TeXhax Digest Monday March 3, 1989 Volume 89 : Issue 32 Moderators: Tiina Modisett and Pierre MacKay %% The TeXhax digest is brought to you as a service of the TeX Users Group %%% %% in cooperation with the UnixTeX distribution service at the %%% %% University of Washington %%% Today's Topics: General information regarding Unix TeX Wanted: tree formatter TeX vs. LaTeX Needed: Metafont sources for Chinese and Japanese fonts Phonetic font and large files Re: Concerning nested conditionals (More on) concatentation of `fil' and `L' Pictex in latex; bitnet exclusion from texhax distribution ------------------------------------------------------------------------- Date: Sun, 2 Apr 89 11:49:36 PDT From: mackay (Pierre MacKay) Subject: General information regarding Unix TeX Keywords: UNIX, TeX The base price for a full distribution of TeX, is $140.00 for 1/2 inch 9-track tapes, $165.00 for 4-track 1/4 inch cartridge tapes. This is for prepaid orders, sent within the Continental U.S., by UPS, surface routing, with delivery in eight days from dispatch. For shipping charges to other sites and for rush orders, see the table below. We can also accept purchase orders, with invoice after delivery, but there will be an extra charge of $10.00, owing to the invoice processing charges we incur. We are required to have written confirmation of orders (no phone confirmation) and we do not have fax facilities. Direct payment by wire can be made to UNIVERSITY of WASHINGTON Account # 002-138-0641 For: Northwest Computing Support Center, # 14-0449 in payment of UW INVOICE # ------------- Rainier Bank -- University Branch 1300 - NE 45th Street Seattle, WA 98105 U.S.A. Payment by wire also involves an extra charge of $10.00 to cover bank handling charges, but you do not need to pay the $10.00 invoice charge if we do not have to send an invoice. TOTAL costs, including shipping charges are: ----- (add $10.00 for invoicing or payment by wire) |-------|-------|--------|--------|---------------|-----------| | Local | Local | Canada | Canada | Overseas | Overseas | | UPS | DHL | UPS* |Airborne|Air Parcel Post|DHL Courier| |-------|-------|--------|--------|---------------|-----------| magtape |$140.00|$148.00|$140.00 |$160.00 | $160.00 | $170.00 | |-------|-------|--------|--------|---------------|-----------| cartridge|$165.00|$173.00|$165.00 |$185.00 | $185.00 | $195.00 | |-------|-------|--------|--------|---------------|-----------| *NOTE: UPS shipments to Canada only available to Province of Ontario and to metropolitan areas of Montreal, Vancouver and Victoria. Checks should be in U.S. dollars, payable to The University of Washington, (IRS Tax number 91-6001537) and sent to: The Director Northwest Computing Support Center, DW-10 University of Washington Seattle, Washington 98195 The normal distribution is a tar tape, blocked 20, 1600 bpi, on an industry standard 2400 foot half-inch reel. If you need 1/4 inch streamer cartridges, be sure to tell us. The 1/4 inch physical format is either QIC-11, 8000 bpi, or QIC-24, 4-track serpentine recording. We cannot write nine-track on cartridges, nor can we write TK50 cartridges for the DEC Microvax. SystemV tapes can be written in cpio format, blocked 5120 bytes, ASCII headers, but we prefer not to, since cpio format is extremely slow and wastes a great deal of tape on inter-record gaps. Again, please specify this format if you want it, and make sure which of the several QIC formats you can read. Most systems can read both QIC-11 and QIC-24, but very old systems can read only QIC-11, and some 3Bx s will only read QIC-24. We can also offer the distrtibution on a NeXT format laser-disk. The original organization of the distribution reflected the use of pascal for all compilations of TeX, TeXware, BibTeX, METAFONT and MFware. This has now been supplemented by a more convenient and more generally portable WEB-to-C compilation for TeX, TeXware, BibTeX, METAFONT and MFware (except for GFtoDVI, which is currently being rewritten by Donald Knuth). LaTeX and AMSTeX are macro packages processed by TeX, and do not require any further compilation. Current versions of standard programs in the distribution: (in the ./tex82 path) TeX 2.98 (implying the use of the new cm fonts.) (plain.tex version 2.94) LaTeX 2.09 (release of 4 Aug, 1988) (also SliTeX 2.09) tangle 2.9 