From - Mon Dec 1 17:32:16 2003 Return-Path: Received: from sunshine.math.utah.edu (sunshine.math.utah.edu [128.110.198.2]) by nef.ens.fr (8.12.10/1.01.28121999) with ESMTP id hB1GWBaW075693 for ; Mon, 1 Dec 2003 17:32:11 +0100 (CET) Received: from plot79.math.utah.edu (IDENT:3f7/pwBZtEFWp81nTAviNuxTBRduUV3i@plot79.math.utah.edu [128.110.198.20]) by sunshine.math.utah.edu (8.12.10/8.12.10) with ESMTP id hB1GVuhx009899; Mon, 1 Dec 2003 09:31:57 -0700 (MST) Received: from plot79.math.utah.edu (IDENT:/GyzjVFquERmgtdxK+CUD9NLw9YnYQjH@localhost [127.0.0.1]) by plot79.math.utah.edu (8.12.10/8.12.10) with ESMTP id hB1GVueV019463; Mon, 1 Dec 2003 09:31:56 -0700 (MST) Received: (from beebe@localhost) by plot79.math.utah.edu (8.12.10/8.12.10/Submit) id hB1GVs55019461; Mon, 1 Dec 2003 09:31:55 -0700 (MST) Date: Mon, 1 Dec 2003 09:31:55 -0700 (MST) From: "Nelson H. F. Beebe" To: metafont@ens.fr, tex-fonts@math.utah.edu Cc: beebe@math.utah.edu X-US-Mail: "Center for Scientific Computing, Department of Mathematics, 110 LCB, University of Utah, 155 S 1400 E RM 233, Salt Lake City, UT 84112-0090, USA" X-Telephone: +1 801 581 5254 X-FAX: +1 801 585 1640, +1 801 581 4148 X-URL: http://www.math.utah.edu/~beebe Subject: News: GNU package for 3-D drawing with Metapost Message-ID: This interesting item appears on the info-gnu mailing list: >> ... >> From: Laurence Finston >> Subject: 3DLDF: 3D drawing with MetaPost output >> Date: Sun, 30 Nov 2003 16:58:22 +0100 (MET) >> To: info-gnu@gnu.org >> >> 3DLDF is a new GNU package for three-dimensional drawing with MetaPost >> output. It is written in C++ using CWEB and includes the _3DLDF User and >> Reference Manual_ written using Texinfo, and documented program >> code in book form (_3DLDF: The Program_), which was created using cweave. >> >> 3DLDF is intended, among other things, to provide a convenient way of >> creating 3D graphics for inclusion in TeX files. >> >> 3DLDF 1.1.2 (the current version) is currently available for >> downloading from: >> http://wwwuser.gwdg.de/~lfinsto1 >> >> The _3DLDF User and Reference Manual_ and _3DLDF: The Program_ are >> downloadable >> separately. >> >> 3DLDF will be made available at >> http://ftp.gnu.org/gnu/3DLDF >> as soon as possible. >> The official 3DLDF website will be >> http://www.gnu.org/software/3DLDF >> but it hasn't been set up yet. >> >> The following mailing lists should be made available soon: >> >> info-3DLDF@gnu.org For announcements. >> help-3DLDF@gnu.org For users to help one another >> bug-3DLDF@gnu.org For bug reports. >> >> Until they are available, please contact the author (see below). >> >> 3DLDF does not currently include an input routine, so that user code must >> be written in C++, compiled, and linked with the rest of the program. >> The files README and INSTALL, and the _3DLDF User and Reference Manual_ >> contain detailed instructions for doing this. >> >> Laurence Finston is the maintainer and, so far, the only developer of >> 3DLDF. >> He can be contacted at: >> lfinsto1@gwdg.de >> ... ------------------------------------------------------------------------------- - Nelson H. F. Beebe Tel: +1 801 581 5254 - - Center for Scientific Computing FAX: +1 801 581 4148 - - University of Utah Internet e-mail: beebe@math.utah.edu - - Department of Mathematics, 110 LCB beebe@acm.org beebe@computer.org - - 155 S 1400 E RM 233 - - Salt Lake City, UT 84112-0090, USA URL: http://www.math.utah.edu/~beebe - ------------------------------------------------------------------------------- From - Tue Dec 9 21:46:34 2003 Return-Path: Received: from smtp-out.tiscali.be (spoolo1.tiscali.be [62.235.13.210]) by nef.ens.fr (8.12.10/1.01.28121999) with ESMTP id hB9KkVaW028699 for ; Tue, 9 Dec 2003 21:46:31 +0100 (CET) Received: from [62.235.116.57] (helo=adelmotte.largevision.com) by smtp-out.tiscali.be with esmtp (Tiscali.be http://www.tiscali.be) id 1ATokX-0005Qv-Iq for ; Tue, 09 Dec 2003 21:46:31 +0100 Message-Id: <5.0.2.1.2.20031209213419.009efd70@pop3.pro.tiscali.be> X-Sender: tpm17153@pop3.pro.tiscali.be (Unverified) X-Mailer: QUALCOMM Windows Eudora Version 5.0.2 Date: Tue, 09 Dec 2003 21:38:26 +0100 To: liste metafont From: Alain Delmotte Subject: Accent Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1"; format=flowed Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by nef.ens.fr id hB9KkVaW028699 Hello, In the Metafont/(La)TeX system, where are the accent defined? They are created by metafont, but are they normal characters with a normal width? And if yes how are they correctly placed atop of any other character? What about the vertical position, it should not be the same for a lowercase letter and for a capital? Is it possible to imagine to create a character without width so that it overlap the preceding or following characters? Thanks, Alain Delmotte avenue du Marathon, 6 B 1348 Louvain-la-Neuve Belgique tél./fax 00 (32) 10 45 44 44 tél. privé 00 (32) 10 45 11 92 From - Tue Dec 9 22:15:41 2003 Return-Path: Received: from smarthost1.mail.easynet.fr (smarthost1.mail.easynet.fr [212.180.1.68]) by nef.ens.fr (8.12.10/1.01.28121999) with ESMTP id hB9LFcaW041250 for ; Tue, 9 Dec 2003 22:15:38 +0100 (CET) Received: from [212.180.15.41] (helo=easyconnect.fr) by smarthost1.mail.easynet.fr with esmtp (Exim 3.35 #1 (Debian)) id 1ATpCk-0003tE-00; Tue, 09 Dec 2003 22:15:38 +0100 Message-ID: <3FD63B79.4070402@easyconnect.fr> Date: Tue, 09 Dec 2003 22:15:37 +0100 From: Bernard Desgraupes User-Agent: Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.4) Gecko/20030624 Netscape/7.1 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Alain Delmotte CC: liste metafont Subject: Re: [metafont] Accent References: <5.0.2.1.2.20031209213419.009efd70@pop3.pro.tiscali.be> In-Reply-To: <5.0.2.1.2.20031209213419.009efd70@pop3.pro.tiscali.be> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit > > In the Metafont/(La)TeX system, where are the accent defined? In the fonts. > > They are created by metafont, but are they normal characters with > a normal width? And if yes how are they correctly placed atop of any > other character? It depends on the encoding you are using, on the accent itself etc. In OT1 encoding (128 code values) accents are separate characters and are placed on the top of characters thanks to adhoc macros defined in the specifictation of the encoding (file .enc read by TeX) In T1 (256 code values) you have both accents and accented letters. There are other encodings, other scripts (cyrillic, greek etc.), so that one can find quite a wide variety of situations. > > What about the vertical position, it should not be the same for a > lowercase letter and for a capital? > > Is it possible to imagine to create a character without width so that it > overlap the preceding or following characters? Yes it is easy to do in Metafont. > From - Wed Dec 10 11:46:32 2003 Return-Path: Received: from ifi.informatik.uni-stuttgart.de (ifi.informatik.uni-stuttgart.de [129.69.211.1]) by nef.ens.fr (8.12.10/1.01.28121999) with ESMTP id hBAAkSaW080554 for ; Wed, 10 Dec 2003 11:46:28 +0100 (CET) Received: from isar.informatik.uni-stuttgart.de (isar [129.69.215.232]) by ifi.informatik.uni-stuttgart.de (8.9.3p2/2.2) with ESMTP id LAA08688; Wed, 10 Dec 2003 11:46:27 +0100 (MET) Received: (from raichle@localhost) by isar.informatik.uni-stuttgart.de (8.9.3p2/2.2) id LAA01406; Wed, 10 Dec 2003 11:46:26 +0100 (MET) From: Bernd Raichle MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-ID: <16342.63874.140712.413114@is.informatik.uni-stuttgart.de> Date: Wed, 10 Dec 2003 11:46:26 +0100 To: liste metafont Subject: Re: [metafont] Accent In-Reply-To: <5.0.2.1.2.20031209213419.009efd70@pop3.pro.tiscali.be> References: <5.0.2.1.2.20031209213419.009efd70@pop3.pro.tiscali.be> X-Mailer: VM 7.17 under Emacs 21.2.2 On Tuesday, 9 December 2003 21:38:26 +0100, Alain Delmotte writes: > Hello, > > In the Metafont/(La)TeX system, where are the accent defined? In the font and/or in TeX, the program. Depending on the font encoding, fonts can contain accented characters, e.g. an e accent grave in an T1 encoded font. These characters are created by Metafont and are "normal" glyphs. For fonts (e.g. OT1 encoded) which do not have these accented characters you need an accent glyph and a "base" character glyph to built them. > They are created by metafont, but are they normal characters with > a normal width? And if yes how are they correctly placed atop of any > other character? TeX provides the \accent primitive to put an accent glyph over another glyph. (It is possible that these two characters are from different font). > What about the vertical position, it should not be the same for a > lowercase letter and for a capital? The accent glyph is a "normal" glyph character in the font. It should be created in such a way that its placement is correct for an non-slanted font and an lowercase character without ascenders, i.e., the height of the base characters is equal to 1ex (e.g. "x" :-). If the height of the base character differs from 1ex, \accent raises (or lowers) the accent glyph accordingly by height_basechar - 1ex. The horizontal placement is done in such a way that the accent glyph is centered over the base glyph. Thus both characters should have a proper width. If a font is slanted, the horizontal placement is fixed accordingly using the distance by which the accent is raised (or lowered). All these computations are done by TeX's \accent primitive and can not be changed. If you want to change the accent placement, you have to do it by changing the glyph dimensions ... or by fooling TeX about the height of 1ex (this trick is used by "german.sty"/Babel's "german" to lower the umlaut accents a little bit). > Is it possible to imagine to create a character without width so that it > overlap the preceding or following characters? Yes, this is possible. TeX only uses the glyph dimensions to place them. TeX does not know where a (Meta)font designer puts the ink on the paper. Btw. overlapping can be done using TeX's macros, too (cf. the cedille macro \c in plain.tex). Best wishes, -bernd _____________________________________________________________________ Bernd Raichle "Le langage est source Autor des `german.sty' (aktuell: v2.5e) de malentendus" DE-TeX-FAQ: http://www.dante.de/faq/de-tex-faq/ (A. de Saint-Exupery) From - Wed Dec 10 18:05:07 2003 Return-Path: Received: from mailhost.uark.edu (mail.uark.edu [130.184.5.66]) by nef.ens.fr (8.12.10/1.01.28121999) with ESMTP id hBAH54aW099792 for ; Wed, 10 Dec 2003 18:05:04 +0100 (CET) Received: from comp.uark.edu (comp.uark.edu [130.184.5.197]) by mailhost.uark.edu (iPlanet Messaging Server 5.2 HotFix 1.21 (built Sep 8 2003)) with ESMTPS id <0HPO00GLVUY5AV@mailhost.uark.edu> for metafont@ens.fr; Wed, 10 Dec 2003 10:54:05 -0600 (CST) Date: Wed, 10 Dec 2003 10:54:05 -0600 (CST) From: "Daniel H. Luecking" Subject: Re: [metafont] Accent In-reply-to: <3FD63B79.4070402@easyconnect.fr> To: liste metafont Message-id: MIME-version: 1.0 Content-type: TEXT/PLAIN; charset=US-ASCII Content-transfer-encoding: 7BIT References: <5.0.2.1.2.20031209213419.009efd70@pop3.pro.tiscali.be> <3FD63B79.4070402@easyconnect.fr> On Tue, 9 Dec 2003, Bernard Desgraupes wrote: > > > > > > In the Metafont/(La)TeX system, where are the accent defined? > > In the fonts. > > > > > They are created by metafont, but are they normal characters with > > a normal width? And if yes how are they correctly placed atop of any > > other character? > > It depends on the encoding you are using, on the accent itself etc. > In OT1 encoding (128 code values) accents are separate characters and > are placed on the top of characters thanks to adhoc macros defined in > the specifictation of the encoding (file .enc read by TeX) As far as LaTeX is concerned, no .enc files are read. LaTeX's fontenc package reads *enc.def. Plain TeX has fixed encodings, based on the actual positions in the cm fonts. > In T1 (256 code values) you have both accents and accented letters. > There are other encodings, other scripts (cyrillic, greek etc.), so that > one can find quite a wide variety of situations. > > > > > What about the vertical position, it should not be the same for a > > lowercase letter and for a capital? > > > > Is it possible to imagine to create a character without width so that it > > overlap the preceding or following characters? > > Yes it is easy to do in Metafont. In the cm fonts, the only place (that I know of) where this is used is for \not (the slash making \not= work). TeX has \accent and \mathaccent primatives that behave as Bernd Raichle describes in another post. Regards, -- Dan Luecking Dept. of Mathematical Sciences luecking@uark.edu University of Arkansas http://comp.uark.edu/~luecking/ Fayetteville, AR 72101 From - Wed Dec 10 18:35:16 2003 Return-Path: Received: from smarthost1.mail.easynet.fr (smarthost1.mail.easynet.fr [212.180.1.68]) by nef.ens.fr (8.12.10/1.01.28121999) with ESMTP id hBAHZFaW009636 for ; Wed, 10 Dec 2003 18:35:15 +0100 (CET) Received: from [212.180.15.203] (helo=easyconnect.fr) by smarthost1.mail.easynet.fr with esmtp (Exim 3.35 #1 (Debian)) id 1AU8Ex-00086b-00 for ; Wed, 10 Dec 2003 18:35:11 +0100 Message-ID: <3FD7594F.8040607@easyconnect.fr> Date: Wed, 10 Dec 2003 18:35:11 +0100 From: Bernard Desgraupes User-Agent: Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.4) Gecko/20030624 Netscape/7.1 X-Accept-Language: en-us, en MIME-Version: 1.0 To: liste metafont Subject: Re: [metafont] Accent References: <5.0.2.1.2.20031209213419.009efd70@pop3.pro.tiscali.be> <3FD63B79.4070402@easyconnect.fr> In-Reply-To: Content-Type: multipart/alternative; boundary="------------090204060409040101000409" This is a multi-part message in MIME format. --------------090204060409040101000409 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Daniel H. Luecking wrote: >As far as LaTeX is concerned, no .enc files are read. LaTeX's fontenc >package reads *enc.def. Plain TeX has fixed encodings, based on the >actual positions in the cm fonts. > > This is what I wanted to say. I was thinking of ot1enc.def et al. but my fingers typed too fast. Thanks for the correction. >>>Is it possible to imagine to create a character without width so that it >>>overlap the preceding or following characters? >>> >>> >>Yes it is easy to do in Metafont. >> >> > >In the cm fonts, the only place (that I know of) where this is used >is for \not (the slash making \not= work). TeX has \accent and >\mathaccent primatives that behave as Bernd Raichle describes in another >post. > > I think the original message was asking about creating a char with zero width with Metafont, not hacking cm. Cheers Bernard --------------090204060409040101000409 Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit

Daniel H. Luecking wrote:


As far as LaTeX is concerned, no .enc files are read. LaTeX's fontenc
package reads *enc.def.  Plain TeX has fixed encodings, based on the
actual positions in the cm fonts.

