Received: from mail.proteosys.com ([62.225.9.49]) by nummer-3.proteosys with Microsoft SMTPSVC(5.0.2195.5329); Mon, 28 Apr 2003 15:28:02 +0200 Received: by mail.proteosys.com (8.12.9/8.12.2) with ESMTP id h3SDRwPQ002225 for ; Mon, 28 Apr 2003 15:28:00 +0200 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----_=_NextPart_001_01C30D89.FE061D00" Received: from sun.dante.de (root@sun.dante.de [134.100.9.52]) by rzdspc1.informatik.uni-hamburg.de (8.12.9/8.12.9) with ESMTP id h3SDEKa4013377 (version=TLSv1/SSLv3 cipher=EDH-RSA-DES-CBC3-SHA bits=168 verify=NO); Mon, 28 Apr 2003 15:14:21 +0200 (CEST) Received: from rzdspc1.informatik.uni-hamburg.de (root@rzdspc1.informatik.uni-hamburg.de [134.100.9.61]) by sun.dante.de (8.12.9/8.12.9) with ESMTP id h3SDE0e1027363; Mon, 28 Apr 2003 15:14:01 +0200 (CEST) X-MimeOLE: Produced By Microsoft Exchange V6.5 Received: from abel.math.umu.se (abel.math.umu.se [130.239.20.139]) by rzdspc1.informatik.uni-hamburg.de (8.12.9/8.12.9) with ESMTP id h3SDDua4013358 (version=TLSv1/SSLv3 cipher=EDH-RSA-DES-CBC3-SHA bits=168 verify=NO); Mon, 28 Apr 2003 15:13:56 +0200 (CEST) Received: from [130.239.20.144] (mac144.math.umu.se [130.239.20.144]) by abel.math.umu.se (8.12.3/8.12.3/Debian-6.3) with ESMTP id h3SDDrDq021113; Mon, 28 Apr 2003 15:13:53 +0200 In-Reply-To: <200304271207.h3RC7Ppt028313@postman.arcor-online.net> Return-Path: X-OriginalArrivalTime: 28 Apr 2003 13:28:03.0475 (UTC) FILETIME=[FEE72E30:01C30D89] X-Sender: lars@abel.math.umu.se X-Scanned-By: MIMEDefang 2.28 (www . roaringpenguin . com / mimedefang) X-Spam-Score: -19.3 () EMAIL_ATTRIBUTION,IN_REP_TO,MIME_LONG_LINE_QP,QUOTED_EMAIL_TEXT,REPLY_WITH_QUOTES Content-class: urn:content-classes:message Subject: Re: textcomp once more Date: Mon, 28 Apr 2003 14:12:34 +0100 Message-ID: X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: textcomp once more Thread-Index: AcMNif8E8JoDrkfISDOdNcxdjchFBA== From: =?iso-8859-1?Q?Lars_Hellstr=F6m?= To: "Walter Schmidt" Cc: "Frank Mittelbach" , , Status: R X-Status: X-Keywords: X-UID: 4602 This is a multi-part message in MIME format. ------_=_NextPart_001_01C30D89.FE061D00 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable At 13.46 +0200 2003-04-27, Walter Schmidt wrote: >On Sat, 26 Apr 2003 23:30:00 +0200, Frank Mittelbach wrote: > >>After talking with Walter at the DANTE meeting in Bremen we came to = the >>conclusion that perhaps it isn't such a bad idea to make TS1 safe for = users >>(or at least he gave in on me :-) > >Indeed, the need for yet another configuration file >(textcomp.cfg) does not make me happy. Now that I see the >actual implmentation, I really doubt whether all of this is >a good and useful idea. One could probably have some of these declarations in textcomp.sty (e.g. those for CM families), use the "full TS1" as default, and suggest that font-loading packages contain explicit declarations when they don't = provide this default. textcomp.sty could still look for textcomp.cfg, but there wouldn't by default be such a file. (This is more on the lines with the other .cfg files.) >> - is there a point in having a logic that even numbers mean a = matching >>subset >> while odd numbers mean that the subset is not complete? Do you mean that e.g. 3 would mean "somewhere between 2 and 4, all = commands for the larger of the two will work"? It could be useful to have a standardized way of saying that, but I'm not sure how useful it would be for LaTeX to take special actions in this case. >> - is there a point in the generalisation implemented that would allow = using >> this even for other font encodings, eg T1? > >Oh no! The use isn't worth the effort. And I guess it >would slow doen LaTeX. Not sure