bugFree UCS Outline Fonts - Bugs: bug #38802, Alternative glyphs in Greek

 
 

bug #38802: Alternative glyphs in Greek

Submitter:  Rosella Capriotti <rosy58>
Submitted:  Mon 22 Apr 2013 12:17:42 PM UTC
   
 
Category:  individual character(s) Severity:  3 - Normal
Item Group:  character substitution issue Status:  Fix posted
Privacy:  Public Assigned to:  Stevan_White
Open/Closed:  Open Release:  * 2012-05-03
* Mandatory Fields

Add a New Comment Rich Markup
   

Jump to the original submission

Sun 22 Sep 2013 07:55:06 PM UTC, comment #27: 

I think you could complete the series, for reasons of consistency with:

- Variant PHI U+03D5
- Variant PI U+03D6

But these are really marginal and almost never used glyphs, but apart from that, kudos! :)

Rosella Capriotti <rosy58>
Sun 22 Sep 2013 07:28:00 PM UTC, comment #26: 

OK,

I made the title even shorter.

Could I get you to comment on the progress so far, outside of the suggestion about the lunate sigma?


Steve White <Stevan_White>
Group administrator
Sun 22 Sep 2013 06:21:01 PM UTC, comment #25: 

This topic has been expanded from its original intent, to cover the whole range of Greek alternate glyphs, therefore I suggest that you change the heading to something like "Alternate Glyphs in Greek Range".

Thanks

Rosella Capriotti <rosy58>
Tue 17 Sep 2013 06:37:21 PM UTC, comment #24: 

I think it'd be proper to have cv05 change normal Sigma to lunate Sigma only on capitals, because lunate Sigma on small letters I've never seen it while on capitals is common in Greece.

Rosella Capriotti <rosy58>
Thu 22 Aug 2013 12:17:19 PM UTC, comment #23: 

Beta



Right.  The curly form is not usual typography.  Took it out of SVN.

cv01 kappa
----------
All the character variants enable a replacement of the Unicode letters.  This one enables replacement of Unicode kappa with "script" kappa.  It is intended for cases when the default 'locl' replacements are disabled.

I'm trying to keep the "usage.txt" note up to date regarding the exact replacements being done.

Steve White <Stevan_White>
Group administrator
Thu 22 Aug 2013 11:55:12 AM UTC, comment #22: 

My comments about your choices:

default substitutions
---------------------

  • italic: script versions of beta, theta, rho. Why beta too? Wasn't script beta supposed to be enabled by cv02 and used only in non-initial positions?


discretionaary variants
-----------------------
cv01: script kappa. You mean latin-style kappa, don't you? Script kappa is enabled by default (see above), so the character variant should be put to use to enable the latin-style kappa.

Rosella Capriotti <rosy58>
Thu 22 Aug 2013 08:37:28 AM UTC, comment #21: 

Here is the latest understanding of the proper substitutions.

To maintain compliance with the standard, and to make math work as expected with no extra fiddling, glyphs in text not marked as Greek will reflect the Unicode samples.

Simply marking the text as Greek will activate several substitutions, which are different between upright and italic faces.  The intent is to bring Greek text as near modern conventions as possible.

Furthermore, various substitutions can be done individually using several character variants, if desired.  (It being also possible to disable the default 'locl' substitutions.)  These are meant to reflect alternative or historical styles.

default substitutions
---------------------

  • all faces: script version of kappa
  • italic: script versions of beta, theta, rho


discretionaary variants
-----------------------
cv01: script kappa
cv02: script beta in non-initial positions (French style)
cv03: script theta in initial positions (French style)
cv04: script theta everywhere
cv05: lunate sigma everywhere

references
----------

"Letters" by Nick Nicholas
http://www.tlg.uci.edu/~opoudjis/unicode/letters.html

"From Unicode to Typography, a Case Study: the Greek Script" by Yannis Haralambous
http://scriptsource.org/cms/scripts/page.php?item_id=source_detail&uid=tz9ucg6rx4

Steve White <Stevan_White>
Group administrator
Wed 31 Jul 2013 03:02:50 PM UTC, comment #20: 

Rosella,

OK, that's almost backward from my previous guess as to your meaning.

My understanding is that the "symbol" character is really the handwritten form.

OK, I've posted something to SVN in the non-bold serif faces, regarding kappa and theta.  Give it a shot.

I probably will make the two rho's different -- they really should be.  But maybe I can do something like the above for rho too.

