From - Sat Nov 10 17:38:59 2001 Return-Path: Received: from smtpzilla2.xs4all.nl (smtpzilla2.xs4all.nl [194.109.127.138]) by nef.ens.fr (8.10.1/1.01.28121999) with ESMTP id fAAGclq65930 for ; Sat, 10 Nov 2001 17:38:47 +0100 (CET) Received: from server-1.pragma-ade.nl (s340-isdn2507.dial.xs4all.nl [194.109.189.203]) by smtpzilla2.xs4all.nl (8.12.0/8.12.0) with ESMTP id fAAGcXwi013764; Sat, 10 Nov 2001 17:38:41 +0100 (CET) Received: from laptop-2.wxs.nl (laptop-1.pragma-ade.nl [200.1.1.26]) by server-1.pragma-ade.nl (8.11.2/8.9.3) with ESMTP id fAAEo5e09254; Sat, 10 Nov 2001 15:50:05 +0100 Message-Id: <5.1.0.14.1.20011109224505.02eec0e8@server-1> X-Sender: hagen@server-1 X-Mailer: QUALCOMM Windows Eudora Version 5.1 Date: Fri, 09 Nov 2001 22:48:15 +0100 To: metafont@ens.fr From: Hans Hagen Cc: Fabrice POPINEAU , taco@elvenkind.com In-Reply-To: <200110280244.f9S2ihM94458@coxeter.math.toronto.edu> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed X-Virus-Scanner: AMaVis 0.2.0-pre6 / Virus Scan X-Loop: metafont@nef.ens.fr X-Sequence: 493 Precedence: list Subject: [metafont] possible bug in metapost Hi, I just ran into a strange situation: every 62th "input" (or eof test) fails I used to have: def readfile (expr name) = if (readfrom (name) <> EOF) : scantokens("input " & name & " ") fi enddef ; which fails at read 62/124/etc The solution is: def readfile (expr name) = if (readfrom (name) <> EOF) : scantokens("input " & name & " ") elseif (readfrom (name) <> EOF) : scantokens("input " & name & " ") fi enddef ; This maybe windows/mpost specific, but somehow the magic 62 puzzles me. Hans ------------------------------------------------------------------------- Hans Hagen | PRAGMA ADE | pragma@wxs.nl Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: +31 (0)38 477 53 69 | fax: +31 (0)38 477 53 74 | www.pragma-ade.com ------------------------------------------------------------------------- From - Mon Nov 12 21:18:43 2001 Return-Path: Received: from moutvdom00.kundenserver.de (moutvdom00.kundenserver.de [195.20.224.149]) by nef.ens.fr (8.10.1/1.01.28121999) with ESMTP id fACKIYq97803 for ; Mon, 12 Nov 2001 21:18:34 +0100 (CET) Received: from [195.20.224.220] (helo=mrvdom04.kundenserver.de) by moutvdom00.kundenserver.de with esmtp (Exim 2.12 #2) id 163NXN-0001c8-00 for metafont@ens.fr; Mon, 12 Nov 2001 21:18:33 +0100 Received: from pd90104a9.dip0.t-ipconnect.de ([217.1.4.169] helo=1und1/pt1094-284) by mrvdom04.kundenserver.de with esmtp (Exim 2.12 #2) id 163NXM-0001w1-01 for metafont@ens.fr; Mon, 12 Nov 2001 21:18:33 +0100 Reply-To: From: "Wolfgang Werners-Lucchini" To: Date: Mon, 12 Nov 2001 20:52:09 +0100 X-Msmail-Priority: Normal X-Priority: 3 X-Mailer: Microsoft Internet Mail 4.70.1132 Mime-Version: 1.0 Content-Type: text/plain; charset=Default Content-Transfer-Encoding: 8bit Message-Id: X-Virus-Scanner: AMaVis 0.2.0-pre6 / Virus Scan X-Loop: metafont@nef.ens.fr X-Sequence: 494 Precedence: list Subject: [metafont] metapost, ghostscript and german umlaut Hallo! I would like to use mp and ghostscript (WITHOUT tex,latex and dvips). How can I print german umlaut (äöüßÄÖÜ). Thanks, Wolfgang From - Wed Nov 14 18:14:33 2001 Return-Path: Received: from coxeter.math.toronto.edu (coxeter.math.toronto.edu [128.100.68.3]) by nef.ens.fr (8.10.1/1.01.28121999) with ESMTP id fAEHELU12593 for ; Wed, 14 Nov 2001 18:14:21 +0100 (CET) Received: (from laurent@localhost) by coxeter.math.toronto.edu (8.11.0/8.11.0/UTMath 1.0) id fAEHEBh79334; Wed, 14 Nov 2001 12:14:11 -0500 Date: Wed, 14 Nov 2001 12:14:11 -0500 From: Larry Siebenmann Message-Id: <200111141714.fAEHEBh79334@coxeter.math.toronto.edu> To: metafont@ens.fr Subject: Re: [metafont] does MF draw NURBs ?? Cc: ??@math.toronto.edu, MF@math.toronto.edu, NURBs@math.toronto.edu, Re:, [metafont]@math.toronto.edu, does@math.toronto.edu, draw@math.toronto.edu X-Virus-Scanner: AMaVis 0.2.0-pre6 / Virus Scan X-Loop: metafont@nef.ens.fr X-Sequence: 495 Precedence: list Hi all, Recall that NURB means "non-uniform rational B-spline" (in the plane for us). As Nelson indicated, it is, at heart, just a projective polynomial parametrised plane curve. In other words there exist real polynomials p,q,r in the real parameter t such that the curve is expressible as: t |--> ( p(t)/r(t) , q(t)/r(t) ) in R^2 The term NURBS refers to a "control point" apparatus for deforming it --- one more complex then the the bezier-casteljau apparatus that Knuth & Hobbly use. Notice that this curve is the projection from the origin to the plane z=1 in R^3 of the polynomial space curve: t |--> ( p(t), q(t), r(t) ) in R^3 Santiago Muelas took up on my title "does MF draw NURBs?" and pointed out (if I am reading right) that with one interpretation of the title, yes, Metafont and and Metapost can at least calculate successively the points. Indeed, restricting to degree 3 (as MF does) and projecting to the x-y and y-z planes we have two cubic plane curves t |--> ( p(t), q(t) ) t |--> ( q(t), r(t) ) Recall that these are are both bezier with uniquely determined control points and so MF can give the R^3 coordinates of any point on the space curve, and then the centrally projected point on the plane z=1. That's a bit painful