character) has to be given its pronounciation, or
transliteration will be impossible.

Maybe BibTeX is for you, but I tend to run into some of the problems
every time I decide to use it...  On the other hand, I think that if
a really useful BibTeX++ should be written, the world's linguists
have to help.

Yours,
Thomas

BNF Thomas Martin Widmann |http://ling.hum.aau.dk/~viralbus| T. +995/32-224918
Barnovi str. 185, apt. 3  |    viralbus@access.sanet.ge    | Lernu Esperanton!
380062 Tbilisi    Georgia |  CAESAR*NON*SVPRA*GRAMMATICOS  | Stud (Ling. & CS)

20th September 96: Re: Caveat BibTeX!

From: cthiele@ccs.carleton.ca (Christina Thiele)
Subject: Re: Caveat BibTeX!

Thomas Martin Widmann writes:
> 
> I have been following the discussion of BibTeX with great interest
> (unfortunately I have not been able to join the discussion before now,
> but I am studying Georgian (and some linguistics) in Tbilisi, Georgia,
> and the computer revolution has yet to cross the Caucasus!)...  It seems
> to me that BibTeX is a bit like candyfloss:  It looks great and it is OK
> in the beginning, but after a short time you get problems and at some
> point you get stuck. :-)
> 
> ...
>

Many thanks, Thomas, for all the comments and cautions! I'm sure that
it will be very helpful for people to know that BibTeX (like most
things, I guess) isn't the complete and universal answer to all our
bibliographic problems.

It would be good to hear from others who have BibTeX experience, to
find out if and how they've found similar road-blocks, or perhaps even
found creative solutions to these and other sticking points.

Nevertheless, I would hate to see this discussion come to a screeching
stop just because there are potential problems ahead. The warnings are
necessary to the discussion, perhaps even moreso than the rosy theory
;-) But let's not now feel the situation is totally bleak and
hopeless!

So I'd invite more posts from the ling-tex community on the subject.

Ch.

20th September 96: Re: Caveat BibTeX!

From: "Harold F. Schiffman" 
Subject: Re: Caveat BibTeX!

On Fri, 20 Sep 1996, Christina Thiele wrote:

> Thomas Martin Widmann writes:
> > 
> > I have been following the discussion of BibTeX with great interest
> > (unfortunately I have not been able to join the discussion before now,
> > but I am studying Georgian (and some linguistics) in Tbilisi, Georgia,
> > and the computer revolution has yet to cross the Caucasus!)...  It seems
> > to me that BibTeX is a bit like candyfloss:  It looks great and it is OK
> > in the beginning, but after a short time you get problems and at some
> > point you get stuck. :-)
> > 
> > ...
> >
> 
> 
> It would be good to hear from others who have BibTeX experience, to
> find out if and how they've found similar road-blocks, or perhaps even
> found creative solutions to these and other sticking points.
> 
> Nevertheless, I would hate to see this discussion come to a screeching
> stop just because there are potential problems ahead. The warnings are
> necessary to the discussion, perhaps even moreso than the rosy theory
> ;-) But let's not now feel the situation is totally bleak and
> hopeless!
> 
> So I'd invite more posts from the ling-tex community on the subject.
> 
> Ch.


Well, since you ask...

I tried to use BibTeX for a book I was working on that went to press in 
1995 and has since been published, but with no help from BibTeX.  I had 
personally formatted the whole volume in LaTeX, and created the *.bib 
file as I went along, adding to it from various sources.

My greatest problem, of course, was knowing how to deal with anything 
that was not just a vanilla-type entry, i.e. book by so and so, pub. by 
soand so, year, title, author.  There were many entries involving edited 
volumes, dissertations, special issues of journals, and most of these 
just fell through the net.  When I tried to run a final output file, 
these things simply did not appear.  And if one comma was out of place, 
things did not appear.  The worst thing was that I was unaware of this 
until too late, and I in fact ended up without a useful bibliography 
file, and had to send my publisher (Routledge) the raw 
source file as my bibliography (they wanted the book in machine-readable 
form, but not in LaTeX, so they had to strip out LaTeX codes).  

This book was not particularly "linguistic" in format; it's a book on the 
subject of language policy, so I didn't have too many "linguistics" type 
bibliographic entries that would have messed things up; but I must say 
there is many a slip between cup and lip.

I really think BibTeX ought to be in a data-base format, not entered as 
text. As I say, one little comma out of place, and all fails...