Steve White <Stevan_White>
Group administrator
Wed 31 Jul 2013 02:46:46 PM UTC, comment #19: 

Hello Steve,

My point is that I've never seen something homoglyphic with U+03D1 GREEK THETA SYMBOL in the upright style, I mean not in Greek text, be it manuscript or printed, it only occurs in the cursive, so does Old Standard TT, which has a cursive style (Theano Modern doesn't). My suggestion is to make this the glyph for italic THETA.

Forget about making the two rho's different: the italic tailed rho is one of the few stylish things the Greek range has, you need to increase these niceties, not cut them down, after all, it's a serified font, isn't it? What we need is script theta as an italic glyph for THETA.

I understand your position concerning script KAPPA.

Rosella Capriotti <rosy58>
Wed 31 Jul 2013 02:01:22 PM UTC, comment #18: 

Hi

On Wed, Jul 31, 2013 at 1:32 PM, Rosella Capriotti  wrote:

> Follow-up Comment #17, bug #38802 (project freefont):
>
> I correct my statement: I've never seen any script THETA in the upright style
> in any manuscript OR book, never, and I've seen a lot of Greek text, be it
> classical, medieval or modern.
>

Maybe I'm not following your meaning of "script" in "script theta".
Are we referring to U+03B8 (theta) or U+03D1 (theta symbol) here?

Or are you referring to the "italic" (AKA cursive) face?

> I don't understand what's the matter with mathematical symbols, those who
> typeset mathematics rarely if ever do so in Greek


Math text is still written in Greek (mostly for Greek students).  But that isn't what I was referring to.

In math and physics the symbols from the main Greek range are used as symbols, all the time.  This is part of the reason the Unicode Greek range is so messed up.

>and for sure, they'll never use italic,


In some math styles, Greek is always upright, in others it is cursive.  Some styles use upright for some text and cursive for others.

> make it like the RHO, you have a precedent, follow it consistently.
>

I haven't decided what to do about that.  I only noticed the inconsistency during this discussion.

I may have to make sure the two rho's are different.

> About script KAPPA: I do insist, please make the latinate KAPPA a variant,
> nobody would seriously consider typesetting Greek with that KAPPA.


I can only imagine that we've somehow misunderstood one another here. 

FreeFont is first and foremost a Unicode font.

The Unicode samples, which are our primary guide, as well as the TheanoModern font, as well as many other fonts I can find, have the short "latinate" kappa at U+03BA, and all have the "script" or "symbol" kappa at U+03F0, as does FreeFont.

If you really think the glyphs ought to be swapped, you should consider to bring your case to the Unicode consortium.

The idea to use styles was taken from TheanoModern, and for now seems a good approach for proper typesetting. Nicolas' page says

   "As with upsilon, the kappa symbol form is merely the apla glyph for kappa, and is quite widely used in Greece. My impression is that it is less frequent among Western Classicists. It originates in the cursive form of kappa, and already appears in the 2nd century AD."

So he's saying it's a popular style, and so a style lookup seems appropriate.

> Apart from
> Bodoni-style fonts (like Theano Modern), but your font looks much more like
> Theano Didot, doesn't it?
>

Not in the sense of... not in any sense I can see.  Didiot is much more styled face, that (I think) follows the French school of the previous century. 

I see that in TheanoDidiot does use the same glyph for theta and theta symbol.  This would make it problematic to use for math typesetting, and that is a major use case for the users of FreeFont.  So we really can't use anything like this.

Cheers!

Steve White <Stevan_White>
Group administrator
Wed 31 Jul 2013 11:32:53 AM UTC, comment #17: 

Hello Steve,

I correct my statement: I've never seen any script THETA in the upright style in any manuscript OR book, never, and I've seen a lot of Greek text, be it classical, medieval or modern.

I don't understand what's the matter with mathematical symbols, those who typeset mathematics rarely if ever do so in Greek and for sure, they'll never use italic, make it like the RHO, you have a precedent, follow it consistently.

About script KAPPA: I do insist, please make the latinate KAPPA a variant, nobody would seriously consider typesetting Greek with that KAPPA. Apart from Bodoni-style fonts (like Theano Modern), but your font looks much more like Theano Didot, doesn't it?

As for lunate SIGMA, if you really want to keep it, it's up to you!

Rosella Capriotti <rosy58>
Wed 31 Jul 2013 10:28:04 AM UTC, comment #16: 

Hi

On Wed, Jul 31, 2013 at 12:26 AM, Rosella Capriotti

> Follow-up Comment #15, bug #38802 (project freefont):
>

First, let's keep in mind that the Greek range is a terrible mess in Unicode.

