TeXhax Digest Tuesday, December 22, 1987 Volume 87 : Issue 104 [SCORE.STANFORD.EDU]TEXHAX104.87 Editor: Malcolm Brown Today's Topics: tpic LaTeX Notes, Part I (Re: TeXhax Digest V87 #101) TUGboat Vol 8 No 3 Article: Floating Figures at the Right... MF misnaming gf files weird behaviour with \newtheorem in Latex 2.09 TeXfonts for Indian languages dvi -> postscript DVI Previewer for MS-WINDOWS TeX in Foreign Languages. (TeXhax Digest V87 #101) TeXhax Digest V87 #101 LaTeX Notes, Part II (Re: TeXhax Digest V87 #102) DVI to Xerox print server? Re: Pound signs and digest V87 #102 mode_def needed for sun QMS laser printer blank pages / overwriting : the solution. Stanford fonts dvi2* drivers 118 dpi fonts (TeXhax Digest V87 #102) ---------------------------------------------------------------------- Date: Tue, 15 Dec 87 14:01:23 EST From: kgk%cs.brown.edu@RELAY.CS.NET Subject: tpic I would like to know about ``tpic'', and where to obtain it. Thanks. Keiji Kanazawa ------------------------------ Date: Tue, 15 Dec 87 12:31:05 pst From: lamport@src.dec.com (Leslie Lamport) Subject: LaTeX Notes, Part I (Re: TeXhax Digest V87 #101) Thomas H. Cormen writes I believe I've encountered a bug in LaTeX captions. I have defined a macro to make a vector given two endpoints: $\vector{a}{b}$ outputs ab with a rightarrow above them. \renewcommand{\vector}[2]{\overrightarrow{#1#2}} This macro works fine in running text, but in a caption, the basline is too high. In particular, if I use $\vector{p}{q}$, the very bottom of the p and q are at the baseline of the rest of the caption line, instead of the baseline of the p and q being at the baseline of the rest of the line. Indeed, he has discovered a problem with the figure environment. Until he gets a corrected version of LaTeX, he can get his \vector command to work by changing it to \renewcommand{\vector}[2]{{\boxmaxdepth=1000pt\relax\overrightarrow{#1#2}}} For the TeXperts, LaTeX currently sets \boxmaxdepth to 0pt inside a figure environment, which screws up the Plain TeX \overrightarrow command (and probably other commands as well). The trick of locally resetting \boxmaxdepth to a large value will correct any other such problems that may arise. Marc Auslander writes: Using LaTeX, I've placed the \pagestyle right after \documentstyle. It still doesn't effect the first page. What now? I'm not sure what he's trying to do, but page 162 of the LaTeX manual informs us that The right head information is always null for the first page of a document. If this is a problem, generate a blank first page with the titlepage environment. Leslie Lamport ------------------------------ Date: Tue, 15 Dec 1987 14:20:15 CST From: "Thomas J. Reid" Subject: TUGboat Vol 8 No 3 Article: Floating Figures at the Right... A co-worker of mine, Don Berryman, pointed out an error in my article "Floating Figures at the Right --- and --- Some Random Text for Testing" which appears in TUGboat Volume 8, Number 3. The error appears on page 316. The last three lines of the second column read: {\everypar={\relax}\setbox0=\lastbox \parskip=\wd0 \parskip=0pt \par \penalty-10001 \leavevmode}% The lines should be: {\everypar={\relax}\setbox0=\lastbox \parindent=\wd0 \parskip=0pt \par \penalty-10001 \leavevmode}% The first \parskip on the middle line should be a \parindent. This is an error on my part and I apologize for it. The purpose of these lines is to "back out" of a paragraph that has just started, enter vertical mode, issue a penalty, then restart the para- graph that was just backed-out. The control sequences appear within a macro that is called by \everypar, so the \everypar is temporarily reset to prevent recursion. These commands work as follows: (1) {\everypar{\relax} starts a new group and sets up a dummy \everypar. (2) \setbox0=\lastbox takes the glue added at the start of the para- graph and saves it in \box0. It also removes the glue from the list. This is needed since the \everypar token list is called after horizontal mode is entered and any paragraph indentation glue is added to the list. (3) \parindent=\wd0 sets the width of the last box (the paragraph indent glue) as the \parindent glue. (4) \parskip=0pt suppresses adding more vertical glue to a the vertical list when the paragraph is restarted; the vertical glue from the aborted paragraph has not been removed from the list. (5) \par causes vertical mode to be re-entered. (6) \penalty-10001 causes the \output routine to