weave 3.1 dvitype 2.9 pltotf 2.3 tftopl 2.5 metafont 1.7 gftodvi 1.7 gftopk 1.4 gftype 2.2 pktype 2.2 pktogf 1.0 bibtex 0.99c mft 0.3 (a formatting program for metafont source files.) also: (chtopx gftopxl pktopx pxtoch pxtopk --- but the use of these obsolescent pxl-related programs is discouraged) Various foreign language utilities, German, Greek, Portuguese, Swedish, and a first run of TeX-XeT for Semitic languages. Lots of other goodies of a generally similar nature. Fonts in tfm (TeX Font Metric) format. This is the format that describes fonts for TeX. It includes all useful information about font characters except what they will actually look like on paper. TFM format is size independent and device independent. and can therefore be used on any system that runs TeX. It will give you a valid DVI (DeVice Independent) output file, but will be of no use to give you readable hard (or CRT) copy output. Fonts in mf source format. The full Computer Modern as released from Stanford. Utility fonts for character proofs etc. (not made with cmbase.mf), LaTeX and SliTeX fonts are also supplied in mf source format. Fonts in PK format. This format produces characters that can actually be printed on paper. It is very device-dependent, and can therefore only be supplied for a small range of common output devices. All the standard shapes and sizes as declared in plain.tex, lfonts.tex, sfonts.tex and webmac.tex are here, in 118, 200, 240, and 300 PK(dpi) series. The 300dpi fonts are sent out in write-black and write-white flavors (CanonCX and Ricoh 4080). No provision is made for the global magnifications of LaTeX style files (such as bk11.sty). If your site needs these or similar magnifications you will have to run METAFONT to get them. The PK files for the principal LaTeX and SliTeX alphanumeric and symbol fonts are in a separate list. The Euler fonts (includes Fraktur) come in PK format only. (If you want [euler].mf source files, you should get in touch with the American Mathematical Society, which will make these available under license.) All these fonts may be converted to gf format (the format produced by the METAFONT program) by using the MFware program pktogf. Old METAFONT fonts. AMS fonts-- Cyrillc, and special symbols (created with old METAFONT-in-SAIL). Converted to PK format. There is no use in the mf files for these unless you are running a DEC10 or DEC20 with a SAIL compiler. Fonts in pxl format. No more. All the best drivers use gf or pk these days. If you really need pxl, you will have to compile and run pktopx. NOTE. for those who wish to go on using am series fonts, the files am_plain.tex, am_lfonts.tex and am_webmac.tex are provided. Drivers for dvi output. imagen(2 styles), qms(2 styles), LaserWriter (PostScript), LN03. These all produce intermediate files of one sort or another. Get in touch with Nelson Beebe Beebe@CS.UTAH.EDU if you need an HP LaserJet driver. Symbolics LGP, Ln01, Versatec. (These are all obsolescent, and are no longer maintained.) Compilation is now based entirely on Web-to-C, but the pascal compilation files have been left on the tape in compressed format (not updated). After about 6 months, they will be removed, unless there is some urgent reason for keeping them. Here is a partial list of successful compilations under Web-to-C: Sun-2, SunOS 3.2 Sun-3, SunOS 3.2 (no register variables), SunOS 3.4, 3.5, 4.0FCS (with register variables), all using the Sun-supplied C compiler, cc. The -O4 optimization level seems to work on both Sun-3's and Sun-4's in SunOS 4.0 for TeX, but not for pltotf.c, mf4.c, and mf9.c. With some early versions of gcc, you must define "shar" as "int" in site.h to avoid optimizer bugs (and then TeX's speed and size are comparable to the code produced by cc). With recent versions of gcc, changing schar is not necessary. Sun-4, SunOS 3.2-4, using cc -O, and SunOS 4.0FCS with cc -O4. Gcc 1.28 also passes the trip test, but the execution time of the resulting program is the same or slower than that compiled with cc. Gcc 1.30 seems to produce code which fails the trip test. Sun-386i, SunOS 4.0, cc -O, gargantuan version. Sequent Balance, Dynix 2.1.1, using cc, in both the Berkeley and ATT universes. Gcc 1.21 and earlier failed even to produce assembly code which would assemble, much less execute. 