This is what I wanted to say. I was thinking of ot1enc.def et al. but my fingers typed too fast. Thanks for the correction.

Is it possible to imagine to create a character without width so that it
overlap the preceding or following characters?
      
Yes it is easy to do in Metafont.
    

In the cm fonts, the only place (that I know of) where this is used
is for \not (the slash making \not= work). TeX has \accent and
\mathaccent primatives that behave as Bernd Raichle describes in another
post.
I think the original message was asking about creating a char with zero width with Metafont, not hacking cm.

Cheers

Bernard
--------------090204060409040101000409-- From - Wed Dec 10 21:56:02 2003 Return-Path: Received: from smtp-out.tiscali.be (spoolo2.tiscali.be [62.235.13.211]) by nef.ens.fr (8.12.10/1.01.28121999) with ESMTP id hBAKtuaW070608 for ; Wed, 10 Dec 2003 21:55:57 +0100 (CET) Received: from [62.235.116.5] (helo=adelmotte.largevision.com) by smtp-out.tiscali.be with esmtp (Tiscali.be http://www.tiscali.be) id 1AUBMx-0000TA-SU for ; Wed, 10 Dec 2003 21:55:41 +0100 Message-Id: <5.0.2.1.2.20031210214525.00a51230@pop3.pro.tiscali.be> X-Sender: tpm17153@pop3.pro.tiscali.be (Unverified) X-Mailer: QUALCOMM Windows Eudora Version 5.0.2 Date: Wed, 10 Dec 2003 21:53:35 +0100 To: liste metafont From: Alain Delmotte Subject: Re: [metafont] Accent In-Reply-To: <16342.63874.140712.413114@is.informatik.uni-stuttgart.de> References: <5.0.2.1.2.20031209213419.009efd70@pop3.pro.tiscali.be> <5.0.2.1.2.20031209213419.009efd70@pop3.pro.tiscali.be> Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1"; format=flowed Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by nef.ens.fr id hBAKtuaW070608 Thanks to everybody for the answers given. A 11:46 10/12/2003 +0100, vous avez écrit : >For fonts (e.g. OT1 encoded) which do not have these accented >characters you need an accent glyph and a "base" character glyph to >built them. > > > They are created by metafont, but are they normal characters with > > a normal width? And if yes how are they correctly placed atop of any > > other character? > >TeX provides the \accent primitive to put an accent glyph over another >glyph. (It is possible that these two characters are from different >font). So, if I understand well for the OT1: the accent is a glyph on its own and it should have a small height as it will go above another character (if it were of the height of the font, it would go far above the character!!) I also think of putting the accents I'll need at the position of the normal accents in cm fonts to benefit of the existing macros \'x, ... Correct? (this is for a thai font and the accent are in fact the voyelles put above the other letters, sometimes in two levels - unless somebody knows of a full operational system, I went to CTAN and downloaded the usl-thai.mf file). Thanks again, Alain Delmotte avenue du Marathon, 6 B 1348 Louvain-la-Neuve Belgique tél./fax 00 (32) 10 45 44 44 tél. privé 00 (32) 10 45 11 92 From - Wed Dec 10 22:13:06 2003 Return-Path: Received: from mailhost.uark.edu (mailhost.uark.edu [130.184.5.66]) by nef.ens.fr (8.12.10/1.01.28121999) with ESMTP id hBALD0aW075910 for ; Wed, 10 Dec 2003 22:13:03 +0100 (CET) Received: from comp.uark.edu (comp.uark.edu [130.184.5.197]) by mailhost.uark.edu (iPlanet Messaging Server 5.2 HotFix 1.21 (built Sep 8 2003)) with ESMTPS id <0HPP00EE669S9N@mailhost.uark.edu> for metafont@ens.fr; Wed, 10 Dec 2003 14:58:41 -0600 (CST) Date: Wed, 10 Dec 2003 14:58:40 -0600 (CST) From: "Daniel H. Luecking" Subject: Re: [metafont] Accent In-reply-to: <3FD7594F.8040607@easyconnect.fr> To: liste metafont Message-id: MIME-version: 1.0 Content-type: TEXT/PLAIN; charset=US-ASCII Content-transfer-encoding: 7BIT References: <5.0.2.1.2.20031209213419.009efd70@pop3.pro.tiscali.be> <3FD63B79.4070402@easyconnect.fr> <3FD7594F.8040607@easyconnect.fr> On Wed, 10 Dec 2003, Bernard Desgraupes wrote: > > >>>Is it possible to imagine to create a character without width so that it > >>>overlap the preceding or following characters? > >>> > >>Yes it is easy to do in Metafont. > >> > >In the cm fonts, the only place (that I know of) where this is used > >is for \not (the slash making \not= work). TeX has \accent and > >\mathaccent primatives that behave as Bernd Raichle describes in another > >post. > > > I think the original message was asking about creating a char with zero > width with Metafont, not hacking cm. I only meant to expand on your words by giving an example of a font created with Metafont that actually used the idea. And to suggest that it is not very common as the rather standard cm fonts don't employ it much, and certainly not for accents. If a zero width character were used to create an accent, the glyph would need a different position (in its 0-width box) for each possible height and width of characters that needed the accent. It is probably for this reason that Knuth gave TeX the \accent primitive. Regards, -- Dan Luecking Dept. of Mathematical Sciences luecking@uark.edu University of Arkansas http://comp.uark.edu/~luecking/ Fayetteville, AR 72101 From - Wed Dec 10 22:25:43 2003 Return-Path: Received: from smarthost2.mail.easynet.fr (smarthost2bis.mail.easynet.fr [212.180.1.37]) by nef.ens.fr (8.12.10/1.01.28121999) with ESMTP id hBALPcaW081565 for ; Wed, 10 Dec 2003 22:25:38 +0100 (CET) Received: from [212.180.15.203] (helo=easyconnect.fr) by smarthost2.mail.easynet.fr with esmtp (Exim 3.35 #1 (Debian)) id 1AUBpy-0004vn-00 for ; Wed, 10 Dec 2003 22:25:38 +0100 Message-ID: <3FD78F52.9070702@easyconnect.fr> Date: Wed, 10 Dec 2003 22:25:38 +0100 From: Bernard Desgraupes User-Agent: Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.4) Gecko/20030624 Netscape/7.1 X-Accept-Language: en-us, en MIME-Version: 1.0 To: liste metafont Subject: Re: [metafont] Accent References: <5.0.2.1.2.20031209213419.009efd70@pop3.pro.tiscali.be> <3FD63B79.4070402@easyconnect.fr> <3FD7594F.8040607@easyconnect.fr> In-Reply-To: Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Daniel H. Luecking wrote: >I only meant to expand on your words by giving an example of a font >created with Metafont that actually used the idea. And to suggest that >it is not very common as the rather standard cm fonts don't employ it >much, and certainly not for accents. > >If a zero width character were used to create an accent, the glyph >would need a different position (in its 0-width box) for each possible >height and width of characters that needed the accent. It is probably >for this reason that Knuth gave TeX the \accent primitive. > > Right. I fully agree. Bernard > > > From - Thu Dec 11 14:21:06 2003 Return-Path: Received: from ifi.informatik.uni-stuttgart.de (ifi.informatik.uni-stuttgart.de [129.69.211.1]) by nef.ens.fr (8.12.10/1.01.28121999) with ESMTP id hBBDL2aW049699 for ; Thu, 11 Dec 2003 14:21:02 +0100 (CET) Received: from isar.informatik.uni-stuttgart.de (isar [129.69.215.232]) by ifi.informatik.uni-stuttgart.de (8.9.3p2/2.2) with ESMTP id OAA27949; Thu, 11 Dec 2003 14:20:59 +0100 (MET) Received: (from raichle@localhost) by isar.informatik.uni-stuttgart.de (8.9.3p2/2.2) id OAA03071; Thu, 11 Dec 2003 14:20:58 +0100 (MET) From: Bernd Raichle MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Message-ID: <16344.28474.125448.213433@is.informatik.uni-stuttgart.de> Date: Thu, 11 Dec 2003 14:20:58 +0100 To: liste metafont Subject: Re: [metafont] Accent In-Reply-To: <5.0.2.1.2.20031210214525.00a51230@pop3.pro.tiscali.be> References: <5.0.2.1.2.20031209213419.009efd70@pop3.pro.tiscali.be> <5.0.2.1.2.20031210214525.00a51230@pop3.pro.tiscali.be> X-Mailer: VM 7.17 under Emacs 21.2.2 Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by nef.ens.fr id hBBDL2aW049699 On Wednesday, 10 December 