Hal Schiffman

=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+
------------------------------------------------------------------------------

Harold F. Schiffman				           Academic Director
Henry R. Luce Professor of Language Learning		Penn Language Center
Dept. of South Asia Regional Studies		   4th Floor, Lauder-Fischer
820 Williams Hall, Box 6305					    Box 6330
					
			University of Pennsylvania
			Philadelphia, PA 19104-6305

Phone:  (215) 898-5825	   		       	              (215) 898-6039
Fax:  (215) 573-2138				          Fax (215) 573-2139

Email:  haroldfs@ccat.sas.upenn			      plc@ccat.sas.upenn.edu
WWW:  http://ccat.sas.upenn.edu/~haroldfs/    http://ccat.sas.upenn.edu/~plc/

The Penn Language Center is a facility supporting language teaching in the
less-commonly-taught languages, as well as research in language pedagogy
and interdisciplinary language-related issues.  
------------------------------------------------------------------------------

23rd September 96: Biblio ref styles

From: Jean-Pierre.Drucbert@onecert.fr (Jean-Pierre Drucbert)
Subject:  Biblio ref styles

 Do not forget that the marvelous custom-bib package exists! It
 could help you for your language problems with BibTeX. This package
 is available on CTAN.
 This package is writen by  "P.W.Daly, MPAe, Lindau, Germany" 
and has been recently updated (17 Sep 1996, I have noted).

25th September 1996: Fwd: Standards for bibliographical references: summary and proposal

From: "Martin Haase" 
Subject: Fwd: Standards for bibliographical references: summary and proposal

Two weeks ago I asked:

>Is there anyone except me who feels the need for a unified way of writing
>bibliographical references?

I received responses from 23 people (I hope nobody has got lost). In
addition, my posting has given rise to a discussion on another list,
ling-tex, the discussion group of linguists that use TeX. Thanks to
Martin Haase for keeping me informed about that development! Thanks
also to everyone who has written to me and apologies to all those
whom
I haven't responded to individually. Here is the list of
correspondents:

Charlie Rowe 
Chungmin Lee 
Deborah D. Kela Ruuskanen 
Dominique Estival 
Donna Lillian 
Ellen Bard 
Glenn Ayres 
Alice Horning 
Ingo Plag 
Laurie Bauer 
James Jenkins  
Karen S. Chung 
Lee Hartman 
Martin Haase 
Adam Meyers meyers@acf2.NYU.EDU (meyers)
Michael Covington 
Ocke-Schwen Bohn 
Jan Odijk 
Peter Daniels 
Christian Bauer 
Tracy Cameron Mansfield 
Gisela Redeker 

The respondents fall into three categories (with some overlaps):

1. Those who remind me of standards that already exist.
2. Those who remind me of the existence of various software which
might solve the problem.
3. Those who simply support the idea.

Accordingly, let's have a look at existing standards and software.

EXISTING BIBLIOGRAPHICAL STANDARDS

The following bibliographical standards are relevant to the
discussion
(others that have been mentioned are in my view too far from
linguistics and linguistic practice to merit serious consideration):

1. The MLA (Modern Language Association) standard.
2. The APA (American Psychological Association) standard.
3. The "Chicago Manual of Style" standard, which is really two
different standards, which I shall call "Chicago I" and "Chicago II".
4. The Language Style Sheet, previously called the LSA Style Sheet.

For reasons of space and typography, I shall not try to describe in
detail what these different standards are like. Here are two web
sites
where descriptions and examples can be found of (1-3):

http://www.lib.montana.edu/STYLES/index.html
http://www.lib.lehigh.edu/footnote/descrip.html

I think it is possible to exclude at once the MLA standard and
Chicago
II, which are fairly similar to each other but which deviate too much
from what is common practice in our field (the main stumbling-stone
being the placement of the year of publication towards the end of the
record).

More serious candidates are the APA and Language style sheets. The
APA
style sheet, as many correspondents have pointed out, is more or less
universally accepted in psychology, and also in quite a number of
publications on the borderline of psychology and linguistics. The
Language style sheet (regrettably, I should say) never seems to have
enjoyed a similar popularity. It seems that the LSA have kept a low
profile on this point, mainly enforcing it in their own
publications. (It is fairly similar to Chicago I.)

EXISTING SOFTWARE

Some of the software which has been mentioned by correspondents is -

bibtex:

A special program (bibtex) transforms a uniform ASCII bibliographic
database into different bibliographic styles (which are defined in so
called bibstyle files), so one does not have to worry at all about
the
final form of the bibliography, as long as there is a bibstyle. The
citation within the text is also organized through style files (style
sheets), the document styles (which contain other formatting
information, so that the same text can easily adapted to the needs of
a specific style sheet).  (Martin Haase)

EndNote Plus
EndNote Plus 2 is a bibliographic database program that manages
bibliographic references and creates bibliographies automatically in
your word processor. More than 100,000 researchers, librarians, and
students use EndNote Plus to maintain a personal library of
references. The program assists academic writers and researchers in
keeping track of bibliographic references and generating
bibliographies for the books and papers that they write. EndNote Plus
is available for Macintosh, Windows and DOS. EndNote Plus comes with
more than 300 pre-defined bibliographic styles for the leading
journals in a wide variety of disciplines, and you can easily create
an unlimited number of your own styles. With Endlink, EndNote Plus's
import module, you can save yourself from typing bibliographic data
by
hand. A variety of bibliographic databases available through online
services and CD-ROMs can give you fast access to thousands of
references. Using EndLink, you can import these references into your
EndNote Plus library without typing a word (mentioned by Christian
Bauer, info from http://www.niles.com) (Comment: the bibliographic
styles do not include the Language style sheet or any other
specifically linguistic ones)