and I am sure Santiago's system has better projective capabilities when it comes to drawing. Of course it might be as fast to have MF turn the handle on the first (rational) formula! I wish Santiago luck in adding such capabilities. But there is serious question whether this would make Metafont/Metapost more useful. Lars Helstrom's question probes in that direction. Cheers Laurent S. From - Wed Nov 14 18:27:48 2001 Return-Path: Received: from coxeter.math.toronto.edu (coxeter.math.toronto.edu [128.100.68.3]) by nef.ens.fr (8.10.1/1.01.28121999) with ESMTP id fAEHRZU14521 for ; Wed, 14 Nov 2001 18:27:36 +0100 (CET) Received: (from laurent@localhost) by coxeter.math.toronto.edu (8.11.0/8.11.0/UTMath 1.0) id fAEHRYv93252 for metafont@ens.fr; Wed, 14 Nov 2001 12:27:34 -0500 Date: Wed, 14 Nov 2001 12:27:34 -0500 From: Larry Siebenmann Message-Id: <200111141727.fAEHRYv93252@coxeter.math.toronto.edu> To: metafont@ens.fr Subject: Re: [metafont] does MF draw NURBS ?? X-Virus-Scanner: AMaVis 0.2.0-pre6 / Virus Scan X-Loop: metafont@nef.ens.fr X-Sequence: 496 Precedence: list Hi all, Let me return to Lars Helstrom's quiery: LH> Metafont is usually drawing by moving an elliptic pen along a path > which is a cubic spline, thus blackening an area whose outline is > not a cubic spline. With Nelson's message in mind, it could however > be worth investigating whether the outline would be a NURBS. And if so of what degree? I expect the answer is *no* and *known* though I don't have it in hand. I wonder if it is answered in the newish book cited by Nelson Beebe. @Book{Farin:1995:NCS, author = "Gerald E. Farin", title = "{NURB} curves and surfaces: from projective geometry to practical use", publisher = pub-A-K-PETERS, address = pub-A-K-PETERS:adr, edition = "Second", pages = "xv + 267", year = "1999", ISBN = "1-56881-084-9", LCCN = "QA224 .F37 1999", bibdate = "Fri Oct 26 08:53:08 2001", acknowledgement = ack-nhfb, libnote = "Not yet in my library.", } Does anyone present have this book? One small adjustment to cope with familiar facts. The outline cannot in general be better than piecewise NURBS. Consider a bezier curve with an interval of curvature exceeding the curvature(s) of the pen. This creates an "armpit" where two smooth pieces of outline meet at an angle. It is not even always true that the pieces are all of the same degree. To see this consider a bezier curve with a cusp singularity and use a circular pen. Outward from the cusp there is a segment of the penstroke outline that is a semicircle and hence of degree 2 while other pieces are certainly of greater degree. In fact it is rumored that a list member claims to know that degree >= 10 would be involved. This behavior would compromise the value of a positive answer -- in that even a good NURBS enhanced MF/MP type design program would generate 'unexpected' complexity. Cheers Laurent S. From - Fri Nov 16 08:06:59 2001 Return-Path: Received: from coxeter.math.toronto.edu (coxeter.math.toronto.edu [128.100.68.3]) by nef.ens.fr (8.10.1/1.01.28121999) with ESMTP id fAG76oU19549 for ; Fri, 16 Nov 2001 08:06:50 +0100 (CET) Received: (from laurent@localhost) by coxeter.math.toronto.edu (8.11.0/8.11.0/UTMath 1.0) id fAG76n915348 for metafont@ens.fr; Fri, 16 Nov 2001 02:06:49 -0500 Date: Fri, 16 Nov 2001 02:06:49 -0500 From: Larry Siebenmann Message-Id: <200111160706.fAG76n915348@coxeter.math.toronto.edu> To: metafont@ens.fr Subject: Re: [metafont] does MF draw NURBS ?? X-Virus-Scanner: AMaVis 0.2.0-pre6 / Virus Scan X-Loop: metafont@nef.ens.fr X-Sequence: 497 Precedence: list [hoping this is not a double, my mailer crashed] Hi all, Let me return to Lars Helstrom's quiery: LH> Metafont is usually drawing by moving an elliptic pen along a path > which is a cubic spline, thus blackening an area whose outline is > not a cubic spline. With Nelson's message in mind, it could however > be worth investigating whether the outline would be a NURBS. And if so of what degree? I expect the answer is *no* and *known* though I don't have it in hand. I wonder if it is answered in the newish book cited by Nelson Beebe. @Book{Farin:1995:NCS, author = "Gerald E. Farin", title = "{NURB} curves and surfaces: from projective geometry to practical use", publisher = pub-A-K-PETERS, address = pub-A-K-PETERS:adr, edition = "Second", pages = "xv + 267", year = "1999", ISBN = "1-56881-084-9", LCCN = "QA224 .F37 1999", bibdate = "Fri Oct 26 08:53:08 2001", acknowledgement = ack-nhfb, libnote = "Not yet in my library.", } Does anyone present have this book? One small adjustment to cope with familiar facts. The outline cannot in general be better than piecewise NURBS. Consider a bezier curve with an interval of curvature exceeding the curvature(s) of the pen. This creates an "armpit" where two smooth pieces of outline meet at an angle. It is not even always true that the pieces are all of the same degree. To see this consider a bezier curve with a cusp singularity and use a circular pen. Outward from the cusp there is a segment of the penstroke outline that is a semicircle and hence of degree 2 while other pieces are certainly of greater degree. In fact it is rumored that a list member claims to know that