1.22's supposed to have this fixed, but the compiler dies with an unexpected signal :-(. Sequent Symmetry, Dynix 3.0.12, using cc -O. Vaxen running 4.2, 4.3BSD, and Ultrix, using cc. gcc 1.17 (and possibly later versions, although 1.25 and later are known not to have this bug) incorrectly compiles the "makemathaccent" routine, but generally produces *much* faster code. The data segment will be bigger, however, because gcc apparently packs the structure fourquarters very unoptimally. Again, this has been fixed in gcc 1.25. Convex: Everything works, but -O breaks web2c itself, so don't use it when compiling that program; it's ok for TeX etc. Amdahl running UTS (SysV). The type "schar" must be "short". Apollo, SR9.7 and SR10.0 (beta). Ridge 32 running ROS 3.5 and C compiler version 2.1B UNIXpc (aka 3b1 or PC7300) running System V version 3.51. MIPS R/1000, compiler version 1.21 Iris workstation (model and OS version not specified). Celerity C1260, UNIX version 3.4.78, without -O. The optimizer bug has been reported to Celerity (now part of FPS). Pyramid 98x, OSx64Q 4.0-870901, C Compiler CCOMP_4.0, large or small TeX's. AT&T 6386 WGS, running System V Release 3.1.1 Interactive 386/ix Version 1.0.5 (Unix System V Release 3.1) running on an i386 machine. See patches in the directory 386-ix. IBM RT/PC running AOS 4.3 (without -O). HP 9000/350. HP's version of lex, and possibly others, types yytext as unsigned char instead of char. Appropriate changes will have to be made in web2c on such systems. AT&T 3b2, small TeX, using the fpcc compiler. The files tex8.c and dvitype.c had to be modified because each contained a case statement too large for the compiler to handle. It may be necessary to remove the "-s" flag when loading virtex if you're going to use undump. ICL Clan 7 running System V (really a CCI Power6/32). HP 9000/840, compiled without "-O" and with "-lmalloc", both big and small versions of TeX. The 9000/850 is supposed to be object-code compatible, but it was not tested. Cromemco System 300 running Unixsoft System V.2 (TeX only). Sony News Workstation, 4.2BSD Unix Release 2.2, using both the native C compiler and gcc (both TeX and MF). Counterpoint System-19K, C-XIX 3.0. Sperry 7000/40 (Tahoe architecture) running 4.3BSD. Encore Multimax (32332's) running Umax 4.2 release 3.3. NeXT OS version 0.8 if you leave out the -s flag to the loader. (The compiler was gcc) Masscomp (but you must undefine REGFIX) Harris 8CX9 running HCX-UX (basically 4.2BSD), using standard cc compiler. (typedef short schar in site.h, and eliminate #include from extra.c) A DOS port is in progress. For now, you should ignore any references to MS_DOS in the code. The site distributions of TeX are coordinated through the TeX Users Group (TUG) as part of the services of TUG to its members. We urge all users of TeX and METAFONT to join the TeX Users Group, to support and benefit from its activities---particularly the journal, TUGboat. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % Call or write to: The TeX Users Group % % P. O. Box 9506 % Pierre A. MacKay % Providence, RI % TUG Site Coordinator for % 02940-9506 % Unix-flavored TeX % (401) 751-7760 % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% ---------------------------------------------------------------------------- Date: Tue, 4 Apr 89 11:51:35 EDT From: pfkrause@ATHENA.MIT.EDU Reply-To: pfkrause@wheaties.ai.mit.edu Subject: Wanted: tree formatter Keywords: LaTeX, tree formatter Here's a problem that's both new and hard, at least so far as I know. What I want to produce is something like this: S / \ NP VP / \ / \ Det N V NP | | | / \ The man saw Det N | | a boy (probably using the slanted lines from LaTeX's \picture font), with an input something like this: \tree{S}{\tree{NP}{\tree{Det}{The}% \tree{N}{man}}% \tree{VP}{\tree{V}{saw}% \tree{NP}{\tree{Det}{a}}% \tree{N}{boy}}} (if you can think up a better input format, I'd be happy to take that too). Does anybody have an approximation to such a monster, or think that they could write it? My thesis thanks you. I thank you. -paul pfkrause@wheaties.ai.mit.edu ------------------------------------------------------------------------------ Date: Wed, 5 Apr 89 02:34 +1000 From: Douglas Miller Subject: TeX vs. LaTeX Keywords: TeX, LaTeX Peter Flynn writes: > We do in fact need both [TeX and LaTeX]. ...there are many