ProCite

Use ProCite to organize references and format bibliographies in any
journal style. Easily maintain databases of research collections and
share data with colleagues.  ProCite is available for Windows, DOS,
and the Macintosh System.  (mentioned by Peter Daniels, info from
http://www.pbsinc.com)

Blackwell Idealist
(mentioned by Annabel Cormack)
Seems to be a general database program
(http://www.blacksci.co.uk/products/idealist/aboutide.htm).

Now, what all these programs seem to be able to do is to store
bibliographical information in a database and then generate formatted
lists of references to be included in publications. This is all very
well - actually, I think any database program with any self-respect
can do that.  The problem is that it does not really solve the
problem, for the following reasons:

1. So far, style files are typically provided only for the major
standards.  Maybe we can convince the publishers who have their own
idiosyncratic style sheets to supply style files, but then we have a
new problem: the proliferation of software... The number of style
files will have to equal the number of style sheets multiplied by the
number of bibliography programs! So even assuming this solution it
wouldn't hurt to introduce unified standards.

2. If the software is not seamlessly integrated with the
word-processing programs people use, it will be very hard to convince
them that it's a good idea. (This is, I guess, the main objection
against LaTeX.)

2. Existing bibliography software seems to be quite good at
GENERATING
lists of references. But there are other tasks that are no less
time-consuming, which those programs don't seem to be able to perform
(correct me if I am wrong!):

a. IMPORTING lists of references, not only ones specially formatted
for the purpose (comma-separated or whatever) but all your old lists
of references in various formats and various degrees of perfection

b. PROOF-READING bibliographical references in text format

c. CONVERTING references from one text format to another (maybe the
same as a)

Imagine for instance what happens when a volume of conference papers
is produced. Twenty or thirty authors with varying backgrounds and
varying ability and readiness to learn new standards are asked to
comply with some publisher's idiosyncratic style sheet. Each paper
contains between ten and fifty references. The editor of the volume
(whether s/he is the person who organized the conference or someone
at
the publisher) has to check maybe a thousand entries. Anyone who has
tried knows what I'm talking about... In this situation, no existing
database will help you.

THE OPTIONS

If you do not believe me, half an hour in your university library
will
convince you that the situation with respect to writing
bibliographies
in linguistics cannot be characterized by any other word than
chaos. There are about ten different ways of writing such a simple
thing as the year of publication, which I won't bore you by
enumerating. Just to give you an idea of what I am talking about,
consider the following. The Journal of Pragmatics writes "Chomsky,
Noam, 1957." when the Journal of Semantics has "Chomsky,
N. (1957),". The prize for idiosyncrasy goes however to the Kluwer
journals (Linguistics and Philosophy, Natural Language and Linguistic
Theory), which write "Chomsky, N.: 1957, ". I can see no rational
reason whatsoever why Kluwer Academic Publishers Group should be
allowed to get away with this. It does not help that they provide a
LaTeX style file on their website (http://www.wkap.nl/).

In my view, there are three options:

1. Join the psychologists, adopting the APA standard.
2. Adopt the LSA/Language Style Sheet as the standard for linguistics
in general.
3. Create a new standard.

My personal leanings are towards the third alternative, although I
realize it is a bit risky, in that it may just become a new
"sectarian" style sheet used by a few people. The reason is that none
of two major existing standards is quite adequate for the needs of
the
Internet age. (The APA style cannot, for instance, be rendered
properly on the Linguist List due to the use of italics.) Below, I
will describe what I call the "No Frills Style Sheet", which is my
own
proposal. It does have some of the look and feel of the LSA Style
Sheet, but there are a number of points where it differs.


THE NO FRILLS STYLE SHEET

The idea behind this style sheet is that it should be maximally easy
to learn and to use, both for humans and computers. The consequences
are a set of simple principles:

1. A list of references should have a clear field-and-record
structure, with consistent field and record delimiters.
2. Punctuation and variation in typefaces should be kept to a
minimum.
3. Context-sensitive rules should be avoided - an entity, for
instance
a name, should be written the same way wherever it appears, as far as
possible.
4. When in doubt, choose the alternative that is closest to common
practice in linguistics (to the extent that it exists).


Here are some examples what the application of these principles leads
to:

Book:
Chomsky, Noam. 1957. Syntactic structures. The Hague: Mouton.