  • the mix of letters that should have been in the Coptic range
  • confusion of "math symbols" with greek stylistic variants
  • incusion of stylistic variants
  • inclusion of editorial marks as though they were letters


For at least one letter, the "symbol" and "alphabetic" letters were switched between versions of Unicode.

This doesn't make it easy to sort out. 
There may be no "right" way.
We are in the position of trying to find a solution that works for most users.

> Yeah, do it, I've never seen script THETA used in the upright style it in any
> manuscript, just use it as a default italic glyph, you have my vote.
>

But... in a manuscript all the letters would usually be script.
The (regular) face isn't meant to reflect manuscript, but plain typography -- as used in newspapers etc.

My question was regarding the "italic" (if that can be used for Greek).

The obvious solution is to just make the theta the same as the theta symbol.

Unfortunately, this would have a bad effect for users who are writing math.  There the association is clear: the main range theta is the upright one, while the "symbol" one is explicitly the other one.  Users writing math sometimes want a slanted upright theta, and very rarely use the curly one.

Right now, the two rho's are the same in italic.

I wonder if there's some way to sort of resolve this using styles or character variants.

> But I still insist on having script KAPPA as the default glyph, it's the
> latinate one thatr should go in the character variant.
>

I appreciate your preference, but this would be a bad idea for a couple of reasons.  The character variant should suffice.

> Lunate sigma is of no use for a general purpose font like yours, you'll find
> it only in Bodoni or Baskerville fonts (or Alexey's Old Style). Reproduction
> of archaic manuscripts.
>

The intent wouldn't be to reproduce Byzantine texts, but to represent them. Anyway, the alternative table doesn't do much harm.

Steve White <Stevan_White>
Group administrator
Tue 30 Jul 2013 10:26:32 PM UTC, comment #15: 

Yeah, do it, I've never seen script THETA used in the upright style it in any manuscript, just use it as a default italic glyph, you have my vote.

But I still insist on having script KAPPA as the default glyph, it's the latinate one thatr should go in the character variant.

Lunate sigma is of no use for a general purpose font like yours, you'll find it only in Bodoni or Baskerville fonts (or Alexey's Old Style). Reproduction of archaic manuscripts.

Rosella Capriotti <rosy58>
Tue 30 Jul 2013 09:29:10 PM UTC, comment #14: 

OK, I made more commits.
The curly beta and theta don't work as default replacements:
I put them under style set 5.
Character variant 00 now produces lunate sigma.
Character variant 01 now produces script kappa.

I'm not completely happy with theta: I wonder if the script form shouldn't be the normal form in italics, as with the rho.

Find xelatex file attached.

(file #28715)

Steve White <Stevan_White>
Group administrator
Tue 30 Jul 2013 01:50:17 PM UTC, comment #13: 

Noooo?  Noooo what?  What exactly are you objecting to?

I made the curled theta to appear at word beginngs and not elsewhere, as in Theano, as you suggested.  The only thing I did differently than Theano, was to not replace an isolated theta, which strikes me as wrong.

The beta now works as in Theano (but by default, not with a style).

I could make a character style to replace the kappa.

The impression I get from the essay I pointed to, is that each of these replacements is sort of a separate preference -- none is part of a grander "style".  This would mean to have separate styles for each replacement (which could be done), but that isn't what Theano does: the beta and theta replacements are in teh same style.

Steve White <Stevan_White>
Group administrator
Tue 30 Jul 2013 11:33:11 AM UTC, comment #12: 

Noooo please implement the theta I always typeset it with curled theta at the beginning, and please also devise a style which allows the user to have:

  • curled beta in the middle
  • script theta at the beginning
  • variant kappa always


This is really fine and beautiful Greek typography, the way I like it, can you get it for me please? :)

Thanks!

Rosella Capriotti <rosy58>
Mon 29 Jul 2013 05:45:19 PM UTC, comment #11: 

Some nice historical perspectives on these questions by Nick Nicholas:  http://www.tlg.uci.edu/~opoudjis/unicode/letters.html



Steve White <Stevan_White>
Group administrator
Mon 29 Jul 2013 03:11:23 PM UTC, comment #10: 

The logic in Theano is:

If style set (ss11) is turned on,
   * any beta except beginning a word changes to beta_curl
   * any theta at the beginning of a word changes to beta_curl

According to the Wikipedia page on Greek,

  • The symbol ϐ ("curled beta") is a cursive variant form of beta (β). In the French tradition of Ancient Greek typography, β is used word-initially, and ϐ is used word-internally.


  • The symbol ϑ ("script theta") is a cursive form of theta (θ), frequent in handwriting, and used with a specialized meaning as a technical symbol.