degree >= 10 would be involved. This behavior would compromise the value of a positive answer -- in that even a good NURBS enhanced MF/MP type design program would generate 'unexpected' complexity. Cheers Laurent S. PS. Are there good NURBS rendering tools with GUI on Wintel, Lintel, Mac etc... ? PPS. I suspect Knuth & Hobby inspected and rejected NURBS for MF/MP. Is this said explicitly somewhere? PPPS. My guesses are more invitations to discussion than strong beliefs. From - Sat Nov 17 21:48:22 2001 Return-Path: Received: from moutvdom01.kundenserver.de (moutvdom01.kundenserver.de [195.20.224.200]) by nef.ens.fr (8.10.1/1.01.28121999) with ESMTP id fAHKmCU05114 for ; Sat, 17 Nov 2001 21:48:12 +0100 (CET) Received: from [195.20.224.204] (helo=mrvdom00.schlund.de) by moutvdom01.kundenserver.de with esmtp (Exim 2.12 #2) id 165CNo-0002ay-00 for metafont@ens.fr; Sat, 17 Nov 2001 21:48:12 +0100 Received: from pd956aa91.dip0.t-ipconnect.de ([217.86.170.145] helo=1und1/pt1094-284) by mrvdom00.schlund.de with esmtp (Exim 2.12 #2) id 165CNn-00054i-00 for metafont@ens.fr; Sat, 17 Nov 2001 21:48:11 +0100 Reply-To: From: "Wolfgang Werners-Lucchini" To: Date: Sat, 17 Nov 2001 21:44:59 +0100 X-Msmail-Priority: Normal X-Priority: 3 X-Mailer: Microsoft Internet Mail 4.70.1132 Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Message-Id: X-Virus-Scanner: AMaVis 0.2.0-pre6 / Virus Scan X-Loop: metafont@nef.ens.fr X-Sequence: 498 Precedence: list Subject: [metafont] Hallo! May be, the following message was going lost. So I try it once more: > I would like to use mp and ghostscript (WITHOUT tex,latex and dvips). How > can I print german umlaut (äöüßÄÖÜ). > > Thanks, Wolfgang From - Mon Nov 19 21:47:11 2001 Return-Path: Received: from moutvdom01.kundenserver.de (moutvdom01.kundenserver.de [195.20.224.200]) by nef.ens.fr (8.10.1/1.01.28121999) with ESMTP id fAJKl2U33689 for ; Mon, 19 Nov 2001 21:47:02 +0100 (CET) Received: from [195.20.224.204] (helo=mrvdom00.schlund.de) by moutvdom01.kundenserver.de with esmtp (Exim 2.12 #2) id 165vJm-0002Iy-00 for metafont@ens.fr; Mon, 19 Nov 2001 21:47:02 +0100 Received: from pd90104d6.dip0.t-ipconnect.de ([217.1.4.214] helo=1und1/pt1094-284) by mrvdom00.schlund.de with esmtp (Exim 2.12 #2) id 165vJl-0002EN-00 for metafont@ens.fr; Mon, 19 Nov 2001 21:47:01 +0100 Reply-To: From: "Wolfgang Werners-Lucchini" To: Subject: Re: [metafont] mp without tex Date: Mon, 19 Nov 2001 21:44:48 +0100 X-Msmail-Priority: Normal X-Priority: 3 X-Mailer: Microsoft Internet Mail 4.70.1132 Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Message-Id: X-Virus-Scanner: AMaVis 0.2.0-pre6 / Virus Scan X-Loop: metafont@nef.ens.fr X-Sequence: 499 Precedence: list Hallo! > > I would like to use mp and ghostscript (WITHOUT tex,latex and dvips). How > > can I print german umlaut (äöüßÄÖÜ). > > Nelson H.F. Beebe wrote: > > Perhaps if you gave a small example of what you are trying to do, > someone (maybe me) could offer comments. > I would like to print the following with ghostscript: ------------------------------------ beginfig(1) % prologues:=2; defaultfont:="putr8r"; %defaultfont:="Utopia-Regular"; defaultscale:=5; z1=(5cm,5cm); draw thelabel.bot("kühl", z1); endfig; end ------------------------------------ Problem: The umlaut 'ü' above is not printed at all! At the moment I have to change 'Standard-Encoding' and to edit the ps-output in the following way: ------------------------------------ %!PS-Adobe-3.0 EPSF-3.0 %%BoundingBox: 106 101 177 139 %%Creator: MetaPost %%CreationDate: 2001.11.09:2325 %%Pages: 1 %%DocumentFonts: Utopia-Regular /putr8a /Utopia-Regular def /fshow {exch findfont exch scalefont setfont show}bind def /ReEncode { exch findfont dup length dict begin { 1 index /FID eq { pop pop } { def } ifelse } forall /Encoding ISOLatin1Encoding def currentdict end definefont pop } bind def putr8a /ISOfont ReEncode %%EndProlog %%Page: 1 1 106.6143 102.02022 moveto (k\374hl) /ISOfont 49.81323 fshow showpage %%EOF ------------------------------------------- What can I do, to teach gs or mp to include similar code. Thank you, Wolfgang From - Sat Nov 24 23:52:52 2001 Return-Path: Received: from moutvdom01.kundenserver.de (moutvdom01.kundenserver.de [195.20.224.200]) by nef.ens.fr (8.10.1/1.01.28121999) with ESMTP id fAOMqdU06915 for ; Sat, 24 Nov 2001 23:52:39 +0100 (CET) Received: from [195.20.224.220] (helo=mrvdom04.kundenserver.de) by moutvdom01.kundenserver.de with esmtp (Exim 2.12 #2) id 167lf5-0007nv-00 for metafont@ens.fr; Sat, 24 Nov 2001 23:52:39 +0100 Received: from pd956aab7.dip0.t-ipconnect.de ([217.86.170.183] helo=1und1/pt1094-284) by mrvdom04.kundenserver.de with esmtp (Exim 2.12 #2) id 167lf4-00018i-00 for metafont@ens.fr; Sat, 24 Nov 2001 23:52:39 +0100 Reply-To: From: "Wolfgang Werners-Lucchini" To: Date: Sat, 24 Nov 2001 23:51:11 +0100 X-Msmail-Priority: Normal X-Priority: 3 X-Mailer: Microsoft Internet Mail 4.70.1132 Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Message-Id: X-Virus-Scanner: AMaVis 0.2.0-pre6 / Virus Scan X-Loop: metafont@nef.ens.fr X-Sequence: 500 Precedence: list Subject: [metafont] ligtable problem Hallo! I've two characters 's', a long and a short one, like in fraktur. The