Article or chapter from edited book:
Traugott, Elizabeth. 1978. On the expression of spatio-temporal
relations in language. In Greenberg, Joseph & Ferguson, Charles &
Moravcsik, Edith, eds., Universals of Human Language, vol. 3, Word
Structure, 369-400. Stanford: Stanford University Press.

Article from journal:
Hopper, Paul & Thompson, Sandra. 1980. Transitivity in grammar and
discourse. Language 56: 251-99.

The major difference from the Language Style Sheet is in the order
some items are written. Names are always written as "Lastname,
Firstname" - this makes sorting and searching much easier and also
removes a source for typing errors. Similarly, the editors of books
are written before rather than after the title.  There are further
details to be decided but this is the general idea of what it should
look like that I would like to present at this point.

WHAT NOW?

I would now like to hear what people think about the alternatives. I
would like to emphasize that whatever option we choose, it will be a
big step forward in relieving us from a lot of totally unnecessary
trouble, if we succeed in getting it generally adopted.

Please write to me and tell me what you think. If this turns out be a
serious business, we should probably get a discussion list of our
own. But let us first see if there is enough interest in the whole
thing.

Oesten Dahl


===================END FORWARDED MESSAGE===================

-----------------------------------------------------------
Dr. Martin Haase, Universitaet FB 7, D-49069 Osnabrueck
Tel.:+49-541-969-4340, FAX:+49-541-969-4256
http://www.cl-ki.Uni-Osnabrueck.de/~haase/

30th September 1996: Re: Caveat BibTeX!

From: Thomas Martin Widmann 
Subject: Re: Caveat BibTeX!

Christina Thiele (cthiele@ccs.carleton.ca) wrote:

> Nevertheless, I would hate to see this discussion come to a screeching
> stop just because there are potential problems ahead. The warnings are
> necessary to the discussion, perhaps even moreso than the rosy theory
> ;-) But let's not now feel the situation is totally bleak and
> hopeless!

Oh, it's not hopeless, I never said so.  I only feel that BibTeX is not
good enough.  It is true, as Jean-Pierre Drucbert noted, that the
custom-bib package exists, but last time I checked, it only solved a few
problems.

As I see it, the major problems with the current BibTeX are:

*  The .bst files are unreadable.  Of course, if you only want to change
   a minor point (e.g., whether the author should be written in small
   caps) it *is* possible to do that, but IMHO it is much too difficult 
   to read and write .bst files (my minor subject is computer science
   and I've learned at least ten programming languages, so if I have
   problems, I'm probably not the only one).
*  It is very monolingual by design (check my previous posting about
   this point).
*  It is not very flexible if non-standard entries are needed (check the
   fine posting by Harold F. Schiffman about this one).

I also see other minor problems, but the above should be reason enough
to redesign BibTeX.  Of course, this list is probably not the right
place to ask for programmers, but what we need now is design.  If we get
the over-all design right in the first place, it should not be hard to
implement.  I very much want to participate in this work myself
(although it is a little impractical that I am in Tbilisi until july),
and--as I wrote in my previous posting--I feel that linguists should be
active in the design phase.

Just now the best thing would maybe be to find out what the problems are
going to be.  I would very much like to see examples of difficult
bibliography entries (e.g., names that should be sorted in a non-obvious
way, titles that should be transliterated or translated, strange types
of books that should not exist in theory, etc. etc.)!

Just my 50 tetri.

Thomas

30th September 1996: Re: Caveat BibTeX!

From: Emma Pease 
Subject: Re: Caveat BibTeX! 

Thomas writes:
> Oh, it's not hopeless, I never said so.  I only feel that BibTeX is not
> good enough.  It is true, as Jean-Pierre Drucbert noted, that the
> custom-bib package exists, but last time I checked, it only solved a few
> problems.

I note the latest version for custom-bib was released June 1996 (3.81)
with minor modifications added up to Sept. 18 so you might want to
double check.  I confess not to use it myself but to be in the habit
of modifying .bst files directly when necessary.

> As I see it, the major problems with the current BibTeX are:
> 
> *  The .bst files are unreadable.  Of course, if you only want to change
>    a minor point (e.g., whether the author should be written in small
>    caps) it *is* possible to do that, but IMHO it is much too difficult 
>    to read and write .bst files (my minor subject is computer science
>    and I've learned at least ten programming languages, so if I have
>    problems, I'm probably not the only one).

My guess is that for most people (non-programmers) something like
custom-bib should be used if modifications are needed.  

> *  It is very monolingual by design (check my previous posting about
>    this point).

A problem.  I think the basic changes should be possible (month names,
etc).  Perhaps working on a tool for bib file manipulation would be
helpful.  I tend to use bibtex mode in emacs [with some local changes
such as using abbrevs]; however, it isn't the greatest.

