From: William Overington (WOverington@ngo.globalnet.co.uk)
Date: Tue Nov 05 2002 - 04:18:55 EST
Thomas Lotze wrote as follows.
>William Overington wrote:
>
>> I don't know for certain but I suspect that it is that font designers
>> do this so that people can use an application such as Microsoft Paint
>> to produce an illustration using the font. In the absence of regular
>> Unicode code points for the ligatures, a font designer has either to
>> use the Private Use Area and be Unicode compatible or make a
>> non-Unicode compatible font, if the font designer wishes people to be
>> able to have direct access to the ligature characters.
>
>Judging from what I' learned by now, this is not true: If a font
>designer wants to make a Unicode-compatible font, he has to use a font
>file format that supports Unicode, and those formats provide means to
>access unencoded glyphs by transforming certain strings of Unicode
>characters into them.
Well, I suppose it depends upon what one means by a file format that
supports Unicode. The TrueType format does not support the ZWJ method and
thus does not "provide means to access unencoded glyphs by transforming
certain strings of Unicode characters into them". I am unsure as to
whether, in formal terms, TrueType is "a file format that supports Unicode"
as it does not allow the ZWJ sequences to be recognized. Please note that
my sentence did have "if the font designer wishes people to be able to have
direct access to the ligature characters". However, certainly, a font
designer using an advanced font format may well not wish people to be able
to have direct access to the ligature characters. The paragraph was
replying to your question as to why someone who wants to set and print out a
page of Fraktur at present is in practice likely to have to use a font with
the ligatures encoded with code points less than 255. Please know that I am
not seeking to be pedantic over the meaning of the phrase "a file format
that supports Unicode", it is just that I get the impression that you might
possibly have not quite understood that some font formats widely used for
Unicode encoded characters, such as the TrueType format, do not support the
ZWJ glyph substitution process or, in fact, any glyph substitution process,
such as noticing the two letter ct sequence and substituting a ct ligature
glyph within the font.
>And if I understand it correctly, Unicode
>compliance can only be achieved with all of compliant documents, fonts,
>and renderers. So there appears to be no need for direct accessibility
>of ligatures, alternates etc.
I said compatible, I did not say compliant and did not mean compliant. I
was meaning compatible, in the sense that, if one wishes to produce a font
using the TrueType format and that font is to include glyphs for ligatures
such as ct and ppe, how does one do it so that the method used does not
conflct with Unicode. Using Private Use Area code points avoids conflicting
with the regular Unicode code points used for other characters.
>> There are some articles about using WordPad and Paint to produce
>> graphic effects with large characters and gold textures and so on in
>> our family webspace, together with the gold texture file and some
>> other texture files too.
>
>And what's the relevance to Unicode of that?
>
Well, in direct terms probably nothing. However, as this is a widely
distributed mailing list it might be that some readers, having read about
the matter of using ligature characters in Paint and the way that one needs
a font with code points less than 255 in order to access the ligature
characters from Paint, might like to have a go at producing such graphics,
so, having available some articles on the matter, I mentioned them.
If one considers the Gutenberg sample font, the ct ligature is available as
well, at Alt 0201 using Paint. One could use Wordpad to get the character
as well. Yet, suppose that one has an advanced format font with a ct glyph
within it yet where the font does not provide a direct code point access
glyph, but only allows a ct ligature to be displayed using a combination of
computer hardware and software which supports the advanced font format. How
is one going to get that ct ligature to display if one does not have access
to that hardware and software combination? Now certainly the attempt has
been made to trivialise the matter by reference to very very old computer
systems, yet here the problems arise with PCs manufactured in 1999.
May I add that this posting is trying to be helpful to answer questions
which you have posed, I am not seeking to reopen the discussion of whether
the Unicode Technical Committee should encode any more precomposed
ligatures. I raised that issue before the August 2002 meeting of the
Unicode Technical Committee, the committee discussed the matter at the
meeting, formed a consensus view and that consensus view was minuted and the
minutes have been published. It is simply a matter that the Unicode
Technical Committee is not going to encode any more ligatures, I have my
golden ligatures collection on the web and if people choose to ignore the
golden ligatures collection, then that is up to them and if people choose to
use the golden ligatures collection then that too is up to them. My feeling
is, that used with care, the golden ligatures collection does assist some
facilities, such as producing stylish graphics using ligatures on older
computers, to be available, hopefully giving pleasure to the people making
the graphics. I have also had great enjoyment in preparing the golden
ligatures collection in that I have learned a lot about ligatures which were
used by printers in days gone by. Hopefully, even those people who
criticise the golden ligatures collection strongly recognise that I take
great care in the documents to point out that there are alternative ways to
encode ligatures: I find no problem in doing that as my intention is to
offer opportunities not to mislead anyone by holding back information which
is needed so as to make an objective decision.
William Overington
5 November 2002
This archive was generated by hypermail 2.1.5 : Tue Nov 05 2002 - 05:07:08 EST