[NTG-context] \nocap does not work

Khaled Hosny khaledhosny at eglug.org
Tue Apr 13 13:03:23 CEST 2010

On Tue, Apr 13, 2010 at 12:13:49PM +0200, Hans Hagen wrote:
> On 13-4-2010 12:58, Khaled Hosny wrote:
> >On Mon, Apr 12, 2010 at 11:41:59PM +0200, Hans Hagen wrote:
> >>On 12-4-2010 9:43, Khaled Hosny wrote:
> >>>On Mon, Apr 12, 2010 at 09:23:02PM +0200, Marco wrote:
> >>>>>With MkIV \cap and \nocap is no longer necessary when you have a
> >>>>>opentype font
> >>>>>with the smcp (smallcapitals) and c2sc (capitals to smallcapitals)
> >>>>>features.
> >>>>Thanks for your explanation. But why doesn't your example work with the
> >>>>latin modern fonts? Don't they have these features? I'd be surprised
> >>>>when they don't have small caps.
> >>>
> >>>They don't have a small caps OpenType feature (smcp) for a reason beyond
> >>>to me, they however have an old-fashioned separate small caps font.
> >>
> >>this is because they are cm compatible (the initial objective of the
> >>project was a merge of all those variants)
> >>
> >>in the future we might have extra lm fonts that have those features
> >>but the team is still not sure if this should be done ... it makes
> >>much sense for consistency to have smallcaps merged in the normal
> >>shapes but even then there would be a separate smallcaps font as
> >>well
> >
> >A separate small caps font made sense for Type1 fonts (where CM
> >compatibility would matter), but for OpenType fonts I fail to see the
> >reasoning for this.
> this is not the place to elaborate on this (more for the usual font
> bof session at bachotek)

I don't go to TeX conferences, so mailing lists is the only place I can
discuss such stuff, and TeX Gyre fonts don't have a mailing list, AFAIK.

> but to some extend it's a packaging issue:
> why a full smallcaps embedded but not an italic; a similar
> discussion is taking place with regards to the lm math fonts: either
> or not to embed the three sizes as in practice one might want to
> combine different design sizes anyway

Small caps are not usually treated as separate font faces in OpenType
applications, and the common practice is to embed it into the font,
which is not the case with Italic (also one can argue that Italic and
Roman are not mere styles, but faces of each own, unlike small caps).

> >>>I would liked to send "patches" to LM fonts project, but they don't have
> >>>a source repository and their work flow depends on proprietary software
> >>>that I cannot afford.
> >>
> >>concerning the workflow you're wrong ... they use public tools like
> >>metatype1 which is free ... and they're currently building a tool
> >>chain for the math fonts
> >
> >I was referring to OpenType programming where they use Adobe AFDKO which
> >is a proprietary software that run only on proprietary OS, and with lack
> >of sources I can't even provide patches against the feature files, I can
> >make modified fonts but I don't think they will accept it. Even
> >MetaType1, it is written with only Windows users in mind; I can't use
> >the supplied batch files and I can't understand them (with no
> >documentation at all) to write a replacement.
> the team is considering to use fontforge in script mode but then the
> oppososite is happening as fontforge does not run that well on
> windows (needs mingw or cygwin + some x server) which is not much
> different from running AFDKO in wine or so

Fontforge in script mode does not require X, mingw is enough, and AFAIK
latest FF CVS can be compiled under mingw without patching. One can
always run FF under some virtual machine etc. But to run Windows I need
a Windows license which I cannot afford.

> and just like metatype1's toolkit is written with windows (actually
> dos) in mind (the toolkit evolved over many years an at that time
> there was no easy to use linux around anyway) many other tools in
> the tex suite of programs have a strong bias towards unix so it's a
> all matter of history and as the gust foundation team is the
> official team for tex related font development they have been given
> complete freedom in choosing their tools

They have the very right to pick the tools they won't, but this is
alienating many potential contributors, an that what I was complaining

> >>concerning the repos you're right ... given earlier experiences with
> >>lack of quality assurance in public fonts the lm/gyre project
> >>follows strickt procedures and only the core team can patch ... of
> >>course you can send suggestions and patches but the core team
> >>decides
> >
> >Closed development is not the solution for lack of proper QA, proper QA
> >is the solution; you can have open development model with public repos,
> >bug tracker, roadmaps etc. and still maintain a strict QA process, there
> >are many free software projects with very high QA standards but still
> >running an open development model. With open development you encourage
> >potential contributors who may or my not very valuable to you project.
> sure, but there has been many discussions about this (mostly at
> bachotex i admit) and other font experts present (in addition to the
> team) there also have their worries about sensitive things like
> patching fonts ... of course a patch for a otf generator is not the
> same as a patch for a shape and it's the shapes that i was talking
> of .. it's already somewhat debatable to add shapes to artistic work

They already shipped glyphs with such poor artistic value, being in the
Ghostscript fonts is not an excuse, they could have resorted to the
original URW fonts, which were released under GPL, from the start if
quality is so high priority to them. So, I don't see how welcoming
potential contributers can be any worse, it might actually improve the
miserable state of some areas of the fonts.

> >>fyi: the gyre team (gust font foundation) has now reverted to the
> >>core urw fonts for which they got the copyright and it made it
> >>possible to get rid of some ugly artefacts and glyphs (again, lack
> >>of qa had rendered the latest public urw fonts somewhat useless)
> >
> >There is still a long way to go, the Greek glyph for example are ranging
> >from suboptimal to pure crap.
> indeed, greek and cyrillic are the shapes that will be kicked out
> and redone as they're bugged and crappy (something that happens when
> anyone can add to a font and the main reason why the team wanted to
> get the copyright on the originals so that they can kick out bad
> stuff cq. start from scratch)

As I said above, nothing prevented them from kicking bad glyphs out, the
fonts, including original URW++ fonts, were all under GPL (which TeX
Gyre violated for a while).


 Khaled Hosny
 Arabic localiser and member of Arabeyes.org team
 Free font developer

More information about the ntg-context mailing list