be invoked with the special penalty value. (7) \leavevmode causes \parskip glue to be added to the vertical list, horizontal mode to be re-entered, and \parindent glue to be added to the horizontal list. These use the temporarily altered \parskip and \parindent values. (8) }% ends the temporary group and resets \everypar, \parskip and \parindent. Why do all this work? Most of the time, this code simply sets \parindent equal to what it is already set to. However, in cases where a paragraph was started with \noindent, the paragraph indent glue will have zero width while \parindent may be non-zero. This technique makes the re- started paragraph the same as the original one. Impact of Bug: In the erroneous code, \parindent never gets changed. Hence, the problem will not show up unless a paragraph was started with \noindent. The paragraph will instead be restarted with the normal \parindent glue. This affects any paragraphs started with \noindent from the paragraph after the one where the figure started printing through the first para- graph after the figure has been covered. Thomas Reid ------------------------------ From: Julian Bradfield Date: Wed, 16 Dec 87 12:03:21 GMT Subject: MF misnaming gf files I have a problem with Metafont on Sun3s (the standard distribution, as far as I know). I have some programs for Computer Modern Old English letters (edh, thorn, yogh), and when I generate the italic version (which I call cmoeti10), MF puts the gf output onto a file with a 2602gf extension, no matter what mode or magnification I specified. The contents of the gf file are correct; it's just the name that is wrong. This does not happen with cmoer10 ! Before I start looking at it in detail, has anybody else had this problem (and solved it!)? The METAFONT in question was Version 1.2 for Berkeley UNIX. Julian Bradfield. jcb@lfcs.ed.ac.uk (arpa) ------------------------------ Date: Wed, 16 Dec 87 11:54:47 PST From: mcgeer%unicorn.Berkeley.EDU@cad.Berkeley.EDU (Rick McGeer) Subject: weird behaviour with \newtheorem in Latex 2.09 File (a) blows up into an infinite loop. File (b), which is file (a) less one line (there is a second \newtheorem in file (a)) works fine. The files are included here, (a) first. file (a): ------------------------------------------------------------------------------ \newtheorem{theorem}{Theorem}[section] \newtheorem{corollary}{Corollary}[section] \documentstyle[11pt]{article} \title{Initial Notes on Hazards and Critical Paths} \author{Patrick McGeer,\\ Department of Electrical Engineering and Computer Sciences,\\ University of California, Berkeley,\\ Berkeley, CA, 94720.} \begin{document} \maketitle \section {Introduction} Will it blow up here? \end {document} ------------------------------------------------------------------------------- and now,...file (b) _______________________________________________________________________________ \newtheorem{theorem}{Theorem}[section] \documentstyle[11pt]{article} \title{Initial Notes on Hazards and Critical Paths} \author{Patrick McGeer,\\ Department of Electrical Engineering and Computer Sciences,\\ University of California, Berkeley,\\ Berkeley, CA, 94720.} \begin{document} \maketitle \section {Introduction} Will it blow up here? \end {document} ------------------------------------------------------------------------------- Now, if I change the line \newtheorem{corollary}{Corollary}[section] to \newtheorem{corollary}{Corollary} in file (a), it also works fine. Can anyone tell me what's going on here? According to the LaTex book (pg 58) this is perfectly OK. I am running LaTex 2.09 on a GPX II under Ultrix. -- Rick ------------------------------ Date: Wed, 16 Dec 87 16:40:29 EST From: Ajit Ranade Subject: TeXfonts for Indian languages Does anyone have info on how to get hold of TeXfonts for Indian languages (eg. Hindi, Bengali, Tamil Marathi etc.) to be used on an IBM-PC? I know that they exist (probably in many places), but have no clue where to look for them? Info about public domain status, licensing, preview program and such things, will also be greatly appreciated. ------------------------------ Date: WED DEC 16, 1987 18.50.42 From: F3WL%MCGILLC.BITNET@CUNYVM.CUNY.EDU PiCTeX The PiCTeX package (TeXhax v87#101) seems very interesting. Unfortunately, some of us cannot reach either june.cs.washington.edu or a.cs.uiuc.edu, so no luck of getting that. Does any one else has an idea of other sites that have that? Or better