Sorting for 8-bib I suspect may require some internal bibtex recoding.
I note there is an 8-bit bibtex out there by Neil Kempson

 ftp.tex.ac.uk:/tex-archive/biblio/bibtex/8-bit

which in its documentation states

    BibTeX now accepts 8-bit characters in its input files and writes 8-bit
    characters to its output files.  The character set is defined by an
    external configuration text file - the codepage and sort order ("CS")
    file.

    The sort order can be defined for the language and character set.  For
    example, in German, the control sequence \"o (o umlaut) should be sorted
    as if it were the letter "o", but after ordinary "o", leading to this
    order:

        Trofer, Tr\"ofer, Trufer

    However, in Swedish, \"o (o umlaut) is treated as the 29th letter of the
    alphabet and these entries would be sorted as:

        Trofer, Trufer, Tr\"ofer

    The sorting order is defined by an external configuration text file  -
    the codepage and sort order ("CS") file.

    This version of BibTeX, coupled with LaTeX2e and its "inputenc" package
    provide a robust means of handling 8-bit character sets.

I haven't played around with it.  


> *  It is not very flexible if non-standard entries are needed (check the
>    fine posting by Harold F. Schiffman about this one).

True. I have been known to recommend hand editing the bbl file when
nothing else works.  However, you can add new types fairly easily to
the bst file.

> I also see other minor problems, but the above should be reason enough
> to redesign BibTeX.  Of course, this list is probably not the right
> place to ask for programmers, but what we need now is design.  If we get
> the over-all design right in the first place, it should not be hard to
> implement.  I very much want to participate in this work myself
> (although it is a little impractical that I am in Tbilisi until july),
> and--as I wrote in my previous posting--I feel that linguists should be
> active in the design phase.

I did a hunt through the comp.text.tex archives (dejanews is an useful
little beast).  Oren P. the original writer of bibtex is suppose to be
working on a complete overhaul of bibtex; however, I think progress
has been very slow maybe even non-existent (he still has an account
op@cs.stanford.edu if we want to ask him).  Neil Kempson is another
person possibly to ask .

Emma

30th September 1996: A new subscriber's question

From: Jouko Lindstedt 
Subject: A new subscriber's question

I used TeX quite a lot in the late eighties but after that I have been 
using other systems -- which have never quite satisfied me after I learnt 
what typography really is. Now I would like to brush up my TeXnique; 
could someone kindly recommend WWW sites or publications? Is there a FAQ 
for this mailing list?

Thank you in advance!

Jouko Lindstedt
Slavonic and Baltic Department, University of Helsinki
e-mail: Jouko.Lindstedt@Helsinki.Fi or jslindst@cc.helsinki.fi
http://www.helsinki.fi/~jslindst/

1st October 1996: BibTeX

From: tplagwit@ix.urz.uni-heidelberg.de (Thomas Plagwitz)
Subject: BibTex

hi, 

here is my "1 pfennig" (i am *not* computer scientist who has learnt 10
languages and still can't get comfortable with BibTeX, i am just a modest
"Germanist", and thus completely lost  in my *.bst-files :-), as you can
imagine).

I would like to point out, that there is a LaTeX macro package on top of 
BibTex, to be found on:

CTAN: /tex/macros/latex/contrib/supported/camel

which seems to be little known, but is supposed to become the citation
engine for LaTeX 3. So far coded for law-quotations (the author is looking
for volunteers supplying modules for other quotation paradigms), the
underlying camel.sty may answer some of the needs of the humanities already,
like: 

-- automated crossreferencing and indexing of quotations, 

-- automated genereation  of separate bibliographies for sources and
secondary         literature,

-- automated "Ib."-functions,

-- a number of BibTex-macros provided to make writing own sty-files easier,

-- and more 

It's only an "alpha" so far, and according to the author development has
been stopped, until BibTex 1.00 will come out :-( -- while from another
"usually well informed source" on de.comp.tex I learnt that BibTex 1.00
might not bring the quantum leap everybody, it seems, yearns for: 2 x :-(
the situation seems to be kind of a mess, doesn't it?


however, together with the 8bibtex by kempson mentioned here recently, 
the camel-package tempted me to have my first look at BibTex, and that means
something :-)

greetings,

thomas 

p.s. many thanks  to all the package-authors for TeX and LaTeX, btw. 

11th October 1996: Re: custombib

From: Thomas Martin Widmann 

Emma Pease (emma@CSLI.STANFORD.EDU) wrote:

> I note the latest version for custom-bib was released June 1996 (3.81)
> with minor modifications added up to Sept. 18 so you might want to
> double check.  I confess not to use it myself but to be in the habit
> of modifying .bst files directly when necessary.

OK, I see I'll have to check it out again.

> > As I see it, the major problems with the current BibTeX are:
> > 
> > *  The .bst files are unreadable.  Of course, if you only want to change
> 
> My guess is that for most people (non-programmers) something like
> custom-bib should be used if modifications are needed.  