I agree there. Since it is all happening at individual = commands, there would be no difference as long as one isn't using one of these commands. The commands in other encodings that may become subjects to = this are very few, hence the difference would be very small indeed. Sourcing textcomp.cfg would probably take more time than all the processing for special symbols put together. > - i left some of the options from earlier implementations, right now = you can > define the "default" [...] > >IMHO, the default subset should be {1} rather than {5}, to >make sure that any "non-standard" symbols (if available) can >be used in additional fonts, without having to add the font >family name to textcomp.cfg. With the default being {5}, I >will be flooded with questions such as "I have installed the >font family xyz using the latest fontinst. Why can't I use >the Euro symbols?" Installing additional fonts is already >very difficult. Users should not be _forced_ to edit also >textcomp.cfg, just to access all existing symbols in a font. I agree here. See also comment above on where to put the declarations. Also, with respect to Euro, it might be possible to fake this (in textcomp.mtx) from `C' and `equals'. Consider the following fakes: \documentclass{article} \usepackage{times} \usepackage[T1]{fontenc} \newcommand{\fakeeuro}[1][\scriptsize]{% \leavevmode \valign{##\cr \vfil\hbox to 0.07em{#1=3D\hss}\vfil\cr \hbox{C}\crcr } } \begin{document} 75\fakeeuro\ \sffamily 75\fakeeuro\ \ttfamily 75\fakeeuro\ % \fontfamily{pad}\selectfont % 75\fakeeuro[]\ \fontfamily{put}\selectfont 75\fakeeuro\ \fontfamily{cmr}\selectfont 75\fakeeuro\ \fontfamily{cmss}\selectfont 75\fakeeuro\ \fontfamily{cmtt}\selectfont 75\fakeeuro[\small]\ \fontfamily{cmvtt}\selectfont 75\fakeeuro\ \end{document} At least to my unaccustomed eyes[*], most of these fake Euros are good enough that one can understand what they are supposed to be. The main problems seem to be (i) the length of the `equals' and (ii) its vertical position, but I suspect those are more easily handled by fontinst than = by LaTeX anyway. For (i), one can scale the `equals' to any size one = pleases, and for (ii) one probably will know the actual positions of the `equals' bounding box top and bottom (TeX ignores the negative depth of an = `equals' when wrapping it up in the \hbox). >Finally, let me repeat a desciption of the TS1 subsets which >I had proposed. Yes, this was quite necessary. >The comments in textcomp.cfg are not very >detailed and they are lacking my reasong (and they are in >German...). I refer to the numbering used in Frank's new >textcomp.sty now: > >#5 =3D those TS1 symbols only, that are also in the ISO-Adobe >character set; without "currency", which is often misused for >the Euro. Older Type1 fonts from the non-TeX world provide >only this subset. > >#4 =3D #5 + "Euro" >Most newer fonts provide this. If Euro can be faked, then there would not be any need for this = distinction. >#3 =3D #4 + "Omega" >Can also be described as TS1 \cap (ISO-Adobe \cup MacRoman). >(Except for the missing "currency".) Is that the only extra character? If memory serves, MacRoman \setminus ISO-Adobe is about a dozen characters. Is Omega really the only one = which is in TS1? Lars Hellstr=F6m PS: [*] The Euro isn't used in Sweden, but there's going to be a referendum = on the matter in September. Trivia of the day: The German pronounciation of "Euro" (which I find preferable to the English) sounds a lot like the Swedish "Ojr=E5" = (Ojr\aa), which is kind of slang, but would translate to English as "Oops!". ------_=_NextPart_001_01C30D89.FE061D00 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Re: textcomp once more