yet, in a listserver from which we can fetch from BITNET. Wai Hung Leung BITNET: F3WL@MCGILLC ARPA: F3WL%MCGILLC.BITNET@CUNYVM.CUNY.EDU ------------------------------ Date: Wednesday, 16 Dec 87 11:27:59 EST From: thompson (Ross Thompson) @ redbud.psy.cmu.edu Subject: dvi -> postscript Sorry if this is common knowledge, but is there a public domain div2ps program around anywhere? Thanks, Ross ------------------------------ Date: Thu, 17 Dec 87 11:14 EST From: "James A. O'Brien (203) 432-4382" Subject: DVI Previewer for MS-WINDOWS a) Does there exist a DVI previewer that runs under the MS-WINDOWS system on MS-DOS machines? If there is (pref. cheap or pd) I would appreciate knowing about it. I recognize that this animal, if it exists, will probably be SLOOOOOW - however it would be nicely device-independent (I think). Please send answers directly to me (if it's not too much trouble) as well as to the net. Jim O'Brien Department of Chemical Engineering Yale University P.O.Box 2159 Yale Station New Haven, CT 06520, U.S.A. +1 203 432 4382 Return Addresses (both equivalent): OBRIEN@YALEVMS BITnet OBRIEN@VENUS.YCC.YALE.EDU internet ------------------------------ Date: Thu, 17 Dec 87 08:27:20 PST From: mackay@june.cs.washington.edu (Pierre MacKay) Subject: TeX in Foreign Languages. (TeXhax Digest V87 #101) For some time now there has been a forlorn directory on the Unix TeX distribution tape known as ./babel. At present it contains only Portuguese, but we hope to get Turkish on there soon. The idea is to collect as many hyphenation tables ant other macros as possible. Only those solutions that work without a modification of TeX can be included in this effort, however. Any system which provides for substitution of hyphenation tables during a run is unfortunately not exactly TeX. Pierre A. MacKay TUG Site Coordinator for Unix-flavored TeX ------------------------------ Date: Thu, 17 Dec 87 08:38:32 PST From: mackay@june.cs.washington.edu (Pierre MacKay) Subject: TeXhax Digest V87 #101 I must publicly apologize to Michael Wichura for a misunderstanding about PiCTeX. The macros are definitely available for general release, but the manual was not supposed to go out. I appeal to those of you who have the manual to order a copy from the University of Chicago, (on the principle of share-ware) If you like PiCTeX, let UChicago recoup some of the costs of developing it. The PiCTeX tar file will be restored to ~ftp/pub on june, but without the Manual. A publication of the final version of the manual is planned for early next year. Once again. This was my error, and the University of Chicago should not be allowed to suffer for it. the email address is wichura@galton.uchicago.edu Pierre MacKay ------------------------------ Date: Thu, 17 Dec 87 09:58:13 pst From: lamport@src.dec.com (Leslie Lamport) Subject: LaTeX Notes, Part II (Re: TeXhax Digest V87 #102) gary%right.mcs@omnigate.clarkson.edu at writes: What kind of box is returned by the verbatim environment? I am interested in centering the result of verbatim, and nothing seems to have much effect. \begin{center} \begin{verbatim} Try a few lines \end{verbatim} \end{center} puts the output at the current margin. Is there any hope? The verbatim environment is a paragraph-making environment (see section C.5 of the manual), so it doesn't produce any kind of box. The enclosing center environment has no effect on it, just as it has no effect on an enclosed itemize environment. If you want to put a verbatim environment in a box, so it can be centered, try putting it in a minipage environment. Leslie Lamport ------------------------------ Date: Thu, 17 Dec 87 22:19:42 EST From: ingr!b11!guy@uunet.UU.NET Subject: DVI to Xerox print server? I have a Xerox 8040 Series printer on a Xerox print server. I have attempted to write a DVI to Interpress converter for it, but the Interpress software on the server is a very limited implementation. I tried defining fonts in the Interpress master, but I either need Interpress functions that are not supported or I run out of stack space, depending on the way I try to do it. I have not figured out any way to define font files and place them permanently on the server for later reference, either. I would be satisfied with .tfm files for the server's fonts, but I really would like to move the CMR fonts to the server. Does anyone have information or advice (or program source) which would help me print TeX output on my Xerox print