🚀 go-pugleaf

All Adobe PostScript fonts include both oe and ISOLatin1 - though not
at the same time, since there is no standard Encoding containing both.
PostScript fonts do not typically contain multiple spaces with
different widths - spacing is done by the PostScript program, not by
font choice.

I think you should not be disappointed with fonts, but with the tools
you are using for layout. Things like non-breaking spaces, special
space widths etc. belong in there - not by special font techniques. 

Aandi Inston

Aandi, thanks for the reply.

It seems as though you are saying that the print engine (Antenna House) should be interpreting these characters and creating the correct spacing and breaking behavior without any glyph in the font. I guess that AH is saying that that is not their job -- we have to find a font with the appropriate glyph, and they will print that glyph. (The fact that changing fonts for this seems to throw off the proper breaking behavior does seem to me to be a bug on the AH side.)

That the font has to supply the glyph seems pretty sensible to me, especially since characters like the thin space are ill-defined (anywhere from 0.25 em to 0.125 em) and you may want to choose the particular rendering of a given font.

I, as a naive user in terms of typography and more particulary PostScript implementation of it, would say that every Unicode character should have a glyph in a font that provides support for that character. Why should spaces be treated in any different way than non-space characters? I want them to work even if I'm just printing the file as a text file. The no-break space has long been treated in this fashion, so why should the others not be? After all, we're only talking about roughly 15 spacing characters.

I realize that most of these spaces expand in certain situations (justification, quadding), and that is part of the print engine's function in terms of padding, but I would say that they all need to have a minimum glyph (even if it is truly minimum -- i.e., zero-width). However, many always have a fixed width for a given font, and it seems to me that particularly in those cases, there should be a glyph for them in a supporting font. Why else would MS Arial Unicode have these glyphs?

The same goes for ligatures -- if they're in Unicode, they should have a glyph in a supporting font. I have read that the Unicode consortium will not accept more ligatures without a very strong case that they are necessary, but that is a different issue -- they are then saying that the creation of those ligatures is a function of the rendering engine. But for any ligatures already in Unicode, it seems to me that there should be a glyph in a supporting font. I don't want to have to rely on PostScript and the vagaries of a particular implementation in order to print the Unicode character that I want.

It may be (as it seems to me from your response) that this was not the expectation of PostScript people (and maybe other rendering technology people) in the past, but that seems to be a perspective from before Unicode. I think that Unicode has (or should have) changed the rules in this regard.

I may be missing the boat on this, but that's the way it seems to me.

Ed Hicks

>I, as a naive user in terms of typography and more particulary PostScript implementation of it, would say 
>that every Unicode character should have a glyph in a font that provides support for that character. 

You should bear in mind that most PostScript fonts in use, and all
those resident in most printers, are not Unicode fonts (if this phrase
has any meaning). They are type 1 fonts, using technology invented in
1985, and limited to 255 characters per encoding.

You may feel that Unicode should have changed the rules, and in a
sense it has. PostScript supported large fonts from around 1990, but
initially for far eastern languages.  The technology has now been
remodelled and recycled into CID fonts and thence into OpenType.  If
you look for OpenType fonts, you may have better luck.  Standard
PostScript fonts cannot be expected to have newly invented characters.
(Though it seems an exception was made for the euro).

> Why should spaces be treated in any different way than non-space characters? 

In font terms, they shouldn't. But I don't think your expectations of
finding a font to do what you want - at least a PostScript font - are
that likely to succeed.  I definitely feel that typography is the job
of the typographic program.


Aandi Inston

Aandi, thanks again for your further explanation.

Up front, I will say that there is no need to respond to this -- I am just decrying the "state of the art".

I realize that:

(a) we're caught in a "technology transform" where Unicode has created expectations that are still far from being met in practical applications (for various reasons), and

(b) there are different ways to achieve a desired result (for example, glyphs versus what I would call "character-generated processing instructions", or what you would say is the typographic program doing the typography).

As an example of (a), I have searched for years for a truly competent Unicode text editor, and I have yet to find one. I have tried many, and UltraEdit is the closest thing to fulfilling my needs, but as of a year ago, it still exhibited too many bugs and limitations for me to use, particularly when trying to search and replace Unicode characters using regular expressions. I fall back on using a byte-oriented editor and putting up with using the UTF-8 encodings for the multi-byte characters in my XML files, and learning to read wierd-looking character groupings as something meaningful. It's reminiscent of reading hex memory dumps from an IBM mainframe in 1969, and being able to fairly adeptly interpret assembly language instruction codes in them. (NOT my idea of fun, in either case!)

I relate this to the state of the art in the font industry at the moment.

Obviously, it is not difficult to add support for the space glyphs in a downloaded font, particularly a TrueType font (and I presume a PostScript Type 2 font, and an OpenType font). It is there in MS Arial Unicode, and the space glyphs work fine in the Antenna House application. Trouble is, we don't want Arial, we want Times.

So I accept your explanations, but they don't really explain why someone -- like Adobe or Microsoft -- isn't doing more to fill the void for Unicode characters in many "standard" fonts. I'm not necessarily talking about full Unicode support, but at least for Latin-alphabet languages -- why should a user have to change fonts just to deal with all those languages which have such a similar base? For example, a somewhat-expanded version of Times New Roman that does this would seem to be trivial to provide, and very useful.

(In fact, I would think that the European Union would be pressing for an expanded version of that, including Greek and probably now Cyrillic characters -- with all that they spend on other esoterica, it would be an insignificant expense. Perhaps they have had this done already for numerous fonts -- but probably not, because they're not that well-organized and probably aren't even making extensive use of Unicode yet. ;-))

And of course even applications that are supposedly Unicode-capable underneath, like MS Office, are not really capable in the user interface, so we've got a long way to go.

As someone who has struggled with supporting just the English and French languages from DOS days in the 1980s, including keyboard support as well as printing and application support, I can say that the amount of time that is still taken up with trying to resolve these sorts of problems around the world is absolutely mind-boggling. Yet the cost of resolving them via decent Unicode support would be trivial in comparison. And computer speed, memory and storage space are no longer significant considerations, nor is bandwidth, especially with UTF-8 for Latin-alphabet languages. It is very frustrating!

One can always hope that there is hope for the future!

Ed Hicks

I think discussion on where fonts are going must focus on OpenType.
But OpenType fonts are NOT PostScript fonts, not even so called
"PostScript OpenType fonts" - you can't use them in a PostScript
program. (They are converted to PostScript dynamically as you print,
just like TrueType fonts).  CID fonts could do things, but nobody in
the west is making them (I think).

Hence, what I've been saying is limited to the PostScript world
inhabited by PostScript programmers. If you want to discuss where
OpenType is coming from and going to, I'd suggest the Typography
forum.

Aandi Inston