You're probably right; I just feel that .bst syntax is unnecessarily
ugly.  Of course, this is no reason to reinvent the wheel, but if we're
going to redesign the whole thing, this is one of the points I'd like to
change at the same time.

> > *  It is very monolingual by design (check my previous posting about
> >    this point).
> 
> A problem.  I think the basic changes should be possible (month names,
> etc).  Perhaps working on a tool for bib file manipulation would be
> helpful.  I tend to use bibtex mode in emacs [with some local changes
> such as using abbrevs]; however, it isn't the greatest.

Yes, basic changes are possible.  Sometimes the changes (especially in
the case of more or less exotic languages) will, however, be so great
that you've got to be an experienced programmer to do it, which is
A BAD THING.

By the way, wouldn't it be nice if the Babel package would change the
month names in the bibliography, too?

A tool for bib file manipulation would be great, but I guess it will be
difficult to make this program run on all the platforms that support
TeX (not all of them have fancy graphical interfaces, for instance).
I normally use Emacs' bibtex mode, too, and I'm not sure that a program
for all TeX-aware computers would be much better.

> Sorting for 8-bib I suspect may require some internal bibtex recoding.
> I note there is an 8-bit bibtex out there by Neil Kempson
> 
>  ftp.tex.ac.uk:/tex-archive/biblio/bibtex/8-bit
> ..... 
> I haven't played around with it.  

Neither have I, but I had a brief look at it half a year ago.  Rather
nice, but it only solves a few of the problems, of course.  As far as I
remember, it offers no solution if you happen to use two alphabets in
your document (e.g., Latin and Cyrillic) and want to use both in the
bibliography.  And of course, if we all migrate to Omega (16-bit TeX) in
one or two years, we'll have the same problems again.


> > *  It is not very flexible if non-standard entries are needed (check the
> >    fine posting by Harold F. Schiffman about this one).
> 
> True. I have been known to recommend hand editing the bbl file when
> nothing else works.  However, you can add new types fairly easily to
> the bst file.

Agreed, but hand editing the bbl file is a work-around (which I hate)
and I'm afraid that a lot of users wouldn't agree that it is fairly easy
to add new types to the bst file (if custom-bib also solves this, I'm
impressed).

> I did a hunt through the comp.text.tex archives (dejanews is an useful
> little beast).  Oren P. the original writer of bibtex is suppose to be
> working on a complete overhaul of bibtex; however, I think progress
> has been very slow maybe even non-existent (he still has an account
> op@cs.stanford.edu if we want to ask him).  Neil Kempson is another
> person possibly to ask .

I'll try to mail Oren Patashnik.  At least, it would be nice to know
whether his new BibTeX is supposed to abound of new features of not.


Thomas Plagwitz (tplagwit@aixterm1.urz.uni-heidelberg.de) wrote:

> I would like to point out, that there is a LaTeX macro package on top of 
> BibTex, to be found on:
> 
> CTAN: /tex/macros/latex/contrib/supported/camel

Thanks a lot, I heard about Camel some time ago, and immediately forgot
about it again.  I'll try to check up.  At least some of the features
sound promising.

> It's only an "alpha" so far, and according to the author development has
> been stopped, until BibTex 1.00 will come out :-( -- while from another
> "usually well informed source" on de.comp.tex I learnt that BibTex 1.00
> might not bring the quantum leap everybody, it seems, yearns for: 2 x :-(
> the situation seems to be kind of a mess, doesn't it?

Well, again this BibTeX 1.00.  I'll definitely mail O.P.
 

Thomas
ONF Thomas Martin Widmann |http://ling.hum.aau.dk/~viralbus| Tlf +45/8616 5464
A. H. Winges Vej 3,st.tv. |viralbus@{ling.hum,daimi}.aau.dk| Lernu Esperanton!
DK-8200  Århus N,  Europa |  CAESAR*NON*SVPRA*GRAMMATICOS  | Stud (Ling. & CS)

29th October 1996: Modifying the label in cm-lingmacros.sty

From: cthiele@ccs.carleton.ca (Christina Thiele)
Subject: modifying the label in cm-lingmacros.sty

I've spent the better part of today being repeatedly informed by LaTeX
that I really have no clue as to how the list environment works ;-((

I'm trying to do what seems to be such a simple thing ... [a familiar
opening refrain, I'm sure!]


Using cm-lingmacros.sty, I have to deal with examples (using both
\enumsentence and \eenumsentence) which will go up into three digits
-- the largest label is something like 128 or thereabouts, with
parentheses around the digit.

The task is to put those digits flush to the left margin, rather than
indented a certain distance in from the left ... and then to ensure
that all example text starts at the same margin. That is, a fixed
labelwidth, but flushleft, rather than flushright ... at least, I
think that's what I want to say ... 

A drawing, at this point ... my heart's desire:

(1)    a. The sample sentence here.
       b. The second part to the example.