2003 21:53:35 +0100, Alain Delmotte writes: > Thanks to everybody for the answers given. > > A 11:46 10/12/2003 +0100, vous avez écrit : > >For fonts (e.g. OT1 encoded) which do not have these accented > >characters you need an accent glyph and a "base" character glyph to > >built them. > > > > > They are created by metafont, but are they normal characters with > > > a normal width? And if yes how are they correctly placed atop of any > > > other character? > > > >TeX provides the \accent primitive to put an accent glyph over another > >glyph. (It is possible that these two characters are from different > >font). > > So, if I understand well for the OT1: the accent is a glyph on its own > and it should have a small height as it will go above another character ^^^^^^^^^^^^^^ > (if it were of the height of the font, it would go far above the character!!) No, it should be positioned in such a way that its place is ok when overprinting an "x" with this accent. This means that its height is usually greater than 1ex. Please have a look into "cmr10" (or Volume E of "Computers&Typesetting" :-) as an example how it is done by Knuth. > I also think of putting the accents I'll need at the position of the normal > accents in cm fonts to benefit of the existing macros \'x, ... Correct? Yes. > (this is for a thai font and the accent are in fact the voyelles put above > the other letters, sometimes in two levels - unless somebody knows > of a full operational system, I went to CTAN and downloaded the > usl-thai.mf file). The \accent primitive does not support more than one accents... but there are tricks how to achieve it (there are some packages for double accent support on CTAN). Best wishes, -bernd From - Thu Dec 11 15:45:52 2003 Return-Path: Received: from ccs.carleton.ca (driftwood.ccs.carleton.ca [134.117.1.49]) by nef.ens.fr (8.12.10/1.01.28121999) with ESMTP id hBBEjkaW076311 for ; Thu, 11 Dec 2003 15:45:46 +0100 (CET) Received: from rideau.YP.nobel (daemon@rideau [134.117.1.17]) by ccs.carleton.ca (8.12.10/8.12.10) with ESMTP id hBBEjkM8007809 for ; Thu, 11 Dec 2003 09:45:46 -0500 (EST) From: Christina Thiele Received: (from cthiele@localhost) by rideau.YP.nobel (8.11.6+Sun/8.9.1) id hBBEjjc22619 for metafont@ens.fr; Thu, 11 Dec 2003 09:45:45 -0500 (EST) Message-Id: <200312111445.hBBEjjc22619@rideau.YP.nobel> Subject: Re: [metafont] Accent To: metafont@ens.fr Date: Thu, 11 Dec 2003 09:45:45 -0500 (EST) In-Reply-To: from "Bernd Raichle" at Dec 11, 2003 02:20:58 PM X-Mailer: ELM [version 2.5 PL5] MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Bernd Raichle writes: > > ... > > (this is for a thai font and the accent are in fact the voyelles put above > > the other letters, sometimes in two levels - unless somebody knows > > of a full operational system, I went to CTAN and downloaded the > > usl-thai.mf file). > > The \accent primitive does not support more than one accents... but > there are tricks how to achieve it (there are some packages for double > accent support on CTAN). > > Best wishes, > -bernd > One of these is a very old macro we devised for CJL back in the 1980s (first published in _TeXniques_ No. 5, 1987 TUG conf.). I believe it's also available to wsuipa users and probably also tipa (Fukui?). But you can copy this one -- still works for me ;-) \def\diatop[#1|#2]{\leavevmode{\setbox1=\hbox{{#1{}}}% \setbox2=\hbox{{#2{}}}% \dimen0=\ifdim\wd1>\wd2\wd1\else\wd2\fi% \dimen1=\ht2\advance\dimen1by-1ex% \setbox1=\hbox to1\dimen0{\hss#1\hss}% \rlap{\raise1\dimen1\box1}% \hbox to1\dimen0{\hss#2\hss}}}% Usage is then: \diatop[<1st accent>|<2nd accent + char.>] Occasionally, the results aren't good ;-) so the first strategy is to just reverse the order of the accents. I think that Phil Taylor once told me why this was so but I've of course forgotten ;-) -- not because it was Phil ;-) but because I'm not very good at understanding TeX's inner workings ;-)) I also recall that Yannis used \diatop is some of his SE Asian language stuff -- it's on one of his TUGboat articles, probably a proceedings issue, perhaps early 90s? -- and I pointed out that I'd found that adding \leavevmode (first line of the def) was a useful refinement (again, I was looking at the results, not the reasons ... ). And I do think that on at least one occasion I've embedded a \diatop within a \diatop to get a third diacritic -- probably would've been a combination of two on top, and one below ... Anyone else with \diatop experiences or workarounds for problem combinations? Ch. From - Fri Dec 12 20:25:14 2003 Return-Path: Received: from mailer.gwdg.de (mailer.gwdg.de [134.76.10.26]) by nef.ens.fr (8.12.10/1.01.28121999) with ESMTP id hBCJPBaW045952 for ; Fri, 12 Dec 2003 20:25:11 +0100 (CET) Received: from gwdu71.gwdg.de ([134.76.8.21]) by mailer.gwdg.de with esmtp (Exim 4.20) id 1AUsuQ-0008Ru-Rl for metafont@ens.fr; Fri, 12 Dec 2003 20:25:06 +0100 Received: from localhost by gwdu71.gwdg.de (8.11.1/1.1.29.3/16Jun03-0924AM) id hBCJP6r0000296506; Fri, 12 Dec 2003 20:25:06 +0100 (MET) X-Authentication-Warning: gwdu71.gwdg.de: lfinsto1 owned process doing -bs Date: Fri, 12 Dec 2003 20:25:06 +0100 (MET) From: Laurence Finston To: liste metafont Subject: 3DLDF 1.1.4 Release (fwd) Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Spam-Level: / X-Spam-Report: Content analysis: 0.0 points, 6.0 required X-Virus-Scanned: (clean) by exiscan+sophie 3DLDF 1.1.4 has been released. 3DLDF 1.1.4 is a GNU package for three-dimensional drawing with MetaPost output. ********************************** Dear Metafont users, It would be fairly easy to get 3DLDF to output Metafont code as well as MetaPost code. I've been thinking about a possible application, namely fonts consisting of characters that are perspective projections of three-dimensional characters. This effect is sometimes used in movie and TV titles, e.g., Monty Python. I believe it would be most appropriate for display fonts, so ``meta-ness'' would not necessarily be required. Would anybody be interested in this sort of font? If so, I'll give a higher priority to making 3DLDF output Metafont code. It would be even nicer if someone was interested in working on it with me. For example, if someone else coded a set of letters, I could write the code for making Metafont output possible. (The letters need to be coded in C++, using the functions of 3DLDF.) Another thing that might be of interest to Metafont and MetaPost users is the input routine. As you may know, 3DLDF doesn't have one yet. I'm about to start work on it. I intend that 3DLDF should behave like Metafont from a user's point of view. The messages it prints to standard output and its error handling should be similar to Metafont's. Ideally, any valid Metafont or Metapost input should be valid 3DLDF input. However, this is a non-trivial task, given the complexity of Metafont's syntax. What makes it easier, or even possible, is that _The METAFONTbook_ explains the syntax very clearly, using tables. It shouldn't be necessary to refer to _METAFONT: The Program_. Another difficulty is that Metafont and 3DLDF are quite different. Knuth programmed Metafont's scanning and parsing routines from scratch, and I plan to use Bison and Flex. However, I believe it should be possible with a reasonable amount of effort, and that it is definitely worth doing. If anyone's interested in helping with either of these tasks, or with 3DLDF in general, please contact me. Sincerely, Laurence Finston 3DLDF maintainer ************************************************************** 3DLDF can be downloaded from http://ftp.gnu.org/gnu/3dldf New in version 1.1.4: The Texinfo manual can now be generated in Info and HTML formats. Bug fix in the shell script 3DLDF-1.1.4/CWEB/tsthdweb that's used by 