TeXhax Digest Monday, October 19, 1987 Volume 87 : Issue 83 [SCORE.STANFORD.EDU]TEXHAX83.87 Editor: Malcolm Brown Today's Topics: tex for Pyramid 4.0 Unix previewer for HP 350 WS Question about BibTeX enhancement AMS-TeX SUN TeX Previewer Re: Problem with BibTeX LaTeX Notes (Re: TeXhax Digest V87 #81) Re: TeX->spell Re: TeX on Gould PowerNode?? LaTeX [c]-type alignment in tabular environment Generating DVI Drivers. Re: Hershey fonts |-> METAFONT ? Hershey fonts Re: Hershey fonts |-> METAFONT ? Hershey fonts revisited ---------------------------------------------------------------------- To: texhax@score Subject: tex for Pyramid 4.0 Unix Date: Thu, 08 Oct 87 18:51:46 PDT From: Richard Roy Has anybody installed TeX successfully on a Pyramid running their 4.0 UNIX? Specifically, are there any problems with their pascal compiler and some of the nonstandard features used in tex such as the others/otherwise clause and possible char/int byte size problems? If so, do appropriate initex.ch, tangle.ch, and weave.ch files exist to handle these problems? R. Roy ------------------------------ To: texhax@score Subject: previewer for HP 350 WS Date: Thu, 08 Oct 87 19:27:30 PDT From: Richard Roy Does anyone have such a program for previewing dvi files on HP 350 workstations (HP 98782A terminals)? RR ------------------------------ To: texhax@score.stanford.edu Subject: Question about BibTeX enhancement Date: Fri, 09 Oct 87 05:09:04 -0400 From: sharat@cvl.umd.edu A while ago, there was some discussing about enhancements to BibTex. I was wondering if there is going to be, or there is already a program that will take a BibTeX database as input and certain keywords and print out, as ouptut, a human readable bibliographic citation as output (that matches all the keywords in some boolean fashion). Something on the lines of the UNIX "lookbib". Any pointers to answers to the above will be greatly appreciated, as I am considering writing some hacks to do exactly that. sharat chandran Electronic: sharat@cvl.umd.edu Arpanet/Domain ...!uunet!mimsy!sharat Uucp ------------------------------ Date: 16 Feb 87 8:54 EST From: ACS_CONNIE%JMUVAX1.BITNET@WISCVM.WISC.EDU To: TeXhax@score.stanford.edu Subject: AMS-TeX Hello, We are having trouble using AMS-TeX on our VAX/VMS system. This is the first anyone has tried to use it, and we get the message that is follows below. According to the message, I need to increase the hash size. How do I do this and which input file defines this constant? What should I increase it to? Is there anything else we need to do (ie. at compile time) when we use AMS-TeX that is different from using LaTeX? Any help would be greatly appreciated. PLEASE REPLY DIRECTLY TO ME. *** ERROR MESSAGE *** This is TeX, Vax/VMS Version 2.0 (preloaded format=lplain 86.5.6) **&lplain paper LaTeX Version 2.09 - Released 7 January 1986 (ACC1:[TEX.AMSTEX]AMSTEX.TEX;3 AmS-TeX - Version 1.1a ! TeX capacity exceeded, sorry [hash size=2500]. l.1181 \mathchardef\ varSigma="0106 If you really absolutely need more capacity, you can ask a wizard to enlarge me. Here is how much of TeX's memory you used: 554 strings out of 2520 5619 string characters out of 29562 34942 words of memory out of 65501 2500 multiletter control sequences out of 2500 18998 words of font info for 72 fonts, out of 35000 for 100 14 hyphenation exceptions out of 307 7i,0n,10p,385b,15s stack positions out of 200i,40n,60p,500b,600s No pages of output. Connie R. Minnick James Madison University Harrisonburg, VA 22807 ACS_CONNIE@JMUVAX1 (Bitnet) ------------------------------ Date: Thu, 8 Oct 87 19:43:18 EDT From: Paul Fishwick To: texhax-request@SCORE.STANFORD.EDU Subject: SUN TeX Previewer Does anyone know of a good TeX previewer for the SUN? (preferably at nominal cost or public domain). We are running U. of Washington's package. Thanks. Respond to: fishwick@fish.cis.ufl.edu -paul ------------------------------ Mail-From: PATASHNIK created at 9-Oct-87 09:16:56 Date: Fri 9 Oct 87 09:16:56-PDT From: Oren Patashnik Subject: Re: Problem with BibTeX To: texhax@Score.Stanford.EDU > From: Peter Galvin > Subject: Problem with bibtex > There seems to be a problem with standard BIBTEX, at least as built > on a Sun workstation. It appears that it fails to read files with > capital letters in their file names. Given as input XXX.bls, > it gives the error: > Unable to open XXX.bls > Am I hallucinating, or does this happen on other systems? A current BibTeX misfeature---insensitivity to uppercase letters in file names---has been eliminated in the new version of BibTeX (to appear shortly). --Oren Patashnik ------------------------------ Date: Fri, 9 Oct 87 11:01:32 pdt From: lamport@src.DEC.COM (Leslie Lamport) To: TeXhax@Score.Stanford.edu Subject: LaTeX Notes (Re: TeXhax Digest V87 #81) Toby Howard writes It's often useful to create index entries on the fly, such as when the thing you're indexing is first defined in a macro. But when there's an underscore involved, you get an