server? Guy Streeter ...uunet!ingr!b11!guy (UUCP) ingr!b11!guy@uunet.uu.net (ARPANET) ------------------------------ Date: 18-DEC-1987 07:19:30 GMT From: ABBOTTP%aston.ac.uk@NSS.Cs.Ucl.AC.UK Subject: Re: Pound signs and digest V87 #102 Peter Flynn is suggesting that the upright pound sign is available and the italic one should not be used. Since the file is standard issue and I checked on score before sending the query what about all the other sites using `standard' files. This applies to those who use LaTeX as a tool only for production and are not interested in the inner workings. I would like to see you correction incorporated in the file at score (I have not yet fetched it again to check). Thanks Peter anyway for a solution. Peter ------------------------------ Date: Fri, 18 Dec 87 10:04 GMT From: Ralph Subject: mode_def needed for sun I want to make some metafont fonts for a Sun workstation - can anyone please advise me on what to use for the mode_def ? If we are being really subtle, its a sun 3/160c. Thanks, Ralph ------------------------------ Date: 18-DEC-1987 15:52:49 GMT +00:00 From: MACALLSTR%vax1.physics.oxford.ac.uk@NSS.Cs.Ucl.AC.UK Subject: QMS laser printer blank pages / overwriting : the solution. Fix for the blank page / overwriting problem on QMS Lasergrafix 800 printers. ----------------------------------------------------------------------------- Symptoms. --------- Blanks pages between print files. Print files 'sticking' in the laser printer and then coming out overprinted by the the following file. Cause. ------ The QUIC command ^DC0001 prints one copy of the current page without ending the QUIC pass. Some software may use the ^, command to force printing and may not encounter the problem. Our DVI-to-QMS driver uses the ^DC0001 command. A QUIC page could be 99999 dots X 99999 dots in size (about 27feetX27feet!). If the ^DC0001 command occurs less than 5391 dots down the page, the page isn't printed; GE 5391 dots down the page and all is fine. I don't know the significance of the 5391/5392 number but that is the breakpoint for the problem. If you have FORM-feed enabled on your printer line, you will force output of the 'stuck' pages but you then get an extra blank page after every normal file. Solution. --------- Instead of sending ^DC0001 to the laser printer, force the vertical pointer down the page to some number larger than 5391. I use ^J90000^DC0001 in place of ^DC0001 . Set your printer line to NO FORM-feeds to eliminate unwanted blank pages. In summary, for a VAX/VMS system, do the following. Use ^J90000^DC001 in place of ^DC0001 . ( in your DVI-to-QMS driver ) SET TERM/PERM/NOFORM for the terminal line to which the laser printer is attached. Easy, isn't it?! John. ------------------------------ Date: Fri, 18 Dec 87 11:37:08 EDT From: Dimitri Vulis Subject: Stanford fonts >Date: Thu, 10 Dec 87 12:31:50 ZONE >From: E. Assouline >Subject: KSTFONTS.......TEX TEXT VIA > Did anyone hear of KSTfonts for TeX, are the Metafont sources >for these fonts available? Here is the info you requested: >--- >Date: Fri, 27 Nov 87 13:55:31 EDT >From: Dimitri Vulis >Subject: Stanford fonts >To: DTP%PLAID@SUN.COM > >I have the following question/request/query: >I have come across a _part_ of a 60-page document entitled: > >--- begin quote --- >Stanford Artificial Intelligence Laboratory >Operating Note 74 May 1976 > FIND a FONT > Les Earnest > >Abstract: Typefaces currently available for the Xerox Graphics Printer >are presented herein, warts and all. Procedures for creating, modifying >and stealing fonts are discussed. >--- end quote --- > >The person who left the document here is no longer around. >The document lists a number of very cute fonts, e.g. Boecklin, >Chess pieces, Tengvar (Tolkien), etc. The samples look very nice. >One section describes the file format (similar to TeX PXL). Another >section supposedly described the way to get the files thru FTP, >but this particular section is missing. >..etc.. ----------------------- >Date: Thu, 10 Dec 87 09:44 EDT >From: "guthery%asc@sdr.slb.com" >Subject: RE: FIND a FONT >To: CLDLV%nasagiss.bitnet@RELAY.CS.NET > >The Austin Code Works (512-258-0785) sells 130 of the fonts in Find a Font >in both the original XGP format and in TeX format. If you have a bitmap >to PostScript or TeX to PostScript weggie you're in. The fonts cost $30 >and you get a complete copy of the paper. Claimer: my wife runs the >Austin Code Works. > >Cheers, Scott ---------------------------- >Date: 11 Dec 87 09:12:00 MST >From: >Subject: stanford fonts >To: "cldlv%nasagiss.bitnet" > >What you describe on netnews sounds like the old stanford font distribution >for the original TeX. I certainly remember the Tengwar, chess pieces, >and Bocklin (though that wasn't what it was called). There was an >American sign language font too. Bunches of weird stuff. All in >old TeX (presumably old Metafont, which exists only in SAIL). >...ETC... ------------------ The Austin Code Works' ad indeed says: "KST Fonts (13,200 characters in 139 mixed fonts: specify TeX or bitmap format)...