more complex > formatting requirements in the real world than LaTeX is suited to. Yes, but even highly complex TeX hacking can be hidden inside LaTeX style and style option files, and accessed by LaTeX-style "structural" commands. This is what I assumed to be the intention of LaTeX; it is your "document designer", which uses TeX as its "typesetter". -------------------------------------------------------------------------- Date: Wed, 5 Apr 89 10:19:37 EDT From: William Y. Huang Subject: Needed: Metafont sources for Chinese and Japanese fonts Keywords: fonts, Chinese, Japanese Does anyone have Metafont sources for Chinese or Japanese Kanji Characters? Thanks in advance. %%% Moderators' note: The only thing that exists now is the experimental %%% coding for old METAFONT-in-SAIL done by John Hobby, in cooperation %%% with Gu Guoan. This would require careful translation, but it appears %%% to be a very good start on a new mf coding. Hobby's stroke-based %%% approach could be associated with the stroke-sequence data developed %%% for the Caldwell SinoType. Both the Chinese mf code and the stroke %%% sequence files are available from the University of Washington. ----------------------------------------------------------------------------- Date: Tue, 4 Apr 89 14:15 EST From: STEVENS@SPEECH.MIT.EDU Subject: Phonetic font and large files Keywords: fonts, LaTeX Please help me if you can. I need to be able to use an phonetics font with LaTeX. Do you know of any that is available? Also, need to increase memory size of LaTeX so that we can process large files with many figures and tables. Presently we get the error: Too many unprocessed floats. Any suggestions? I do not subscribe to TeXhax, so perhaps you could reply direct to ann@speech.mit.edu. Thanks much. Ann or Carol 617-253-7309, 617-253-5957 --------------------------------------------------------------------------- Date: 03 Apr 89 1143 PDT From: Arthur Keller Subject: Re: Concerning nested conditionals Keywords: LaTeX, nested conditionals The problem with nested conditionals affects conditions defined with \def and not conditions defined with \newif Consider the following TeX source file: \newcount\counter \counter=1 \newcount\cnt \cnt=1 \newif\iftest \testfalse \def\condition{\iftest} \message{Case 1. } % use of \ifnum \ifodd\counter \message{True } \ifnum\cnt>0 \message{true } \else \message{false } \fi \else \message{False } \ifnum\cnt>0 \message{true } \else \message{false } \fi \fi \message{Case 2. } % use of \iftest defined by \newif \ifodd\counter \message{True } \iftest \message{true } \else \message{false } \fi \else \message{False } \iftest \message{true } \else \message{false } \fi \fi \message{Case 3. } % use of \condition defined by \def \ifodd\counter \message{True } \condition \message{true } \else \message{false } \fi \else \message{False } \condition \message{true } \else \message{false } \fi \fi \bye * Because \iftest is defined by \newif directly it is set to \iftrue or \iffalse by a \let command. The scanner recognizes it as an \if while skipping the text and handles the nesting properly. * \condition is \def to \iftest, but \condition is not expanded while skipping text for \ifodd, so the nesting is not properly handled. * Note that \def\condition{\if...\else...\fi} will work fine. (It is an exercise left for the reader to figure out why. Here is the output from the above TeX source. This is TeX, WAITS Version 2.98 (preloaded format=plain 89.3.2) 3 APR 1989 11:34 **TMP.TEX (TMP.TEX[1,ARK] \counter=\count25 \cnt=\count26 Case 1. True true Case 2. True false Case 3. True false ! Extra \fi. p.2,l.52 ^^I\fi ? x No pages of output. * Moral: For conditions, always use \newif directly. Arthur ------------------------------------------------------------------------------- Date: 4-APR-1989 10:10:18 GMT -01:00 (BST) From: CHAA006%vaxb.rhbnc.ac.uk@NSS.Cs.Ucl.AC.UK Reply-To: Philip Taylor (RHBNC) Subject: (More on) concatentation of `fil' and `L' Keywords: `fil', `l' As an afterthough to my earlier submission on the concatentation of `fil' and 'L' in constructs such as \hskip 0 pt plus 1 fil Left or Right on reading an earlier edition of `The TeX book', I now see that the present production rule, which explicitly produces `fil', `fill' and `filll', had an earlier, recursive, formulation, something like := fil | l Now, taking that in conjunction with P.268 of `The TeX book', where is written: "We shall use a special convention for keywords, since the actual syntax of a keyword is somewhat technical. Letters in type- writer type [such as] `pt' will stand for