short 's' with characterposition 0, the long 's' at the normal 's'-position. With the following ligature-program the long 's' is changed to short 's' if it is at the end of the word: boundarychar := 255; ligtable "s" : 255 =:| 0, ... etc. This works. But there should too be inserted a negativ kerning between an 'e' and an following short 's'. The code ligtable "e" : 0 kern -.5u#; is working only for short 's' who are inserted originaly ("e\char0 ") the automaticaly changed variant ("es ") is not seen by this ligtable command. But ligtable "e" : "s" kern -.5u#; works for both, the changed long 's' (what I want) and the unchanged long 's' (what I don't want)! Strange! Is'nt it? Wolfgang From - Sun Nov 25 06:03:47 2001 Return-Path: Received: from coxeter.math.toronto.edu (coxeter.math.toronto.edu [128.100.68.3]) by nef.ens.fr (8.10.1/1.01.28121999) with ESMTP id fAP53aU24731 for ; Sun, 25 Nov 2001 06:03:36 +0100 (CET) Received: (from laurent@localhost) by coxeter.math.toronto.edu (8.11.0/8.11.0/UTMath 1.0) id fAP53Id41396; Sun, 25 Nov 2001 00:03:18 -0500 Date: Sun, 25 Nov 2001 00:03:18 -0500 From: Larry Siebenmann Message-Id: <200111250503.fAP53Id41396@coxeter.math.toronto.edu> To: B.Jackowski@GUST.ORG.PL, lcs@math.toronto.edu, metafont@ens.fr, wwl@musensturm.de X-Virus-Scanner: AMaVis 0.2.0-pre6 / Virus Scan X-Loop: metafont@nef.ens.fr X-Sequence: 501 Precedence: list Subject: [metafont] Re: ligtable problem (a) The bad news Hi Wolfgang, You complain that TeX cannot correctly compose "...". (=word boundary char.) I believe this is one of the the nasty and notorious deficiencies of TeX that prevent it from exploiting ligatures to fulfill the needs of languages other than English in the way that the TeXbook originally proposed (and may still). For fun I note that if you write the word backwards, ... things work out since ligatures kern correctly with a *following* ordinary character.:) One might summarize by saying that the present frozen TeX is short-sighted in that it inserts kerns hastily -- before following ligature behavior has been determined. In short is for NTS, Omega etc., not TeX. Cheers Laurent S. PS. The basic reference on this is an article by Jackowsky and Rycko(sp?), Eurotex, Prague circa 1992 (write to B.Jackowski@GUST.ORG.PL); the authors have the copyright, so that article can and should be put on CTAN. From - Sun Nov 25 06:04:23 2001 Return-Path: Received: from coxeter.math.toronto.edu (coxeter.math.toronto.edu [128.100.68.3]) by nef.ens.fr (8.10.1/1.01.28121999) with ESMTP id fAP54EU24843 for ; Sun, 25 Nov 2001 06:04:14 +0100 (CET) Received: (from laurent@localhost) by coxeter.math.toronto.edu (8.11.0/8.11.0/UTMath 1.0) id fAP54Df41398; Sun, 25 Nov 2001 00:04:13 -0500 Date: Sun, 25 Nov 2001 00:04:13 -0500 From: Larry Siebenmann Message-Id: <200111250504.fAP54Df41398@coxeter.math.toronto.edu> To: lcs@math.toronto.edu, metafont@ens.fr, wwl@musensturm.de X-Virus-Scanner: AMaVis 0.2.0-pre6 / Virus Scan X-Loop: metafont@nef.ens.fr X-Sequence: 502 Precedence: list Subject: [metafont] Re: ligtable problem (b) The good news? Hi Wolfgang, You complain that TeX cannot correctly compose "...". (=word boundary char.) In case you cannot hold hold your breath waiting for the ultimate TeX, here is an ad hoc workaround. Construct vf that has enough ligatures like: --> where the latter itself ligatures with to produce the appropriate result. The character would be e appropriately kerned to short s. And similarly --> . The vf will be big and nonstandard, but if DVIcopy is in a good mood it will reduce output to an impeccable standard DVI. Incidentally, does some variant/format of TeX automatically insert s? I do not need/use s myself! Cheers Laurent S. PS. Concerning getting \"u in type1 metapost labels without calling TeX, is it not just a matter of setting up a type1 font with an encoding which features \"u as a single character? (Again something a bit beyond my horizon.) From - Mon Nov 26 17:18:06 2001 Return-Path: Received: from ifi.informatik.uni-stuttgart.de (ifi.informatik.uni-stuttgart.de [129.69.211.1]) by nef.ens.fr (8.10.1/1.01.28121999) with ESMTP id fAQGHtU94352 for ; Mon, 26 Nov 2001 17:17:55 +0100 (CET) Received: from isar.informatik.uni-stuttgart.de (isar [129.69.215.232]) by ifi.informatik.uni-stuttgart.de (8.9.3/2.2) with ESMTP id RAA08499; Mon, 26 Nov 2001 17:17:33 +0100 (MET) Received: (from raichle@localhost) by isar.informatik.uni-stuttgart.de (8.9.3/2.2) id RAA09411; Mon, 26 Nov 2001 17:17:48 +0100 (MET) X-Authentication-Warning: isar.informatik.uni-stuttgart.de: raichle set sender to raichle@Informatik.Uni-Stuttgart.DE using -f From: Bernd Raichle Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-Id: <15362.27435.855846.456894@is.informatik.uni-stuttgart.de> Date: Mon, 26 Nov 2001 17:17:47 +0100 (MET) To: metafont@ens.fr Subject: Re: [metafont] Re: ligtable problem (a) In-Reply-To: <200111250503.fAP53Id41396@coxeter.math.toronto.edu> References: <200111250503.fAP53Id41396@coxeter.math.toronto.edu> X-Mailer: VM 6.75 under Emacs 20.7.1 X-Virus-Scanner: AMaVis 0.2.0-pre6 / Virus Scan X-Loop: metafont@nef.ens.fr X-Sequence: 503 Precedence: list On Sunday, 25 November 2001 00:03:18 -0500, Larry