3DLDF-1.1.4/CWEB/Makefile to control rebuilding. Previously, files were sometimes retangled and recompiled unnecessarily. I believe I've fixed the problem now, but it's tricky, so I'll have to keep an eye on it. Added the function Rectangle::is_rectangular() and made mediate() a member function in class Point. MAX_REAL is now really the second largest float, but only on 32-bit, little-endian architectures. On other architectures, 3dldf will exit with an error message. I plan to make this work on other architectures soon. Laurence Finston 3DLDF maintainer lfinsto1@gwdg.de http://wwwuser.gwdg.de/~lfinsto1 From - Fri Dec 12 22:17:42 2003 Return-Path: Received: from smarthost1.mail.easynet.fr (smarthost1.mail.easynet.fr [212.180.1.68]) by nef.ens.fr (8.12.10/1.01.28121999) with ESMTP id hBCLHaaW087871 for ; Fri, 12 Dec 2003 22:17:36 +0100 (CET) Received: from ipdynerx74-107.clients.easynet.fr ([195.114.74.107] helo=easyconnect.fr) by smarthost1.mail.easynet.fr with esmtp (Exim 3.35 #1 (Debian)) id 1AUufG-0000nW-00 for ; Fri, 12 Dec 2003 22:17:34 +0100 Message-ID: <3FDA306D.7060302@easyconnect.fr> Date: Fri, 12 Dec 2003 22:17:33 +0100 From: Bernard Desgraupes User-Agent: Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.4) Gecko/20030624 Netscape/7.1 X-Accept-Language: en-us, en MIME-Version: 1.0 To: liste metafont Subject: Re: [metafont] 3DLDF 1.1.4 Release (fwd) References: In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 8bit Laurence Finston wrote: >Would anybody be interested in this sort of font? If so, I'll give a >higher priority to making 3DLDF output Metafont code. > Great idea ! >It would be even >nicer if someone was interested in working on it with me. For example, >if someone else coded a set of letters, I could write the >code for making Metafont output possible. (The letters need to be >coded in C++, using the functions of 3DLDF.) > >Another thing that might be of interest to Metafont and MetaPost users is >the input routine. As you may know, 3DLDF doesn't have one yet. I'm about >to start work on it. I intend that 3DLDF should behave like Metafont from >a user's point of view. > Do you mean so that people can write 3DLDF fonts with source code files "à la Mf" ? >The messages it prints to standard output and its >error handling should be similar to Metafont's. Ideally, any valid >Metafont or Metapost input should be valid 3DLDF input. However, this is a >non-trivial task, given the complexity of Metafont's syntax. >What makes it easier, or even possible, is that _The METAFONTbook_ >explains the syntax very clearly, using tables. It shouldn't be necessary >to refer to _METAFONT: The Program_. >Another difficulty is that >Metafont and 3DLDF are quite different. Knuth programmed Metafont's >scanning and parsing routines from scratch, and I plan to use Bison and >Flex. However, I believe it should be possible with a reasonable amount >of effort, and that it is definitely worth doing. > > Maybe you can have a look at the mfpic package which spews Mf code too (it is written in TeX but it is interesting to see what kind of files it outputs and what kind of macros it defines for this purpose). Chhers Bernard From - Sat Dec 13 11:12:58 2003 Return-Path: Received: from mailer.gwdg.de (mailer.gwdg.de [134.76.10.26]) by nef.ens.fr (8.12.10/1.01.28121999) with ESMTP id hBDACuaW044719 for ; Sat, 13 Dec 2003 11:12:56 +0100 (CET) Received: from gwdu71.gwdg.de ([134.76.8.21]) by mailer.gwdg.de with esmtp (Exim 4.20) id 1AV6lV-0006AJ-To for metafont@ens.fr; Sat, 13 Dec 2003 11:12:49 +0100 Received: from localhost by gwdu71.gwdg.de (8.11.1/1.1.29.3/16Jun03-0924AM) id hBDACn40000315130; Sat, 13 Dec 2003 11:12:49 +0100 (MET) X-Authentication-Warning: gwdu71.gwdg.de: lfinsto1 owned process doing -bs Date: Sat, 13 Dec 2003 11:12:49 +0100 (MET) From: Laurence Finston To: liste metafont Subject: Re: [metafont] 3DLDF 1.1.4 Release (fwd) In-Reply-To: <3FDA306D.7060302@easyconnect.fr> Message-ID: References: <3FDA306D.7060302@easyconnect.fr> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=X-UNKNOWN X-Spam-Level: / X-Spam-Report: Content analysis: 0.0 points, 6.0 required X-Virus-Scanned: (clean) by exiscan+sophie Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from QUOTED-PRINTABLE to 8bit by nef.ens.fr id hBDACuaW044719 On Fri, 12 Dec 2003, Bernard Desgraupes wrote: > >Another thing that might be of interest to Metafont and MetaPost users is > >the input routine. As you may know, 3DLDF doesn't have one yet. I'm about > >to start work on it. I intend that 3DLDF should behave like Metafont from > >a user's point of view. > > > Do you mean so that people can write 3DLDF fonts with source code files > "à la Mf" ? > It would be more precise to say that I want people to be able to use 3DLDF to generate Metafont fonts. I would like 3DLDF to behave in a way familiar to users of TeX and Metafont: 1. It should be callable from the command line of a shell, taking input either from standard input or a file. Making this work is more important than writing a GUI. 2. When called interactively, `**' should appear, and typing `\relax' should cause `*' to appear. The meaning of `**' of `*' should be the same as in MF. 3. Errors should cause similar messages to be printed to standard error, and context should appear in the same way. It should be possible to delete and insert tokens as in TeX and MF. This may actually be a bit tricky to code. Per default Flex returns information about the line and column numbers of tokens, but I don't think I have direct access to the buffer where it keeps its input. In particular, I think it discards the lines it's already processed, so I'll have to maintain a second input stream in order to display the context. 4. Ideally, when valid Metafont or Metapost is fed to 3DLDF, it should output identical MF or MP code. This in itself will probably serve no useful purpose, but will make it possible to use MF or MP code as the basis of three-dimensional drawings or fonts. It will also be a useful test of whether I manage to implement MF/MP's syntax in 3DLDF. My purpose is to make 3DLDF fit in well with TeX, MF, MP, and the tools associated with them, dvips, kpathsea, etc. > Maybe you can have a look at the mfpic package which spews Mf code too > (it is written in TeX but it is interesting to see what kind of files it > outputs and what kind of macros it defines for this purpose). Thanks for the suggestion. Laurence Finston 3DLDF maintainer From - Sun Dec 14 20:31:06 2003 Return-Path: Received: from mailhost.uark.edu (mailhost.uark.edu [130.184.5.66]) by nef.ens.fr (8.12.10/1.01.28121999) with ESMTP id hBEJV2aW001981 for ; Sun, 14 Dec 2003 20:31:02 +0100 (CET) Received: from mathfolk.uark.edu ([130.184.197.23]) by mailhost.uark.edu (iPlanet Messaging Server 5.2 HotFix 1.21 (built Sep 8 2003)) with ESMTPSA id <0HPW003PHGKFGG@mailhost.uark.edu> for metafont@ens.fr; Sun, 14 Dec 2003 13:24:15 -0600 (CST) Date: Sun, 14 Dec 2003 13:22:13 -0600 From: Dan Luecking Subject: Re: [metafont] 3DLDF 1.1.4 Release (fwd) In-reply-to: <3FDA306D.7060302@easyconnect.fr> X-Sender: luecking@mail.uark.edu To: liste metafont Message-id: <5.2.1.1.0.20031214131052.00b48ad0@mail.uark.edu> MIME-version: 1.0 X-Mailer: QUALCOMM Windows Eudora Version 5.2.1 Content-type: text/plain; charset=us-ascii; format=flowed Content-transfer-encoding: 7BIT References: At 10:17 PM 12/12/2003 +0100, you wrote: >Laurence Finston wrote: > >>The messages it prints to standard output and its >>error handling should be similar to Metafont's. Ideally, any valid >>Metafont or Metapost input should be valid 3DLDF input. However, this is a >>non-trivial task, given the complexity of Metafont's syntax. >>What makes it easier, or even possible, is that _The METAFONTbook_ >>explains the syntax very clearly, using tables. It shouldn't be necessary >>to refer to _METAFONT: The Program_. >>Another difficulty is that >>Metafont and 3DLDF are quite different. Knuth programmed Metafont's >>scanning and parsing routines from scratch, and I plan to use Bison and >>Flex. However, I believe it should be possible with a reasonable amount >>of effort, and that it is definitely worth doing. >> >Maybe you can have a look at the mfpic package which spews Mf code too (it >is written in TeX but it is interesting to see what kind of files it >outputs and what kind of macros it defines for this purpose). Mfpic's methods are not particularly revealing: Apart from handling optional arguments, its commands simply write a MF (or MP) command that is defined in an auxiliary file grafbase.mf (or .mp) which must be input by the file that mfpic writes. For example, \circle{(0,0),2} simply writes "circle((0,0),2)" and grafbase contains: vardef circle(expr p, r)= fullcircle scaled 2r shifted p enddef Mfpic handles raw MF (or MP) code by a command that writes its contents unchanged. I don't see why 3DLDF couldn't do the equivalent. Dan Daniel H. Luecking Department of Mathematical Sciences University of Arkansas From - Sun Dec 14 21:00:07 2003 Return-Path: Received: from mailhost.uark.edu (mail.uark.edu [130.184.5.66]) by nef.ens.fr (8.12.10/1.01.28121999) with ESMTP id hBEK02aW009794 for ; Sun, 14 Dec 2003 21:00:05 +0100 (CET) Received: from mathfolk.uark.edu ([130.184.197.23]) by mailhost.uark.edu (iPlanet Messaging Server 5.2 HotFix 1.21 (built Sep 8 2003)) with ESMTPSA id <0HPW003CJHHHGG@mailhost.uark.edu> for metafont@ens.fr; Sun, 14 Dec 2003 13:44:06 -0600 (CST) Date: Sun, 14 Dec 2003 13:42:04 -0600 From: Dan Luecking Subject: Re: [metafont] 3DLDF 1.1.4 Release (fwd) In-reply-to: X-Sender: luecking@mail.uark.edu To: liste metafont Message-id: <5.2.1.1.0.20031214132228.00b609e8@mail.uark.edu> MIME-version: 1.0 X-Mailer: QUALCOMM Windows Eudora Version 5.2.1 Content-type: text/plain; charset=us-ascii; format=flowed Content-transfer-encoding: 7BIT References: <3FDA306D.7060302@easyconnect.fr> <3FDA306D.7060302@easyconnect.fr> At 11:12 AM 12/13/2003 +0100, Laurence Finston wrote: >I would like 3DLDF to behave in a way familiar to users of TeX >and Metafont: > >1. It should be callable from the command line of a shell, taking input > either from standard input or a file. Making this > work is more important than writing a GUI. This would be good. In fact, I wouldn't put a GUI on my list of priorities at all. >2. When called interactively, `**' should appear, and typing `\relax' > should cause `*' to appear. The meaning of `**' of `*' should be the > same as in MF. > >3. Errors should cause similar messages to be printed to standard error, > and context should appear in the same way. It should be possible to > delete and insert tokens as in TeX and MF. TeX and MF are uniformly disliked for their error messages. I would urge a more parsable style: An error message should be clearly designated as such, contain the file name and line of the error, and the error message. However the MF/TeX method of showing the actual code where the error was detected, and allowing interactive insertion of tokens would be nice. >4. Ideally, when valid Metafont or Metapost is fed to 3DLDF, it should > output identical MF or MP code. This in itself will probably serve no > useful purpose, Actually, it is a very useful thing. No matter how one designs a high level language, an escape to the lower level always seems to be needed. LaTeX style writer frequently resort to TeX primitives and C programmers often embed assembly code. > but will make it possible to use MF or MP code as > the basis of three-dimensional drawings or fonts. It will also be a > useful test of whether I manage to implement MF/MP's syntax in > 3DLDF. Do you really need to implement the syntax? If your program outputs MF code, what prevents you from simply having a command that instructs 3DLDF to pass along some text verbatim? > > Maybe you can have a look at the mfpic package which spews Mf code too > > (it is written in TeX but it is interesting to see what kind of files it > > outputs and what kind of macros it defines for this purpose). > >Thanks for the suggestion. I wouldn't think mfpic would provide any deep insight. It is hardly more than a 1-1 translation: TeX marcros write similar MF macros. The accompanying definitions of the MF macros do all the drawing. (And I know, a large part of them were written by me following the lead of the previous maintainers.) Dan Daniel H. Luecking Department of Mathematical Sciences University of Arkansas From - Mon Dec 15 13:39:17 2003 Return-Path: Received: from mailer.gwdg.de (mailer.gwdg.de [134.76.10.26]) by nef.ens.fr (8.12.10/1.01.28121999) with ESMTP id hBFCdEaW022176 for ; Mon, 15 Dec 2003 13:39:14 +0100 (CET) Received: from gwdu71.gwdg.de ([134.76.8.21]) by mailer.gwdg.de with esmtp (Exim 4.20) id 1AVs0H-0003Ac-Tp for metafont@ens.fr; Mon, 15 Dec 2003 13:39:13 +0100 Received: from localhost by gwdu71.gwdg.de (8.11.1/1.1.29.3/16Jun03-0924AM) id hBFCdDp0000380997; Mon, 15 Dec 2003 13:39:13 +0100 (MET) X-Authentication-Warning: gwdu71.gwdg.de: lfinsto1 owned process doing -bs Date: Mon, 15 Dec 2003 13:39:11 +0100 (MET) From: Laurence Finston To: liste metafont Subject: Re: [metafont] 3DLDF 1.1.4 Release (fwd) In-Reply-To: <5.2.1.1.0.20031214132228.00b609e8@mail.uark.edu> Message-ID: References: <3FDA306D.7060302@easyconnect.fr> <3FDA306D.7060302@easyconnect.fr> <5.2.1.1.0.20031214132228.00b609e8@mail.uark.edu> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Spam-Level: / X-Spam-Report: Content analysis: 0.0 points, 6.0 required X-Virus-Scanned: (clean) by exiscan+sophie On Sun, 14 Dec 2003, Dan Luecking wrote: > TeX and MF are uniformly disliked for their error messages. I would urge a > more parsable style: One reason I want 3DLDF to act like Metafont is so that it will seem familiar to users of TeX, MF, and MP. They shouldn't have to learn a whole new set of conventions. > An error message should be clearly designated as such, > contain the file name and line of the error, and the error message. > I agree completely, I'll do that. > At 11:12 AM 12/13/2003 +0100, Laurence Finston wrote: > > > but will make it possible to use MF or MP code as > > the basis of three-dimensional drawings or fonts. It will also be a > > useful test of whether I manage to implement MF/MP's syntax in > > 3DLDF. On Sun, 14 Dec 2003, Dan Luecking wrote: > > Do you really need to implement the syntax? If your program outputs MF code, > what prevents you from simply having a command that instructs 3DLDF to pass > along some text verbatim? > What I mean is that the syntax of the 3DLDF language should be similar to that of the Metafont language. It should be so similar that valid Metafont code will cause 3DLDF to output identical Metafont code _after processing it_. This is an example of input that 3DLDF should be able to parse: z0 = origin; z1 = (1cm, 2in, 3pt); z2 = (5, 7, 19.32pt); z3 = z2 shifted z1; draw z0 .. tension 1.5 and 2.. z1{curl0} ... z2 .. z3 --cycle; I haven't tested these values, so it might be appropriate for them to signal an error because of an invalid turning number. `zscaling' is clearly a problem, since it already has a meaning in Metafont. However, I refuse to call the z-axis something else just because of this. I could have called the axes `x_1', `x_2', and `x_3', but I decided to stick with the most familiar names. I may regret this if I decide to implement projections of hypercubes and Klein bottles (this isn't a joke). > On Sun, 14 Dec 2003, Dan Luecking wrote: > > I wouldn't think mfpic would provide any deep insight. It is hardly