(11)   Another sample sentence. 

(111)  a. A third sample sentence.
       b. The second part to the example.


LaTeX's pleasant defaults ... 

  (1)  a. The sample sentence here.
       b. The second part to the example.

 (11)  Another sample sentence. 

(111)  a. A third sample sentence.
       b. The second part to the example.


I thought I'd be able to copy what the description environment does,
but when I tested the default first, by inserting (1), (11), and (111)
into the \item [], the text which follows is not at a uniform distance
.. it `travels', according to the \labelsep value, I suppose. Or
whatever ... sort of like this:

(1)  The sample sentence here.

(11)  Another sample sentence. 

(111)  A third sample sentence.

I've looked at the diagrams in Goossens et al. and in Michel's article
in the Cahiers; I've looked in the Hahn book, in the Lamport book ...
I just cannot seem to get the list environment straight in my mind ...
which has now pretty much seized up. And while Goossens et al does
show how to set up fixed labels, I don't know that it's quite the
solution, given the other coding that's in the cm-lingmacros.sty
macros ...

If there's *anyone* who's using cm-lingmacros.sty and has found a way
to have up to 3 digits in the labels, with the example text (as well
as itemized a., b., sub-examples) all starting at the same position
..  puh-leeze send help!

.. it's starting to smell like burnt brain synapses around here ...
;-(

Ch.

P.S. If it turns out that this *is* covered in cm-ling...,
     then I deserve to have burnt goose! Right now, I'm feeling
     pretty cooked ...

31st October 1996: Round 2: flushleft labels in cm-lingmacros.sty

From: cthiele@ccs.carleton.ca (Christina Thiele)
Subject: Round 2: flushleft labels in cm-lingmacros.sty

Further to my post of yesterday about trying to get
flush-left/left-justified labels using cm-lingmacros.sty ... 

a. I've managed to get exactly what I want for \enumsentence; that is, 
   for examples where there are no \item s.

b. I've managed to get the labels to also be flush-left and
   left-justified for \eenumsentence ... but there's a small
   repercusion ... all \item's after the first one are having
   their labels (b. c. etc) flushed to the text's left margin as well! 

Below is the original code from cm-lingmacros, followed by my modified
version. Perhaps someone here can spot the reason that b. and c. and
so on are not behaving as wished -- instead, doing exactly what my
poor coding is telling them to do ... ;-( 

Original:

\long\def\@enumsentence[#1]#2{\begin{list}{}{%
\advance\leftmargin by\widelabel \advance\labelwidth by \widelabel}
\item[#1] #2
\end{list}}

\def\@mklab#1{\hfil#1}
\def\enummklab#1{\hfil(\eelabel)\hbox to \eeindent{\hfil#1}}
\def\enummakelabel#1{\enummklab{#1}\global\let\makelabel=\@mklab}


My version (lines prefixed with * are diff.):

\long\def\@enumsentence[#1]#2{\begin{list}{}{%
*\itemsep=0pt   \topsep=4truept  \parsep=0pt 
*\labelsep=1em  \leftmargin=3em
\advance\leftmargin by\widelabel \advance\labelwidth by \widelabel}
*\small\baselineskip=11pt
\item[#1] #2
\end{list}}

*\def\@mklab#1{#1\hfil}
*\def\enummklab#1{(\eelabel)\hfil\hbox to \eeindent{\hfil#1}}
\def\enummakelabel#1{\enummklab{#1}\global\let\makelabel=\@mklab}


This is for simple examples, with no sub-parts, and yields the
following output:

(1)    Text starts here. 

(11)   Text starts here.

(111)  Text starts here.

Lovely ... !


Now for the other problem, that's only half-solved:

Original:

\long\def\@eenumsentence[#1]#2{\def\eelabel{#1}\let\holdlabel\makelabel%
\begin{list}{\alph{enumsi}.}{\usecounter{enumsi}%
\advance\leftmargin by \eeindent \advance\leftmargin by \widelabel%
\advance\labelwidth by \eeindent \advance\labelwidth by \widelabel%
\let\makelabel=\enummakelabel}
#2
\end{list}\let\makelabel\holdlabel}


My version:

\eeindent=17pt   %% default setting is 15pt

\long\def\@eenumsentence[#1]#2{\def\eelabel{#1}\let\holdlabel\makelabel%
\begin{list}{\alph{enumsi}.}{\usecounter{enumsi}%
*\small\baselineskip=11pt
*\itemsep=0pt   \topsep=4truept \parsep=0pt
*\labelsep=.5em  \leftmargin=2.7em
\advance\leftmargin by \eeindent \advance\leftmargin by \widelabel%
\advance\labelwidth by \eeindent \advance\labelwidth by \widelabel%
\let\makelabel=\enummakelabel}
#2
\end{list}\let\makelabel\holdlabel}


The output from this is:

(1)   a.
b.