Siebenmann writes: [...] > You complain that TeX cannot correctly compose > "...". (=word boundary char.) [...] > One might summarize by saying that the present frozen TeX > is short-sighted in that it inserts kerns hastily -- > before following ligature behavior has been determined. TeX is short-sighted in the sense that TeX only builds ligatures or inserts kerns between - a char node and a following char node or - a ligature node and a following char node. TeX ignores the fact that the second component in the ligature/kern program of a font can be a ligature, too. This is one of Knuth's over-optimization in the inner loop area simplifying the code and reducing runtime. For TeX this can not be fixed within TeX--The Program, because it will break a lot of documents (and probably some fonts?). For e-TeX and/or NTS it can be fixed with some effort. > In short is for NTS, Omega etc., not TeX. No! Ligatures using the is for TeX as any other ligature in a TeX font. If your statement would be true, the same statement applies to the ligatures '! and '?, i.e., for all ligatures whose glyph differ from the glyph of its first constituent because for these cases the probability is higher that the ligature needs kerning different from the one needed for the first constituent. For almost any ligature in the standard CM fonts this ``bug'' in TeX applies---it is invisible because the ligature glyph needs (almost) the same kerning as the glyph of the first constituent: ``, '', --, ---. [...] Best wishes, -bernd From - Tue Nov 27 23:23:19 2001 Return-Path: Received: from moutvdom00.kundenserver.de (moutvdom00.kundenserver.de [195.20.224.149]) by nef.ens.fr (8.10.1/1.01.28121999) with ESMTP id fARMN6U78502 for ; Tue, 27 Nov 2001 23:23:06 +0100 (CET) Received: from [195.20.224.220] (helo=mrvdom04.kundenserver.de) by moutvdom00.kundenserver.de with esmtp (Exim 2.12 #2) id 168qd7-0007X2-00 for metafont@ens.fr; Tue, 27 Nov 2001 23:23:05 +0100 Received: from pd956aad3.dip0.t-ipconnect.de ([217.86.170.211] helo=1und1/pt1094-284) by mrvdom04.kundenserver.de with esmtp (Exim 2.12 #2) id 168qd7-0004X4-00 for metafont@ens.fr; Tue, 27 Nov 2001 23:23:05 +0100 Reply-To: From: "Wolfgang Werners-Lucchini" To: Subject: Re: [metafont] Re: ligtable problem (b) Date: Tue, 27 Nov 2001 23:21:37 +0100 X-Msmail-Priority: Normal X-Priority: 3 X-Mailer: Microsoft Internet Mail 4.70.1132 Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Message-Id: X-Virus-Scanner: AMaVis 0.2.0-pre6 / Virus Scan X-Loop: metafont@nef.ens.fr X-Sequence: 504 Precedence: list > The good news? > > Hi Wolfgang, > > You complain that TeX cannot correctly compose > "...". (=word boundary char.) > > In case you cannot hold hold your breath waiting for > the ultimate TeX, here is an ad hoc workaround. Construct > vf that has enough ligatures like: --> > where the latter itself ligatures with to produce > the appropriate result. The character would > be e appropriately kerned to short s. And similarly > --> . > > The vf will be big and nonstandard, but if DVIcopy is in a > good mood it will reduce output to an impeccable standard > DVI. 1. Because I have never created a vf (vpl) I would like to create such a thing as an exercise. Do you have some tips? I propably have to start with the pl-translation of my tfm!? 2. At the moment I use a second workaround in metafont. I 've created two more ligatures e and e with proper kern. Than the exchange e --> e is working. Wolfgang From - Wed Nov 28 09:31:37 2001 Return-Path: Received: from linux.bop.com.pl (pg85.gdansk.sdi.tpnet.pl [213.77.190.85]) by nef.ens.fr (8.10.1/1.01.28121999) with ESMTP id fAS8VQU27184 for ; Wed, 28 Nov 2001 09:31:27 +0100 (CET) Received: from localhost (bop@localhost) by linux.bop.com.pl (8.9.3/8.9.3) with ESMTP id JAA05379 for ; Wed, 28 Nov 2001 09:31:10 +0100 X-Authentication-Warning: linux.bop.com.pl: bop owned process doing -bs Date: Wed, 28 Nov 2001 09:31:10 +0100 (CET) From: Boguslaw Jackowski X-Sender: bop@linux.bop.com.pl To: The Discussion List of METAFONT/METAPOST Subject: Re: [metafont] Re: ligtable problem (b) Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanner: AMaVis 0.2.0-pre6 / Virus Scan X-Loop: metafont@nef.ens.fr X-Sequence: 505 Precedence: list Hi, Larry Siebenmann: > TeX cannot correctly compose > "...". (=word boundary char.) > [...] > here is an ad hoc workaround. Construct > vf that has enough ligatures [...] Isn't that true that everything that can be represented by virtual fonts can also be programmed from within METAFONT? I would risk at least the following statement: as far as metric tricks/freaks are considered, vf fonts are not needed. Am I right that virtual fonts prove useful primarily to ``shuffle'' several _already existent_ fonts together? Wolfgang Werners-Lucchini > At the moment I use a second workaround in metafont. > I 've created two more ligatures e and e with proper kern. > Than the exchange e --> e is working. I believe that this is a perfect solution for this particular problem. All the best -- Jacko Ps. Bernd Raichle: > all ligatures whose glyph differ from the glyph > of its first constituent because for these cases the > probability is higher that the ligature needs kerning > different from the one needed for the first constituent. Right. That's exactly the reason why I am obstinately against using ligatures as a mean of accessing diacritical characters. Perhaps the most crazy example of such an approach one may find in sources of WNCYR fonts (see, e.g., cyrfont.mf). =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Bogus\l{}aw Jackowski: B.Jackowski@GUST.ORG.PL ---------------------------------------------------------------- Hofstadter's Law: It always takes longer than you expect, even when you take into account Hofstadter's Law. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- From - Wed Nov 28 11:16:11 2001 Return-Path: Received: from ifi.informatik.uni-stuttgart.de (ifi.informatik.uni-stuttgart.de [129.69.211.1]) by nef.ens.fr (8.10.1/1.01.28121999) with ESMTP id fASAG2U46743 for ; Wed, 28 Nov 2001 11:16:03 +0100 (CET) Received: from isar.informatik.uni-stuttgart.de (isar [129.69.215.232]) by ifi.informatik.uni-stuttgart.de (8.9.3/2.2) with ESMTP id LAA12084; Wed, 28 Nov 2001 11:15:28 +0100 (MET) Received: (from raichle@localhost) by isar.informatik.uni-stuttgart.de (8.9.3/2.2) id LAA22071; Wed, 28 Nov 2001 11:15:43 +0100 (MET) X-Authentication-Warning: isar.informatik.uni-stuttgart.de: raichle set sender to raichle@Informatik.Uni-Stuttgart.DE using -f From: Bernd Raichle Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-Id: <15364.47439.427903.422484@is.informatik.uni-stuttgart.de> Date: Wed, 28 Nov 2001 11:15:43 +0100 (MET) To: The Discussion List of METAFONT/METAPOST Subject: Re: [metafont] Re: ligtable problem (b) In-Reply-To: References: X-Mailer: VM 6.75 under Emacs 20.7.1 X-Virus-Scanner: AMaVis 0.2.0-pre6 / Virus Scan X-Loop: metafont@nef.ens.fr X-Sequence: 506 Precedence: list On Wednesday, 28 November 2001 09:31:10 +0100, Boguslaw Jackowski writes: > Larry Siebenmann: > > TeX cannot correctly compose > > "...". (=word boundary char.) > > [...] > > here is an ad hoc workaround. Construct > > vf that has enough ligatures [...] > > Isn't that true that everything that can be represented by virtual fonts can > also be programmed from within METAFONT? Mmmh, no, this is not true. In a virtual font the code for a virtual glyph is DVI code with some restrictions (e.g. a glyph can not be more than one DVI page ;-). > I would risk at least the following > statement: as far as metric tricks/freaks are considered, vf fonts are not > needed. This is IMHO true. > Am I right that virtual fonts prove useful primarily to ``shuffle'' several > _already existent_ fonts together? This was the use when Knuth invented them and this is their major use nowadays. Nonetheless virtual fonts can be used for more things ... > Wolfgang Werners-Lucchini > > At the moment I use a second workaround in metafont. > > I 've created two more ligatures e and e with proper kern. > > Than the exchange e --> e is working. > > I believe that this is a perfect solution for this particular problem. ...with the restriction that each additional ``helper'' ligature will use one glyph position in the font and Tex can not support more than 256 glyphs per font. Thus if there are not enough free glyph positions or too many combination of ``helper'' ligature, you will loose :-( [...] > Bernd Raichle: > > all ligatures whose glyph differ from the glyph > > of its first constituent because for these cases the > > probability is higher that the ligature needs kerning > > different from the one needed for the first constituent. > > Right. That's exactly the reason why I am obstinately against using ligatures > as a mean of accessing diacritical characters. Perhaps the most crazy example > of such an approach one may find in sources of WNCYR fonts (see, e.g., > cyrfont.mf). Add me, too! IMHO it would be better to distinguish between (a) font ligatures like ``fl'', ``fi'' etc. and (b) a transliteration scheme. In TeX both are mixed, which is one reason of TeX's inability to insert a kern between two ligatures. The transliteration scheme (e.g. ``-'' + ``-'' => en-dash) is needed for typing text into a restricted keyboard giving access to more glyphs, i.e., it works on the text character input level ...whereas font ligature are working on the font glyph level. Best wishes, -bernd From - Thu Nov 29 09:57:51 2001 Return-Path: Received: from coxeter.math.toronto.edu (coxeter.math.toronto.edu [128.100.68.3]) by nef.ens.fr (8.10.1/1.01.28121999) with ESMTP id fAT8vfU83307 for ; Thu, 29 Nov 2001 09:57:41 +0100 (CET) Received: (from laurent@localhost) by coxeter.math.toronto.edu (8.11.0/8.11.0/UTMath 1.0) id fAT8veI126570; Thu, 29 Nov 2001 03:57:40 -0500 Date: Thu, 29 Nov 2001 03:57:40 -0500 From: Larry Siebenmann Message-Id: <200111290857.fAT8veI126570@coxeter.math.toronto.edu> To: laurent@math.toronto.edu, metafont@ens.fr X-Virus-Scanner: AMaVis 0.2.0-pre6 / Virus Scan X-Loop: metafont@nef.ens.fr X-Sequence: 507 Precedence: list Subject: [metafont] Re: ligtable problem(b) -- more good news? Hi Wolfgang, Your first tactic of introducing the exotic ligatures I suggested into your mf source rather than into a new vf seems OK to me, particularly if the font is your very own. If there are "standard tfm's" of