$30". Based on the above, I strongly doubt that the new MF code exists; it is likely that the MF78 programs still exist 'somewhere'. However, in my expreience, trying to translate MF78 into MF82 is sometimes harder than writing a new program for the picture (probably most people disagree with this). DV Acknowledge-To: MAIL ------------------------------ Date: Fri, 18 Dec 87 11:30:55 EST From: Alan Cabrera <13501ADC%MSU.BITNET@CUNYVM.CUNY.EDU> Subject: dvi2* drivers I would like to get my hands on some drivers that will convert DVI files for a PostScript machine and also a Xerox 2700. PASCAL or WEB source is preferable but C is acceptable. I would be able to aquire the drivers via anonymous FTP, or E-mail via BITNET or Internet. Mailing tapes can be done if necessary. Drivers from any VM/CMS sites would make my life easier. Thanks. Alan D. Cabrera Bitnet: 13501adc@msu Computer Laboratory Internet: cabreraad@clvax1.cl.msu.edu Room 400d Computer Center Michigan State University East Lansing, Michigan 48824-1042 ------------------------------ Date: Fri, 18 Dec 87 09:38:01 PST From: mackay@june.cs.washington.edu (Pierre MacKay) Subject: 118 dpi fonts (TeXhax Digest V87 #102) The fonts sent out with the Unix distribution were originally made for the Bitggraph, not for the Sun, but the settings have never been changed. Moreover, no attention has been paid to the fact that all black on white fonts for a CRT are by definition write-white fonts. One of the long range projects ought to be to collect improvements, but the advent of X makes me want to wait and see (If I can ever get X to work) whether X will take over the previewer business. 118 dpi is kind of unfortunate anyway. It was a compromise worked out by trial and error on the BitGraph, and quickly imported onto other devices. 120dpi, which would fit with things like the Imagewriter, and other dot matrix printers would have been better. METAFONT does cough at low resolutions. Paths cross and get tangled and other undesirable things happen. I suspect that is inevitable. There is a script in the present distribution that produces LaTeX fonts (gf118.makefile) that nobly attempts to produce things like cmbx7.84gf and doesn't do too badly, but some characters just fail. More serious than that extreme case is that cmtt7.118gf doesn't work (I have the log files, but not on this machine). I don't have any notion about a cure at this end of the scale. I am hesitant to put too much into it, because I wonder if 118 dpi is an unfortunate fossil, and ought perhaps to be eliminated. Here is the mode_def that was used for the UnixTeX fonts. mode_def bitgraph = % bitgraph and sun screens (the original settings) proofing:=0; % no, we're not making proofs fontmaking:=1; % yes, we are making a font tracingtitles:=0; % no, don't show titles in the log pixels_per_inch:=118; % for the oversize *.590pxl fonts blacker:=.55; % make pens a bit blacker fillin:=.1; % and compensate for diagonal fillin o_correction:=.3; % but don't overshoot much enddef; Here is an indication of why 120 dpi would have been preferable. mode_def imagewriter = % proofing:=0; % no, we're not making proofs fontmaking:=1; % yes, we are making a font tracingtitles:=0; % no, don't show titles in the log pixels_per_inch:=144; % Image-writer is twice screen density blacker:=0; % assume same as Epson owing to ink flow fillin:=0; % and same for diagonal fillin o_correction:=.2; % don't overshoot much enddef; ------------------------------ %%% %%% subscriptions, address changes to: texhax-request@score.stanford.edu %%% please send a valid arpanet address!! %%% %%% submissions to: texhax@score.stanford.edu %%% %%% BITNET redistribution: TEX-L@MARIST.BITNET (list server) %%% %%%\bye %%% ------------------------------ End of TeXhax Digest ************************** -------