Note: doesn't say anything about theta as an inital form, as it appears in Theano.


In SVN:

I chose to implement these replacements by default  (could chang mind later.).  These can be turned off by turning off the 'calt' feature.

I don't like the replacement of the isolated theta, so I didn't implement that.

Steve White <Stevan_White>
Group administrator
Mon 29 Jul 2013 10:33:12 AM UTC, comment #9: 

Very well, I looked into the image you provided, and found it is linked from the page
https://en.wikipedia.org/wiki/Kappa
with no further explanation.  The glyph looks much like U+03F0, which has been in FreeFont for a long time.

Concerning the original report, I have now looked through the tables of the (very fine) TheanoModern font.  It has a lot of tables.  The purposes of some, I understand, some not, a few are questionable.

You seem to be particularly referring to the 'calt' lookup in that font, implemented as a contextual replacement, affecting only beta and theta.

As I read it, this replaces beta_curly and theta_curly with beta and theta, respectively (by some complex logic).  By itself, this doesn't make sense.  So... my guess is, this is meant to selectively "undo" the effect of a preceding style set, which replaces all beta and theta with beta_curly and theta_curly.
I don't think I would do it this way -- I'll experiment with it though.

Steve White <Stevan_White>
Group administrator
Tue 23 Jul 2013 09:18:57 PM UTC, comment #8: 

Is this kappa variant something other than Unicode u+O3F0 ?

Steve White <Stevan_White>
Group administrator
Tue 23 Jul 2013 06:04:09 PM UTC, comment #7: 

By the way, to make FreeSerif really competitive for quality Greek typesetting, you must also add http://upload.wikimedia.org/wikipedia/commons/9/9c/Greek_lowercase_kappa_variant.svg to the glyph repository.

Rosella Capriotti <rosy58>
Mon 22 Apr 2013 09:00:25 PM UTC, comment #6: 

Hi Steve:

Basically Theano Modern has a more "historical" looking style and has more stuff than Didot and OldStyle is even more weird and with all ligatures and stuff enabled is readable only by paleography experts.

So, if you're interested in fine Greek typography, you'd better stick for now to basic features of Didot, first and foremost Inital/Medial Beta and Initial/Medial Beta and Theta (elct ID on table page 19).

If you have some doubts about the pdf, please ask, I'm quite good in those paleography/manuscript stuff though I'm not a pro like Alexey, of course.

The difference between the sigmas is made by keyboard input, yes, I know they're basically the final and medial forms of one same character, but, obviously this happened before OpenType even existed, and considering that support in mainstream software is nonexistent even today, this is not a bad idea.

Actually, even the beta and theta variants were supposed to be typed (indeed as you can see they're encoded in the Greek range) but since there's no keyboard key assigned to them, we'll have to find OpenType solutions for them.

Regards


Rosella Capriotti <rosy58>
Mon 22 Apr 2013 08:24:36 PM UTC, comment #5: 

Hi

On Mon, Apr 22, 2013 at 9:05 PM, Rosella Capriotti

>
> Why TheanoModern? I thought TheanoDidot should be the most appropriate font to
> copy features from, TheanoModern's greek typeface is Bodoni which has more
> complicated and intricated features.
>

Yeesh.  Already TheanoModern has a lot of features (for a Western font).  I was only interested in the tables.  Do you think they are significantly different?

> I suggest you choose "contextual forms" or, a stylistic set to enable those
> variations.
>

OK I'll consider those options.  I also considered implementing it much like Alexey did.
Frankly, Alexey's explanation of these forms, which is several pages long, I found very difficult to penetrate.  It's plain though, this is a very -- lofty -- typographic point.. And also, that it would be better not to apply these at all than to apply them inappropriately.

Am I correct in understanding that the distinction between the medial and final sigma forms is usually made by user input, that is, by typing the different characters for the two forms?

Steve White <Stevan_White>
Group administrator
Mon 22 Apr 2013 07:05:45 PM UTC, comment #4: 

Hi Steve,

Why TheanoModern? I thought TheanoDidot should be the most appropriate font to copy features from, TheanoModern's greek typeface is Bodoni which has more complicated and intricated features.

I suggest you choose "contextual forms" or, a stylistic set to enable those variations.

Regards,


Rosella Capriotti <rosy58>
Mon 22 Apr 2013 02:13:07 PM UTC, comment #3: 

Rosella,

I'm familiar with Alexey's page -- he has been a contributor to FreeFont.