the same name then there is a "polution" problem which makes vfs preferable. Public tfm's should be few and immutable. However, it is worth underlining that the choice between MF and vf's as tool for introducing my exotic ligatures does not make the blindest bit of difference to TeX. Either way, TeX itself sees identical tfm's. > Because I have never created a vf (vpl) I would like > to create such a thing as an exercise. Do you have > some tips? I propably have to start with the > pl-translation of my tfm!? Editing the ASCII versions of tfm and vf files is less 'scary' than editing MF source. Start with a "nul" vf and alter it -- using a sufficiently complex vf to learn the syntax. Simplest stuff. But on this list MF source is more beautiful (;=); and frankly vf support seems patchy to me. Concerning hyphenation, I believe the basic TeX fact is that ligatures are undone while acceptable hyphenation points are determined. I hope this means that if is assigned category "other", then, even with the exotic ligatures we are contemplating, it does not change the word seen by the hyphenation mechanism. That would be another bit of good news for you; could you confirm? Cheers Larry S. PS. Have you considered alternatives to using ? I am guessing that you have a text that may be printed in gothic type and/or in latin type at various times in its career? I confess that you are the first user of I have encountered! From - Thu Nov 29 09:58:22 2001 Return-Path: Received: from coxeter.math.toronto.edu (coxeter.math.toronto.edu [128.100.68.3]) by nef.ens.fr (8.10.1/1.01.28121999) with ESMTP id fAT8w7U83449 for ; Thu, 29 Nov 2001 09:58:08 +0100 (CET) Received: (from laurent@localhost) by coxeter.math.toronto.edu (8.11.0/8.11.0/UTMath 1.0) id fAT8w6t126572; Thu, 29 Nov 2001 03:58:06 -0500 Date: Thu, 29 Nov 2001 03:58:06 -0500 From: Larry Siebenmann Message-Id: <200111290858.fAT8w6t126572@coxeter.math.toronto.edu> To: laurent@math.toronto.edu, metafont@ens.fr X-Virus-Scanner: AMaVis 0.2.0-pre6 / Virus Scan X-Loop: metafont@nef.ens.fr X-Sequence: 508 Precedence: list Subject: [metafont] Re: ligtable problem Hi all, Jacko writes: > Perhaps the most crazy example of such an approach > [ligatures to facilitate typing] one > may find in sources of WNCYR fonts (see, e.g., > cyrfont.mf). That's the truth for year 2001. But, that ASCII typing seheme for Cyrillic is in itself is OK. And Barbara Beeton created it before TeX was out of the oven (circa 1979?). There is a presentation of it with the AMS fonts package. She can scarcely be blamed for believing that TeX can do what the TeXbook said (and still says?). Cheers Laurent S. PS. (PUBLICITY) In the last few years, I have created a very different and rather cleaner ASCII typing scheme for Russian and Ukrainian called ASCII-Cyrillic; it allows optimal typography and is independent of fonts and ligs. (But a perfectly seamless integration with (La)TeX begs for the scantokens feature of e-TeX = 'epsilon TeX':=/) See EuroTeX 2001 http://www.ntg.nl/eurotex/proceedings.html From - Thu Nov 29 09:59:27 2001 Return-Path: Received: from coxeter.math.toronto.edu (coxeter.math.toronto.edu [128.100.68.3]) by nef.ens.fr (8.10.1/1.01.28121999) with ESMTP id fAT8xGU83657 for ; Thu, 29 Nov 2001 09:59:16 +0100 (CET) Received: (from laurent@localhost) by coxeter.math.toronto.edu (8.11.0/8.11.0/UTMath 1.0) id fAT8xFP126574; Thu, 29 Nov 2001 03:59:15 -0500 Date: Thu, 29 Nov 2001 03:59:15 -0500 From: Larry Siebenmann Message-Id: <200111290859.fAT8xFP126574@coxeter.math.toronto.edu> To: laurent@math.toronto.edu, metafont@ens.fr X-Virus-Scanner: AMaVis 0.2.0-pre6 / Virus Scan X-Loop: metafont@nef.ens.fr X-Sequence: 509 Precedence: list Subject: [metafont] Re: ligtable problem Hi Bernd, I wrote concerning the "word boundary character" notion : > In short is for NTS, Omega etc., not TeX. You replied: > No! Ligatures using the is for TeX as > any other ligature in a TeX font. Perhaps I was too short and we have reached the level where "yes!" = "no!" = "yes and no!" :) To get to something objective, might I enquire about *established legitimate uses* of that do not expose the less-than-expert user to bad typography? I mean here uses where *ends* the word. Remember my auxilliary question:- How are the s to be inserted in practice? Was the ligature misfeature of TeX we are discussing really known to those who proposed ? Can you trace the geneology of ? The workaround I proposed to Wolfgang to allow use of with today's TeX has not yet been accused of any sin against good typography. But it is decidedly a "workaround" in that it uses many pairs of ligatures where a single ligature suffices for NTS (and the TeXbook:=)). As you say: > ...each additional > ``helper'' ligature will use one glyph position in the > font and Tex can not support more than 256 glyphs per > font. Indeed, where Wolfgang squeaks by, others may founder. In short, may be suitable for routine use in some NTS, but not in today's TeX. Cheers Laurent S. PS. TeX deserves explicit honorable mention here for placing an appropriate kern *after* ligatured characters. Provided, of course, that it is not followed by another ligatured character! Uses of at the *beginning* of words thus escape the TeX deficiency we are discussing. But are there really any such 'out front' uses??? From - Thu Nov 29 10:58:41 2001 Return-Path: Received: from linux.bop.com.pl (pg85.gdansk.sdi.tpnet.pl [213.77.190.85]) by nef.ens.fr (8.10.1/1.01.28121999) with ESMTP id fAT9wRU93348 for ; Thu, 29 Nov 2001 10:58:27 +0100 (CET) Received: from localhost (bop@localhost) by linux.bop.com.pl (8.9.3/8.9.3) with ESMTP id KAA00937 for ; Thu, 29 Nov 2001 10:44:20 +0100 X-Authentication-Warning: linux.bop.com.pl: bop owned process doing -bs Date: Thu, 29 Nov 2001 10:44:20 +0100 (CET) From: Boguslaw Jackowski X-Sender: bop@linux.bop.com.pl To: The Discussion List of METAFONT/METAPOST Subject: [metafont] Re: ligtable problem Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanner: AMaVis 0.2.0-pre6 / Virus Scan X-Loop: metafont@nef.ens.fr X-Sequence: 510 Precedence: list Hi, Jacko: > Perhaps the most crazy example of such an approach > [ligatures to facilitate typing] one > may find in sources of WNCYR fonts (see, e.g., > cyrfont.mf). Larry: > that ASCII typing seheme for Cyrillic is in itself is OK. And Barbara > Beeton created it before TeX was out of the oven (circa > 1979?). There is a presentation of it with the AMS fonts > package. She can scarcely be blamed for believing that > TeX can do what the TeXbook said (and still says?). I'd never blame Barbara! Oppositely, I consider WNCYR a milestone. Moreover, I wasn't clever enough in that time to propose a better solution, therefore I have no rights to criticize WNCYR seriuosly. But, anyway, WNCYR is an instructive example of a wrong design decision---emphasizing this fact was my intention. If I touched Barbara, I sincerely apologize. Cheers -- Jacko =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Bogus\l{}aw Jackowski: B.Jackowski@GUST.ORG.PL ---------------------------------------------------------------- Hofstadter's Law: It always takes longer than you expect, even when you take into account Hofstadter's Law. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- From - Thu Nov 29 12:00:10 2001 Return-Path: Received: from linux.bop.com.pl (pg85.gdansk.sdi.tpnet.pl [213.77.190.85]) by nef.ens.fr (8.10.1/1.01.28121999) with ESMTP id fATB02U02488 for ; Thu, 29 Nov 2001 12:00:02 +0100 (CET) Received: from localhost (bop@localhost) by linux.bop.com.pl (8.9.3/8.9.3) with ESMTP id LAA01114 for ; Thu, 29 Nov 2001 11:46:04 +0100 X-Authentication-Warning: linux.bop.com.pl: bop owned process doing -bs Date: Thu, 29 Nov 2001 11:46:04 +0100 (CET) From: Boguslaw Jackowski X-Sender: bop@linux.bop.com.pl To: The Discussion List of METAFONT/METAPOST Subject: [metafont] Re: ligtable problem (b) Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanner: AMaVis 0.2.0-pre6 / Virus Scan X-Loop: metafont@nef.ens.fr X-Sequence: 511 Precedence: list Hi, again, Jacko: > Isn't that true that everything that can be represented by virtual fonts > can also be programmed from within METAFONT? Bernd: > Mmmh, no, this is not true. In a virtual font the code for a virtual > glyph is DVI code with some restrictions (e.g. a glyph can not be more > than one DVI page ;-). Agreed. Jacko: > Am I right that virtual fonts prove useful primarily to ``shuffle'' > several _already existent_ fonts together? Bernd: > This was the use when Knuth invented them and this is their major use > nowadays. Nonetheless virtual fonts can be used for more things ... Could you provide some examples of practical applications? > IMHO it would be better to distinguish between (a) font ligatures like > ``fl'', ``fi'' etc. and (b) a transliteration scheme. > In TeX both are mixed True. But, on the other hand, isn't transliteration a proper task for a good text editor, being able to convert on the fly to-and-fro TeX-specific notation and its visual representation? For example, my colleagues programmed emacs in such a way, that the Polish diacritical characters are represented in a source file using our parochial slash-notation (/a stands for aogonek, /c -- for cacute, etc.), but you see them on the screen as usual letters. Of course, it cannot be sort of ``a standard of conversion'' (there are some 6000 languages all over the world!) -- such an interface should be flexible, preferably programmable. I fancy that one could have a possibility of programming (e.g., using METAFONT ;-) the screen representation of a well-specified sequence of characters and associate this sequence with the representation... [Actually, the problem is more complex than it may seem at the first glance, so I better stop here. I'd gladly accept commenting my enunciation, but rather privately, in order to avoid boring to death the participants of the MF/MP list. I'm not an expert in text editors and I cannot exclude that such an idea has already been implemented. Does anybody know about it?] All the best -- Jacko =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Bogus\l{}aw Jackowski: B.Jackowski@GUST.ORG.PL ---------------------------------------------------------------- Hofstadter's Law: It always takes longer than you expect, even when you take into account Hofstadter's Law. =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-