undesired expansion... There are hacks to get around this problem. However, before hacking, it's usually a good idea to step back and rethink the entire problem. I presume that the problem is not the undesired expansion of the macro, but the fact that this expansion messes up the processing of the index file--probably by messing up the alphabetization. The general solution to this problem is the one utilized by the Makeindex program--namely, to allow two separate strings, one to specify the symbol in the index entry and another for use in alphabetizing the index. I believe that this approach is more useful and more general than any hack for fine-tuning the expansion of index entries. Glenn Vanderburg describes an extension of the LaTeX tabular and array environments to simplify the typesetting of columns of numbers with decimal points. Such extensions (which should be made only by adding a document style option, never by changing LaTeX itself) are useful in special applications--for example, for use by an accounting firm that typesets lots of columns of numbers with decimal points. However, one should be aware that changing such a complicated environment as tabular or array is likely to cause problems. Mr. Vanderburg points out several such problems, including the bad interaction with \cline and \multicolumn, but there are probably others that he hasn't discovered. The convenience will outweight these problems only for users who have lots of columns of numbers to enter. Users who only occasionally enter columns of numbers should not use this extension; the convenience offered by the feature will not offset the time they waste when they eventually are bitten by one of the problems. Eric Hildum writes I have looked through the LaTeX and TeX manuals and have not been able to determine how I can specify the length (\linewidth - 2\fboxsep) to \parbox. Does TeX or LaTeX allow sums of lengths to be determined? Section 5.4.1 of the LaTeX manual, entitled "Length", describes the commands \newlength, \setlength, and \addtolength; they provide the necessary functionality. Leslie Lamport ------------------------------ To: TeXhax@Score.Stanford.edu Subject: Re: TeX->spell Reply-To: "Christopher A. Kent" Date: Fri, 09 Oct 87 12:16:25 -0700 From: kent@decwrl.dec.com When I was keyboarding my thesis into LaTeX a couple of years ago, I complained bitterly to our TeXhacker, Dan Trinkle, about the crummy tools for using spell. The detex and delatex that came on the Unix distribution tape just didn't cut it for me. Dan coded up much nicer versions of both programs. They actually got merged into one, finally, and understand \input and \include and lots of other good things. For some reason, nobody outside Purdue seems to know about these programs, even though Dan has tried a couple of times to get the Unix tape builders to distribute them. Dan, are these still available for public FTP? Could you tell the folks how? Maybe public acclaim will get the job done. Cheers, chris %%% Dan Trinkle replies: There are a few people out there that got it and used it. I have made a few enhancements based on suggestions from those users. Specifically, it now allows the user to specify latex environments to ignore (with a default list) and it has a "-n" option to inhibit the processing of \include and \input commands (single file processing). It also detects unterminated modes and environments at EOF. The latest version (2.4 - 06/30/87) is available for anonymous ftp from arthur.cs.purdue.edu (192.5.48.1, 128.10.2.1) in pub/detex.tar (25KB) or pub/detex.tar.Z (11KB). It is also available for UUCP users from purdue in /usr/spool/uucppublic/detex.shar (19KB). Daniel Trinkle trinkle@cs.purdue.edu ARPA Computer Science Department trinkle%purdue.edu@relay.cs.net CSNET Purdue University {ucbvax,decvax,ihnp4}!purdue!trinkle UUCP West Lafayette, IN 47907 (317) 494-7844 PHONE ------------------------------ To: texhax%score.stanford.edu@Q2.ICS.UCI.EDU, jsweet@Q2.ICS.UCI.EDU Subject: Re: TeX on Gould PowerNode?? Date: Fri, 09 Oct 87 14:16:16 -0700 From: Jerry Sweet > Does anyone know of a Gould implementation of TeX, and if so, > how was it done? > > Peter Ilieve peter@memex.co.uk > peter@memex.uucp Something better MUST have come along by now, but just in case it hasn't: In December, 1985, I did a port of UNIX TeX 1.3 to UTX on a Gould Concept 32/67 using Meridian-Pascal (aka LNS-Pascal) version 2.2a. However, it failed one of the trip tests (something to do with a large negative penalty value), and I ran out of time to work on it, so I never released it back to the University of Washington. Nevertheless, it seems to work fine, and the company for which I originally did the port continues to use it to this day. I also sent a copy of this Gould TeX to a fellow who ran it satisfactorily on a PowerNode (sorry, I don't know the model number) at Clarkson University in New York. With your understanding that (1) it is not REALLY TeX until it passes the