" the `l` in the second alternative of the earlier recursive production rule can clearly be preceded by , and can equally clearly in either case. It seems possible that when Knuth changed the production rule to explicitly limit the number of `l's to three, he didn't change the WEB code which looks for all 'l's after the first as semi-independent keywords, capable of absorbing initial . One ( final?) correction: the recursive production rule in `The TeX book' is {\it later} than the non-recursive formulation ::= fil | fill | filll which occurs in the second printing, 1984. The recursive formulation appears in the sixth edition, 1986. I have no intermediate editions against which to check. So my comments on the possibility of Knuth changing the definition without changing the code were unwarranted. It is now not clear to me why the code implements an upper-bound of three 'l's, while the production-rule allows an infinite quantity. To be more precise, I do not understand why Knuth appears to have adopted an infinite production rule for a meta-notion with strictly-limited productions, when the original formulation, while not indicating that spaces could be soaked up, at least made it plain that at most three 'l's could occur. Might not the `best' production rule be something like ::= fil | fill | filll ** Phil. -------------------------------------------------------------------------- Date: Tue, 4 Apr 89 10:17 EST From: Subject: Pictex in latex; bitnet exclusion from texhax distribution Keywords: PiCTeX Dear Texhax, What is going on? Why have I received only issue 12 of this year's edition on bitnet? I have a pictex question that I would like help on but obviously the whole idea of texhax has broken down if everyone on bitnet is cut off. Please respond directly to me in addition to your response to texhax. How does one use pictex in latex? Several people I know who use latex have had memory problems just loading the three pictex macro files needed for latex use. I don't use latex so I have escaped this problem until now but I am unable to advise these people. Second, pictex seems to exceed tex's memory quickly, which I have somewhat gotten around by repeated use of the \savelinesandcurves and \replot commands but for an involved diagram, even this fails. Is pictex a serious macro package or just a halfbaked idea? cutoff and unhappy in Philly, bob jantzen jantzen@vuvaxcom.bitnet ------------------------------------------------------------------------------- %%% The TeXhax digest is brought to you as a service of the TeX Users Group %%% in cooperation with the UnixTeX distribution service at the %%% University of Washington %%% %%% Concerning subscriptions, address changes, unsubscribing: %%% BITNET: send a one-line mail message to LISTSERV@UWAVM %%% SUBSCRIBE TEXHAX % to subscribe %%% or UNSUBSCRIBE TEXHAX %%% %%% All others: send a similar one line mail message to %%% TeXhax-request@cs.washington.edu %%% Please be sure you send a valid internet address!! %%% in the form name@domain or name%routing@domain %%% and use the style of the Bitnet one-line message, so that %%% we can find your subscription request easily. %%% %%% All submissions to: TeXhax@cs.washington.edu %%% %%% Back issues available for FTPing as: %%% machine: directory: filename: %%% JUNE.CS.WASHINGTON.EDU TeXhax/TeXhaxyy.nn %%% yy = last two digits of current year %%% nn = issue number %%% %%% For further information about TeX Users Group services and publications %%% contact Karen at KLB@SEED.AMS.COM or write to TUG at %%% TeX Users Group %%% P.O. Box 9506 %%% Providence, R.I. 02940-9506 %%% Telephone (401) 751-7760 %%% %%% Current versions of the software now in general distribution: %%% TeX 2.95 (2.98 coming) metafont 1.7 %%% plain.tex 2.94 plain.mf 1.7 %%% LaTeX 2.09 ( 8/10/88) cmbase.mf see cm85.bug %%% SliTeX 2.09 gftodvi 1.7 %%% tangle 2.9 gftopk 1.4 %%% weave 2.9 gftype 2.2 %%% dvitype 2.9 pktype 2.2 %%% pltotf 2.3 pktogf 1.0 %%% tftopl 2.5 mft 0.3 %%% BibTeX 0.99c %%% AmSTeX 1.1d %%%\bye %%% End of TeXhax Digest ************************** -------