more > than a 1-1 translation: TeX marcros write similar MF macros. The > accompanying definitions of the MF macros do all the drawing. > (And I know, a large part of them were written by me following the lead of > the previous maintainers.) > 3DLDF already outputs Metapost code, so it shouldn't be too difficult to make it output Metafont code. I'll need `beginchar()' and `endchar()' functions in addition to the `beginfig()' and `endfig()' functions, and a means of switching between `Metafont_mode' and `Metapost_mode'. I will have to change the `output()' functions so that they behave properly, and make various other changes. The real difficulty is that 3DLDF currently does no scan converting. It operates on the assumption that the two-dimensional projection of a path will look okay as long as there are enough points on it. Paths consisting of straight segments are no problem. I have books that describe how to implement Bezier curves in three-dimensions, and I plan to start doing this soon. I need to find one thing out: If I project the control points of a 3D Bezier curve, is the 2D Bezier curve defined by the projected points the projection of the 3D curve? If it is, this means less work for me, but I suspect it isn't. Does anybody know of the top of his or her head? Once I get this to work, it will be possible to specify 2D Metafont/Metapost paths that represent true projections of the 3D 3DLDF Paths. Then, it would be possible to program a ``3D font'', as long as all paths either consist of straight segments only, or are specified using `.. controls p and q ..', where p and q are points. Metafont can output this information for paths specified in other ways, so this might be useful for the first attempt at making a 3D font. However, ultimately I would like to implement `curl' and `tension'. I know that John Hobby has published a paper about the algorithm that implements these features, but I haven't read it. I don't know whether he extends it to three-dimensions, or if I'll understand it if he does. I'm pretty sure that I will try to program a 3D font myself eventually; probably when I want to use this feature in a drawing. However, as you probably all know, programming even a simple font with no meta-ness is a lot of work. I've gotten to a point with 3DLDF that I can't do everything myself, at least, not as quickly as I'd like. Therefore, I'm actively looking for volunteers. I'm much more likely to implement support for 3D fonts sooner rather than later if I know that people are going to want to use it, and especially if someone else wanted to code the input for the font itself. Laurence Finston 3DLDF Maintainer Home page: http://wwwuser.gwdg.de/~lfinsto1/LDF.html email: lfinsto1@gwdg.de From - Mon Dec 15 16:44:46 2003 Return-Path: Received: from mailhost.uark.edu (mailhost.uark.edu [130.184.5.66]) by nef.ens.fr (8.12.10/1.01.28121999) with ESMTP id hBFFifaW084849 for ; Mon, 15 Dec 2003 16:44:41 +0100 (CET) Received: from comp.uark.edu (comp.uark.edu [130.184.5.197]) by mailhost.uark.edu (iPlanet Messaging Server 5.2 HotFix 1.21 (built Sep 8 2003)) with ESMTPS id <0HPY007300LONI@mailhost.uark.edu> for metafont@ens.fr; Mon, 15 Dec 2003 09:34:36 -0600 (CST) Date: Mon, 15 Dec 2003 09:34:36 -0600 (CST) From: "Daniel H. Luecking" Subject: Re: [metafont] 3DLDF 1.1.4 Release (fwd) In-reply-to: To: liste metafont Message-id: MIME-version: 1.0 Content-type: TEXT/PLAIN; charset=US-ASCII Content-transfer-encoding: 7BIT References: <3FDA306D.7060302@easyconnect.fr> <3FDA306D.7060302@easyconnect.fr> <5.2.1.1.0.20031214132228.00b609e8@mail.uark.edu> On Mon, 15 Dec 2003, Laurence Finston wrote: > > What I mean is that the syntax of the 3DLDF language should be similar > to that of the Metafont language. It should be so similar that valid > Metafont code will cause 3DLDF to output identical Metafont code _after > processing it_. > > This is an example of input that 3DLDF should be able to parse: > > z0 = origin; > z1 = (1cm, 2in, 3pt); > z2 = (5, 7, 19.32pt); > z3 = z2 shifted z1; > draw z0 .. tension 1.5 and 2.. z1{curl0} ... z2 .. z3 --cycle; > > I haven't tested these values, so it might be appropriate for them > to signal an error because of an invalid turning number. > > `zscaling' is clearly a problem, since it already has a meaning in > Metafont. However, I refuse to call the z-axis something else just because > of this. I could have called the axes `x_1', `x_2', and `x_3', but I > decided to stick with the most familiar names. I may regret this if I > decide to implement projections of hypercubes and Klein bottles (this > isn't a joke). This sounds like a resonable idea. How do you intend to distinguish. 3d points from color? It would seem that one has to make the languages (3DLDF vs MetaPost) pretty different in some respects. > The real difficulty is that 3DLDF currently does no scan converting. It > operates on the assumption that the two-dimensional projection of a path > will look okay as long as there are enough points on it. Paths consisting > of straight segments are no problem. > > I have books that describe how to implement Bezier curves in > three-dimensions, and I plan to start doing this soon. I need to find one > thing out: If I project the control points of a 3D Bezier curve, is the 2D > Bezier curve defined by the projected points the projection of the 3D > curve? If it is, this means less work for me, but I suspect it isn't. Does > anybody know of the top of his or her head? If the projection is a linear projection or an affine projection (linear+shift), then yes, the projection of the curve coincides with the Bezier with the projected control points. Not true for perspective projections. It is also not true that a standard path like A'..B'..C' (where A' is the projection of A, etc., and the control points are calculated by MF/MP) is the correct projection of A..B..C (where 3DLDF would have to calculate the controls in 3D). Not even with linear projections. A smooth curve can have a projection with cusps and A'..B'..C' will always be smooth. I've thought about a few of these things in trying to work up a 3D version of mfpic. > > Once I get this to work, it will be possible to specify 2D Metafont/Metapost > paths that represent true projections of the 3D 3DLDF Paths. Then, it > would be possible to program a ``3D font'', as long as all paths either > consist of straight segments only, or are specified using > `.. controls p and q ..', where p and q are points. > > However, ultimately I would like to implement `curl' and `tension'. I know > that John Hobby has published a paper about the algorithm that implements > these features, but I haven't read it. I don't know whether he extends it > to three-dimensions, or if I'll understand it if he does. The algorithms, as Knuth explains things in the MFbook, acts on direction angles, which don't have an obvious 3D counterpart. -- Dan Luecking Dept. of Mathematical Sciences luecking@uark.edu University of Arkansas http://comp.uark.edu/~luecking/ Fayetteville, AR 72101 From - Mon Dec 15 19:55:20 2003 Return-Path: Received: from mailer.gwdg.de (mailer.gwdg.de [134.76.10.26]) by nef.ens.fr (8.12.10/1.01.28121999) with ESMTP id hBFItEaW044231 for ; Mon, 15 Dec 2003 19:55:14 +0100 (CET) Received: from gwdu71.gwdg.de ([134.76.8.21]) by mailer.gwdg.de with esmtp (Exim 4.20) id 1AVxs9-0004V4-A0 for metafont@ens.fr; Mon, 15 Dec 2003 19:55:13 +0100 Received: from localhost by gwdu71.gwdg.de (8.11.1/1.1.29.3/16Jun03-0924AM) id hBFItAG0000394499; Mon, 15 Dec 2003 19:55:10 +0100 (MET) X-Authentication-Warning: gwdu71.gwdg.de: lfinsto1 owned process doing -bs Date: Mon, 15 Dec 2003 19:55:10 +0100 (MET) From: Laurence Finston To: liste metafont Subject: Re: [metafont] 3DLDF 1.1.4 Release (fwd) In-Reply-To: Message-ID: References: <3FDA306D.7060302@easyconnect.fr> <3FDA306D.7060302@easyconnect.fr> <5.2.1.1.0.20031214132228.00b609e8@mail.uark.edu> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Spam-Level: / X-Spam-Report: Content analysis: 0.0 points, 6.0 required X-Virus-Scanned: (clean) by exiscan+sophie On Mon, 15 Dec 2003, Daniel H. Luecking wrote: > > How do you intend to distinguish. > 3d points from color? It would seem that one has to make the languages > (3DLDF vs MetaPost) pretty different in some respects. > I don't understand your question. Inside 3DLDF, Color and Point are both classes. When I implement the scanner and parser, a declaration like `color c(255, 255, 255);' in the input file will cause an object of type Color to be allocated and stored in a data structure (probably a map), together with the string "c". `point p(1.5, 2, 33);' will be treated analogously. 