last trip test; (2) absolutely no warranty is offered, I am willing to make tape copies of this version of Gould TeX to interested parties. You should send a 9-track reel tape, a self-addressed box in which the tape is to be returned, and return postage (preferably actual U.S. postage stamps) to: Jerry N. Sweet 30 Willow Tree Ln. Irvine, CA 92715 USA The Meridian-Pascal compiler is sold by Gould; I regret that I cannot provide it. -jns ------------------------------ From: Gordon Howell Date: Tue, 6 Oct 87 17:12:42 GMT To: TeXhax@score.stanford.edu Subject: LaTeX [c]-type alignment in tabular environment I am trying to build a table who's definition looks like: \begin{tabular}{|c||p{2in}|p{2in}|} The first column item is a single word usually, while the next two columns are usually short paragraphs. I want my output to look like: ------------------------------------------- | this is a 2in | this is a 2in | | parbox which | parbox which | | can be several| can be several| | lines long | lines long | ENTRY | But I want it | But I want it | | aligned with | aligned with | | the centre of | the centre of | | the box in col| the box in col| | number one. | number one. | ------------------------------------------- In normal \parbox, you can specify box alignment as [b] [c] or [t]; but according to the LaTeX manual, these parboxes are forced [t] style; thus the column one entry "ENTRY" goes to the top. How can I get it the way I want it? (assume different line heights for different lines) Is there a parameter which defines the height of the strut for a particular row of entries? Thanks. Gordon Howell gordon@hci.hw.ac.uk Scottish Centre for Human Computer Interaction Edinburgh, Scotland ------------------------------ Date: 07 Oct 87 16:21:56 bst From: mcvax!ed.ac.uk!G.Toal@uunet.UU.NET Subject: Generating DVI Drivers. To: texhax , info-tex@mail.aston.ac.uk Is there a recommended way for generating DVI to device X drivers? I have done one by hacking apart a DVIType - removing all writelns and replacing calls to give info by calls to an external device driver module, BUT it has most of DVItype still in there doing nothing (well, all the consistency checks) except slowing it down. My question is, is there a skeleton device driver in existence which does the job properly? I'm a bit worried that my effort will go stiff if the next version of DVIType changes (for instance). I thought I should ask now as I intend to convert from PXL files to GF files in the near(ish) future. Graham Toal. P.S. My hacked DVIType was at least hacked by a web change file - not by editing the Web or the pascal - hence why I'm a bit worried that my changes may go stiff if DVIType changes underfoot. ------------------------------ To: texhax@score.stanford.edu Subject: Re: Hershey fonts |-> METAFONT ? Date: Sat, 10 Oct 87 23:33:00 -0400 From: Ken Yap > How come no one to the best of my knowledge has ever converted Hershey fonts > to MF format? (Hershey fonts: a PD collection of about 1500 characters in > vector format, designed by someone named Hershey for Nat. Bureau of Standards > a while ago. They include Old English, Fraktur, Carolinian?, Cyrillic and a > lot of useful symbols not usually available with TeX) I've tried it myself with > limited success: I converted several OE characters (a very straightforward > procedure), but then I quit because the only METAFONT I had access to was a > self-port to a PC that was unbearably slow, and I'm not a graphics artist, and > I plainly don't have the time (I'm a grad student burdened by 4 courses and a > job.) However, for someone into METAFONT and with plenty of time, this project > would be a piece of cake (relatively), and I'm willing to help a little. > The results SHOULD be freely available to everyone. > > |---> Dimitri Vulis (cldlv%nasagiss.bitnet@wiscvm.wisc.edu ;) I embarked on this project a while back and got some reasonable preliminary results but hit some snags. The major problem is that Hershey characters are defined by a series of strokes with minimal pen lifting and it is hard to distinguish between a vertex where the pen takes a curve and where the pen begins a totally unrelated path. I tried adding some heuristics to detect the difference, but they doesn't work well. Turning on spline fitting in MF makes the characters look smoother and nicer but makes simple characters like the numeral 1 worse, which becomes a big curve extending outside the bounding box. Another thing the Hershey fonts require is an estimation of the character baseline. I believe the only thing to do might be to do hand tuning. I may be wrong and it may be easier than I think but since I was short of spare time, I shelved the project. It would have been nice to have MF descriptions for the oriental characters. It is my opinion that there is a limit to how good Hershey fonts can look because they are characters designed for pen plotters. What do you expect for something free? Ken ------------------------------ Date: Sun, 11 Oct 87 10:57:48 EDT From: Dimitri Vulis Subject: Hershey fonts To: TeXHaX@score.stanford.edu Sure Ken, that's what I meant when I said 'plenty of time'. Let's consider Old English uppercase D as a case study. First, a program reads Hershey font and generates MF program that lookes sort of like z0=(10,10); % all breakpoints in segs z1=(30,30); .... pickpen ... draw z0--z1--...cycle; Note that I use -- for the first approximation. Then I print this out (gf2dvi) with points labeled and see visually what lines have to be deleted altogether, because they are used to fill dark areas, which lines are supposed to be reversed and joined to make filldraw and where I can change -- to .. and eliminate some points and add penpos's. Suppose you have a shape like the one on 1 ** 2 the right. In MF, you want to use 3 penpos's ... and a filldraw (at least, I would). In ..... Hershey, you have 3 segement: z1--z3--z5 ...... z2--z4--z6 and the one in the middle to 3 *..*...*4 signify fill. It's easy to see but I ........ really don't see how this can be done 5 *..*...* 6 by a program. The results were very good looking, but FEW, because on the equipment I had it takes more than an hour per character, most of it in waiting---for MF, for GF2DVI, for printer... I'm not sure if baseline is a problem---I thought all H characters have the baseline at the same level? Also, in the data I have, the segments are all separated. Thus, if there is someone out there who has MF on a VAX or SUN and has plenty of time, it would be nice to have all those Gothic characters available for TeX (wishful thinking, probably). It would be a fair amount of work, and I guess quite a few people have tried it and gave up. Perhaps someone who teaches could give it to a student as a project? What kind of Oriental characters did you have in mind? There was a paper long time ago in Byte about Devanghari in MF; and I'm sure someone has Kanji by now; and of course there is the famous paper about Chinese characters. I doubt if anyone has a large number of Chinese characters written in MF. As I said, Hershey characters do look good, and a good graphics artist can probably tune them up and make them look even better; and yes, some of the best things in life (incl. computer-related) are free. Another possible source of cheap low-quality MF fonts are the myriads of PD raster fonts available from BBS's etc. One can fit a spline on the white-black boundary and fine-tune it by hand. There was a ver interesting paper I think in Pat.Recog. by some people from Stanford (don't have the reference offhand). I'm going to look into it myself, since this is related to what I'm doing, but I don't think I can actually get any fonts this way again. Dimitri ------------------------------ Date: Mon, 12 Oct 87 12:09:33 CDT From: William LeFebvre Subject: Re: Hershey fonts |-> METAFONT ? To: TeXhax@score.stanford.edu > How come no one to the best of my knowledge has ever converted > Hershey fonts to MF format? It's an interesting possibility. Two problems: (1) because of the limitations in the Hershey specification (straight lines drawn between points on a relatively coarse grid), the results would probably only be acceptable as a starting point (i.e.: further refinement would be necessary to make them look good), and (2) the Hershey fonts are pretty poor to begin with. But at least there's a script font in the Hershey set---something I've wanted for quite some time. William LeFebvre Department of Computer Science Rice University ------------------------------ Date: Mon, 12 Oct 87 16:00:47 EDT From: Dimitri Vulis Subject: Hershey fonts revisited To: William LeFebvre Dear William: You're right of course, one can't just take them (Hershey fonts) and plug them into METAFONT 'as is' and hope to get decent results. I described the (slow&painful) procedure that I used in my letter to Ken Yap which I cc:. H-fonts are just a bunch of crude plotter fonts; converting them to MF involves seriously tuning up the shapes, adding kerning etc. This requires time and drawing ability, neither of which I have, but it's nothing overly intellectual. Certainly, it would be very nice to have script, old english, planet symbols etc freely available for TeX (well, some of this is available NOW from AmS, but it costs $$$ and comes w/o MF source.) So, the reasonable question is, why is this guy (me) raising this topic (Hershey fonts) if he's got no time to do it himself? Well, I was hoping that someplace on bitnet, there an (angel? undergraduate looking for a project?) who'd jump at the idea, once aware of it. 2 e=mc Dimitri ------------------------------ %%% %%% subscriptions, address changes to: texhax-request@score.stanford.edu %%% %%% submissions to: texhax@score.stanford.edu %%% please send a valid arpanet address!! %%% %%% BITNET redistribution: TEX-L@TAMVM1.BITNET (list server) %%% %%%\bye %%% ------------------------------ End of TeXhax Digest ************************** -------