(11)  a.
b.

(111) a.
b.

So ... I do get the digits to be left-justified, but this has been a
tad over-generalised to now include item b. c. and so on  ... ;-(

Why do subsequent \item s inside \eenumsentence no longer line up
under the first item -- a. -- and how can they be made to behave
normally?

Any suggestions? 

Ch.

1st November 1996: Final round: solution found

From: cthiele@ccs.carleton.ca (Christina Thiele)
Subject: Final round: solution found

In the end, the solution to the problem of having example numbers
flushed to the left margin, leaving a fixed space with room for up to
3 digits, is embarrassingly simple:

To achieve:
            (1)   Text begins.
            (11)  Text begins.
            (111) Text begins.

the following line in the \long\def for \@enumsentence:

   \item[#1] #2 

should simply have an \hfill inserted:

   \item[#1\hfill] #2


To achieve:
            (1)   a. Text begins.
                  b. Text begins.
            (11)  a. Text begins.
                  b. Text begins.
            (111) a. Text begins.
                  b. Text begins.

the following command:

   \def\enummklab#1{\hfil(\eelabel)\hbox to \eeindent{\hfil#1}}

only needs to have \hfil(\eelabel) switched around:

   \def\enummklab#1{(\eelabel)\hfil\hbox to \eeindent{\hfil#1}}


.. sooo ... only two small changes -- a matter of seconds -- will get
the required results which other cm-lingmacros users may also find
useful on occasion ...

Ch.  ... can't hang the head in embarrassment low enough .. ;-(( 

P.S. It would be reassuring if someone else could reproduce these
     results ... since it my solution is the result of the
     hit-and-miss approach to TeX macros ... 

30th November 1996: Transliteration standards

Transliteration standards

I am the new chair of the International Organization for Standardization
subcommittee responsible for transliteration (ISO/TC46/SC2: Conversion of
Written Languages). The new secretary (Evangelos Melagrakis from Greece) and
I hope to make it far more visible and far more relevant to end users than it
has been in the past.

To enable this, an electronic mailing list for ISO/TC46/SC2 (tc46sc2@elot.gr)
has now been set up by ELOT (the Greek national standards body). We hope this
list will attract researchers and scientists who can add useful information
which might assist in developing standards on the Conversion of Written
Languages.

We also hope to have an emphasis on issues of using computers to do
appropriate transformations necessary in automated transliteration, and also
look forward to having regular contact with those on this list who are
interested in such issues.

There are quite a few with an interest in transliteration in library
catalogues on the list, but there are other potential users of
transliteration too. People are now begining to realist that
transliteration may have more indirect impact on other aspects of
multilingual computing than they had previously realised.

There are now over 200 subscribers to tc46sc2@elot.gr, from 38 countries
and territories - one of the most international lists around.

One major advantage of email is the ability to involve far more people in
the development of a common purpose than were involved before, to get user
feedback, and expert opinion from various sources.


Subscribing to the mailing list for ISO/TC46/SC2

In order to join the list you should be actively involved in using
transliteration systems, or in developing transliteration systems, and should
be prepared to contribute to the list from time to time.

If you meet these criteria, and wish to join the list, send an email to
majordomo@elot.gr with this message in the body of the text:

        subscribe tc46sc2 your@email.address

(but with your real email address replacing the string your@email.address).
To find out further commands you can use, send the command "help" as the text
of an email either to tc46sc2-request@elot.gr or to: majordomo@elot.gr
To unsubscribe, send the command "unsubscribe" instead, omitting the "quotes"
marks in both cases. This will tell you how to obtain copies of past
messages etc., and other useful features.

Once you are subscribed, you can send messages to tc46sc2@elot.gr and receive
messages from other members of the list. Please reply where possible to the
list as a whole, so that all can benefit: using the Group Reply function
(pressing G on some email software) is the simplest way to achieve this.

Other members will also be interested to see who else is joining the list, so
it is useful to send a brief introduction (say, one or two short paragraphs)
to tc46sc2@elot.gr at the outset, saying what languages, scripts and other
things you are involved in. That is the most likely way to stimulate others
to write on the subjects you are interested in!

You should also inform your national standards body to express your interest
in participating in this list. I can provide some information on details of
your national member body of ISO, if you send me an email requesting this.

I look forward to seeing new participants on this list. Please feel free to
forward this to anyone else who may be interested in transliteration
standardisation issues, and to send any queries about the list to me.

                                   Yours sincerely


                                     John Clews
--
   John Clews (Chairman of ISO/TC46/SC2: Conversion of Written Languages)

SESAME Computer Projects, 8 Avenue Rd.  *  email: Converse@sesame.demon.co.uk
Harrogate, HG2 7PG, United Kingdom      *     telephone: +44 (0) 1423 888 432

Last updated 7th April 1998 by Dag Langmyhr.