At 13.46 +0200 2003-04-27, Walter Schmidt = wrote:
>On Sat, 26 Apr 2003 23:30:00 +0200, Frank = Mittelbach wrote:
>
>>After talking with Walter at the DANTE = meeting in Bremen we came to the
>>conclusion that perhaps it isn't such a bad = idea to make TS1 safe for users
>>(or at least he gave in on me :-)
>
>Indeed, the need for yet another configuration = file
>(textcomp.cfg) does not make me happy.  Now = that I see the
>actual implmentation, I really doubt whether all = of this is
>a good and useful idea.

One could probably have some of these declarations in = textcomp.sty (e.g.
those for CM families), use the "full TS1" = as default, and suggest that
font-loading packages contain explicit declarations = when they don't provide
this default. textcomp.sty could still look for = textcomp.cfg, but there
wouldn't by default be such a file. (This is more on = the lines with the
other .cfg files.)

>> - is there a point in having a logic that = even numbers mean a matching
>>subset
>>   while odd numbers mean that the = subset is not complete?

Do you mean that e.g. 3 would mean "somewhere = between 2 and 4, all commands
for the larger of the two will work"? It could = be useful to have a
standardized way of saying that, but I'm not sure how = useful it would be
for LaTeX to take special actions in this = case.

>> - is there a point in the generalisation = implemented that would allow using
>>   this even for other font = encodings, eg T1?
>
>Oh no!  The use isn't worth the = effort.  And I guess it
>would slow doen LaTeX.

Not sure I agree there. Since it is all happening at = individual commands,
there would be no difference as long as one isn't = using one of these
commands. The commands in other encodings that may = become subjects to this
are very few, hence the difference would be very = small indeed. Sourcing
textcomp.cfg would probably take more time than all = the processing for
special symbols put together.

> - i left some of the options from earlier = implementations, right now you can
>   define the "default" = [...]
>
>IMHO, the default subset should be {1} rather = than {5}, to
>make sure that any "non-standard" = symbols (if available) can
>be used in additional fonts, without having to = add the font
>family name to textcomp.cfg.  With the = default being {5}, I
>will be flooded with questions such as "I = have installed the
>font family xyz using the latest fontinst.  = Why can't I use
>the Euro symbols?" Installing additional = fonts is already
>very difficult.  Users should not be = _forced_ to edit also
>textcomp.cfg, just to access all existing symbols = in a font.

I agree here. See also comment above on where to put = the declarations.

Also, with respect to Euro, it might be possible to = fake this (in
textcomp.mtx) from `C' and `equals'. Consider the = following fakes:

\documentclass{article}

\usepackage{times}
\usepackage[T1]{fontenc}

\newcommand{\fakeeuro}[1][\scriptsize]{%
   \leavevmode
   \valign{##\cr
      \vfil\hbox to = 0.07em{#1=3D\hss}\vfil\cr
      \hbox{C}\crcr
   }
}
\begin{document}

75\fakeeuro\
\sffamily
75\fakeeuro\
\ttfamily
75\fakeeuro\
% \fontfamily{pad}\selectfont
% 75\fakeeuro[]\
\fontfamily{put}\selectfont
75\fakeeuro\
\fontfamily{cmr}\selectfont
75\fakeeuro\
\fontfamily{cmss}\selectfont
75\fakeeuro\
\fontfamily{cmtt}\selectfont
75\fakeeuro[\small]\
\fontfamily{cmvtt}\selectfont
75\fakeeuro\

\end{document}

At least to my unaccustomed eyes[*], most of these = fake Euros are good
enough that one can understand what they are supposed = to be. The main
problems seem to be (i) the length of the `equals' = and (ii) its vertical
position, but I suspect those are more easily handled = by fontinst than by
LaTeX anyway. For (i), one can scale the `equals' to = any size one pleases,
and for (ii) one probably will know the actual = positions of the `equals'
bounding box top and bottom (TeX ignores the negative = depth of an `equals'
when wrapping it up in the \hbox).

>Finally, let me repeat a desciption of the TS1 = subsets which
>I had proposed.

Yes, this was quite necessary.

>The comments in textcomp.cfg are not very
>detailed and they are lacking my reasong (and = they are in
>German...).  I refer to the numbering used = in Frank's new
>textcomp.sty now:
>
>#5 =3D those TS1 symbols only, that are also in = the ISO-Adobe
>character set; without "currency", = which is often misused for
>the Euro.  Older Type1 fonts from the = non-TeX world provide
>only this subset.
>
>#4 =3D #5 + "Euro"
>Most newer fonts provide this.

If Euro can be faked, then there would not be any need = for this distinction.

>#3 =3D #4 + "Omega"
>Can also be described as TS1 \cap (ISO-Adobe \cup = MacRoman).
>(Except for the missing = "currency".)

Is that the only extra character? If memory serves, = MacRoman \setminus
ISO-Adobe is about a dozen characters. Is Omega = really the only one which
is in TS1?

Lars Hellstr=F6m

PS:
[*] The Euro isn't used in Sweden, but there's going = to be a referendum on
the matter in September.

Trivia of the day: The German pronounciation of = "Euro" (which I find
preferable to the English) sounds a lot like the = Swedish "Ojr=E5" (Ojr\aa),
which is kind of slang, but would translate to = English as "Oops!".


------_=_NextPart_001_01C30D89.FE061D00--