Looking at his TheanoModern font, I see

1) It doesn't use the 'medi' and 'fina' font features, which would make such things easy.  My observation too, is that many font rendering systems have enabled these only for Arabic (unfortunately--I would say, wrongly).

2) There is an inital form for beta, and a medial/final form for theta, using a complex contextual chaining substitution.

3) There is also a style set "contextual forms", but this seems to replace beta and theta with the alternate forms wherenever the style set is enabled.

4) There is also a 'salt' Stylistic alternatives table that maps most Greek letters that have variants, to their variants.

I do not find a table for automatically mapping sigma to its final form.  Why would that be?  I read on Wikipedia that formerly there was a rule like the one I mentioned for German, but that it isn't adhered to any more.

Steve White <Stevan_White>
Group administrator
Mon 22 Apr 2013 01:37:17 PM UTC, comment #2: 

Hi Steve!

Transformations are automatic and always applied.

The only free fonts who do this I know of are Old Standard TT and the Theano Family by: http://www.thessalonica.org.ru/en/

I attached the pdfs of these four fonts (the Theano Family comes in three versions: Didot, Modern and OldStyle, but I think you'll only be interested to Didot because other fonts have even more complicated variants).

Just go to the greek chapter of both pdf and you'll find the more exhaustive source of information about this topic I've been able to find out.

I hope you'll enjoy your reading.

Regards

(file #27925, file #27926)

Rosella Capriotti <rosy58>
Mon 22 Apr 2013 01:30:33 PM UTC, comment #1: 

Hi again Rosella!

This is again a rather bigger subject than it might first appear, but perhaps I can do something.

I was aware of the final form of the letter sigma, but frankly didn't know about the others.

Let me explain the technical details:
There are "font features" of OpenType fonts specifically to distinguish between a medial and final form of a letter. 
However, the font itself doesn't do anything: it is data.  The feature may be used by the font rendering software on your computer, or... not.  I will have to experiment.

I can tell you that in German, historically there were medial and final forms for the letter s.  These cannot be implemented by a font, unfortunately, because the form also depends on where the syllable is broken. 

Could something like this be the case in Greek?  Or are these transformations always applied, regardless of the letter's position in a syllable. 

Moreover, I've seen systems in which the medial-final features aren't enabled at all for Latin, because the systems' author thought they were exclusively for Arabic.

Can you point me to another font (preferably free) which behaves in the way you want?

Also, please point me to a web page that explains these Greek typographic issues.

Thanks!

Steve White <Stevan_White>
Group administrator
Mon 22 Apr 2013 12:17:42 PM UTC, original submission:  

Hi Steve,

I'd like to inform you about the need of contextual variants for Greek BETA and THETA (the first one is mandatory in fine Greek typography).

You already have those glyphs encoded at, U+0320 and U+0321. These are called curled beta and script theta.

Basically, the curled beta should occur in medial and final positions, while script theta in initial positions.

It would be good if you implement the substitution through alternates or a specific stylistic set.

Regards

Rosella Capriotti <rosy58>

 

(Note: upload size limit is set to 16384 kB, after insertion of the required escape characters.)

Attach Files:
   
   
Comment:
   

Attached Files
file #28715:  Greek-styles.tex added by Stevan_White (2KiB - application/x-tex - to be run under xelatex)
file #27926:  theano-specimen.pdf added by rosy58 (342KiB - application/pdf)
file #27925:  oldstand-manual.pdf added by rosy58 (565KiB - application/pdf)

 

Depends on the following items: None found

Items that depend on this one: None found

 

Carbon-Copy List
  • -email is unavailable- added by Stevan_White (Posted a comment)
  • -email is unavailable- added by rosy58 (Submitted the item)
  •  

    There are 0 votes so far. Votes easily highlight which items people would like to see resolved in priority, independently of the priority of the item set by tracker managers.

    Only logged-in users can vote.

     

    Follow 8 latest changes.

    Date Changed by Updated Field Previous Value => Replaced by
    2013-09-22 Stevan_White SummaryContextual variants for Greek THETA and BETA Alternative glyphs in Greek
    2013-07-30 Stevan_White Attached File- Added Greek-styles.tex, #28715
    2013-07-29 Stevan_White StatusProceeding Fix posted
    2013-04-22 Stevan_White StatusNeed info Proceeding
    2013-04-22 rosy58 Attached File- Added oldstand-manual.pdf, #27925
        Attached File- Added theano-specimen.pdf, #27926
    2013-04-22 Stevan_White StatusNone Need info
        Assigned toNone Stevan_White

    Back to the top

    Powered by Savane 3.13-d3ae.
    Corresponding source code