3DLDF does not build on MP the way MP builds on MF. I'm only interested in simulating MF/MP's behavior in certain clearly defined areas. Clearly, the 3DLDF language will have to differ from the Metapost language in some respects. For example, in 3DLDF, `point' will play the role of `pair' in Metapost. `pair' will be equivalent to `point' and have three coordinates. This is necessary, in order for 3DLDF to handle Metapost code properly. I like this, because it appeals to my sense of the absurd. It may be a little unclear why I've said both `Point' and `point' above. `Point' is a data type in 3DLDF. 3DLDF is usable now, it just doesn't have an input routine. User code must be written in C++, and `Point' is one of the data types that are defined in 3DLDF. `point' will be the name of a data type in what I've been calling the ``3DLDF language''. It will be one of the tokens recognized by the scanner, and will cause the parser to perform appropriate actions. Once the scanner and parser are implemented, 3DLDF will then be a compiler in the same way that TeX, MF, and MP are, and it will make sense to speak of a ``3DLDF language''. I don't want to limit the syntax of the 3DLDF language to that of MF/MP's; I just want it to handle MF and MP code correctly. > > If the projection is a linear projection or an affine projection > (linear+shift), then yes, the projection of the curve coincides with > the Bezier with the projected control points. Not true for perspective > projections. It is also not true that a standard path like > A'..B'..C' > (where A' is the projection of A, etc., and the control points are > calculated by MF/MP) is the correct projection of A..B..C (where 3DLDF > would have to calculate the controls in 3D). Not even with linear > projections. A smooth curve can have a projection with cusps and > A'..B'..C' will always be smooth. > > Thanks for the information. It would have been too good to be true, the other way. Laurence Finston 3DLDF maintainer website: http://wwwuser.gwdg.de/~lfinsto1/LDF.html email: lfinsto1@gwdg.de From - Mon Dec 15 23:35:12 2003 Return-Path: Received: from mailhost.uark.edu (mail.uark.edu [130.184.5.66]) by nef.ens.fr (8.12.10/1.01.28121999) with ESMTP id hBFMZ9aW005568 for ; Mon, 15 Dec 2003 23:35:09 +0100 (CET) Received: from comp.uark.edu (comp.uark.edu [130.184.5.197]) by mailhost.uark.edu (iPlanet Messaging Server 5.2 HotFix 1.21 (built Sep 8 2003)) with ESMTPS id <0HPY00781K2KS2@mailhost.uark.edu> for metafont@ens.fr; Mon, 15 Dec 2003 16:35:09 -0600 (CST) Date: Mon, 15 Dec 2003 16:35:08 -0600 (CST) From: "Daniel H. Luecking" Subject: Re: [metafont] 3DLDF 1.1.4 Release (fwd) In-reply-to: To: liste metafont Message-id: MIME-version: 1.0 Content-type: TEXT/PLAIN; charset=US-ASCII Content-transfer-encoding: 7BIT References: <3FDA306D.7060302@easyconnect.fr> <3FDA306D.7060302@easyconnect.fr> <5.2.1.1.0.20031214132228.00b609e8@mail.uark.edu> On Mon, 15 Dec 2003, Laurence Finston wrote: > On Mon, 15 Dec 2003, Daniel H. Luecking wrote: > > > > > How do you intend to distinguish. > > 3d points from color? It would seem that one has to make the languages > > (3DLDF vs MetaPost) pretty different in some respects. > > > I don't understand your question. Inside 3DLDF, Color and Point are both > classes. When I implement the scanner and parser, a declaration like > `color c(255, 255, 255);' I only meant that valid MP code can contain color assignments, in which rgb triples appear. If, as you said before, you want to handle MP code properly, you have to either disallow certain constructs or change the way they are entered. The above, for example, is not how MP would declare a color variable. > 3DLDF does not build on MP the way MP builds on MF. > I'm only interested > in simulating MF/MP's behavior in certain clearly defined areas. And color isn't necessarily one of them. OK. > > Clearly, the 3DLDF language will have to differ from the Metapost > language in some respects. For example, in 3DLDF, `point' will play the > role of `pair' in Metapost. `pair' will be > equivalent to `point' and have three coordinates. This is necessary, in > order for 3DLDF to handle Metapost code properly. > I like this, because it appeals to my sense of the absurd. That's the sort of thing I meant. As to color, I just wondered if you were going to try to make code like the following work. color X; X := (.5,.6,.3); path P; P := (0,0,0)..(1,1,2)..(0,1,1); draw P withcolor ((1,1,1)-X); Triples could be parsed to mean colors where colors are needed and points where points are needed, much like TeX allows skips to be dimensions and dimensions to be numbers. Or one could use a different syntax for entering and using colors. -- Dan Luecking Dept. of Mathematical Sciences luecking@uark.edu University of Arkansas http://comp.uark.edu/~luecking/ Fayetteville, AR 72101 From - Tue Dec 16 13:02:06 2003 Return-Path: Received: from mailer.gwdg.de (mailer.gwdg.de [134.76.10.26]) by nef.ens.fr (8.12.10/1.01.28121999) with ESMTP id hBGC22aW024570 for ; Tue, 16 Dec 2003 13:02:04 +0100 (CET) Received: from gwdu71.gwdg.de ([134.76.8.21]) by mailer.gwdg.de with esmtp (Exim 4.20) id 1AWDtq-00032d-An for metafont@ens.fr; Tue, 16 Dec 2003 13:02:02 +0100 Received: from localhost by gwdu71.gwdg.de (8.11.1/1.1.29.3/16Jun03-0924AM) id hBGC21B0000425524; Tue, 16 Dec 2003 13:02:01 +0100 (MET) X-Authentication-Warning: gwdu71.gwdg.de: lfinsto1 owned process doing -bs Date: Tue, 16 Dec 2003 13:02:00 +0100 (MET) From: Laurence Finston To: liste metafont Subject: Re: [metafont] 3DLDF 1.1.4 Release (fwd) In-Reply-To: Message-ID: References: <3FDA306D.7060302@easyconnect.fr> <3FDA306D.7060302@easyconnect.fr> <5.2.1.1.0.20031214132228.00b609e8@mail.uark.edu> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Spam-Level: / X-Spam-Report: Content analysis: 0.0 points, 6.0 required X-Virus-Scanned: (clean) by exiscan+sophie On Mon, 15 Dec 2003, Daniel H. Luecking wrote: > On Mon, 15 Dec 2003, Laurence Finston wrote: > > > > > > `color c(255, 255, 255);' > > The above, for example, is not how > MP would declare a color variable. > Oops. > > That's the sort of thing I meant. As to color, I just wondered if > you were going to try to make code like the following work. > color X; > X := (.5,.6,.3); > path P; P := (0,0,0)..(1,1,2)..(0,1,1); > draw P withcolor ((1,1,1)-X); > Yes. The parser (or a function called by it) will allocate memory for a Color and store a pointer to it in a map with the string "X" as its key. When the parser reads `X:=(.5, .6, .3);', it will look in the map for "X". Then, it will call the function that represents the assignment operation. Since C++ allows overloading of functions, the proper function can be found at run-time. The same applies to Path. Laurence Finston 3DLDF Maintainer website: http://wwwuser.gwdg.de/~lfinsto1/LDF.html email: lfinsto1@gwdg.de