Wed 02 Nov 2011 04:26:25 PM UTC, comment #25:
Well, not quite stylistic, since it entails a visual difference that enables the viewer to distinguish between two different characters (which is not the case in UCS Fonts), of course in order for this change to be consistent you'd have to change the macron shape everywhere not just in CYRILLIC I.
But, as I said, utilization of combining diacritics in Italic shape can be regarded as a typographical experiment at best, as an aberration at worst, at least as far as the Serbian case is concerned.
It's up to you.
|
Wed 02 Nov 2011 03:31:48 PM UTC, comment #24:
I understand you to be saying this is a stylistic difference.
In this case, I would leave the overline as a macron.
One could make an alternate argument in some cases that macrons ought to be of uniform length so as to be recognizible as the same mark.
|
Wed 02 Nov 2011 02:42:46 PM UTC, comment #23:
The information I've been able to gather points towards the uselessness of having Italic glyphs with combining diacritics. Such combinations should occur in texts or portions of text that are specifically technical in scope and therefore, not liable to typographical "embellishments" like cursive script.
So, the supposed identicalness of Italic I+macron and Italic PE (Serbian), should not bother us, BUT, if we care enough for it, DejaVu show there is a difference as you can see in the attached pdf. Typographically, DejaVu distinguishes between the actual combining macron placed above I which is shorter and "higher" than the line placed over PE.
(file #24275)
|
Sat 29 Oct 2011 10:43:12 PM UTC, comment #22:
Hi Alessandro
Are you sure the double-grave isn't used in any other language?
> I don't get it wrong. In any case, downsizing the accents would be a pretty
> good choice, what do you think about it?
>
Nope. That would look goofy. (There are a few unfortunate letters where such things become necessary -- but it looks goofy.)
Based on some advice here about the Hungarian umlaut:
Double Acute/"Hungarumlaut"
http://typophile.com/node/78623
diacritics project article
http://diacritics.typo.cz/index.php?id=24
I relieved the skew angles of the accents, giving the two graves slightly different angles, to make the whole mark less extremely wide. I think it helps a lot.
BTW: You might be interested to contribute to the Diacritics project: They are pretty Czech-centric, maybe you could add some Serbian content!
> deb package ...
It's your system. But I really don't recommend this.
If all things were as they ought to be, it should be possible to un-install the default FreeFont from your system. Unfortunately, it's tangled up with other packages (we're working on un-tangling that.)
We are messing with the distribution software, unfortunately. Just fooling it with a deb package won't make all the headaches go away.
But it is your time to spend. If you get it working, I'd be glad to hear how to do it!
|
Sat 29 Oct 2011 09:32:33 PM UTC, comment #21:
The double grave, as far as I know, is only used in Serbo-Croatian dictionaries (maybe you're confusing it with double acute which is used in Hungarian). The Latin glyphs from U+0200 to U+0217 were added in order to support typesetting Croatian dictionaries and FL literature, that's it, no other language uses double grave. (Serbian and Croatian are based on the same shtokavian dialect, so their prosody is virtually identical)
I think you should take a look at how the shape is realized in DejaVu, it seems to me the glyph is not exactly a juxtaposition of two grave accents, if I don't get it wrong. In any case, downsizing the accents would be a pretty good choice, what do you think about it?
deb package would be very useful in order to tell the system I'm installing an UPDATED version of the font. Since all these fonts (UCS, DejaVu) are bundled with Ubuntu, I can spare lots of headaches if the system overwrites by default the older versions, which otherwise can cause lots of conflicts in XeLaTeX as I've been experiencing of late.
|
Sat 29 Oct 2011 09:09:29 PM UTC, comment #20:
Hi!
I agree one of the inverted breve's was a little misplaced. Fixed now.
Moving the double-grave to the left won't help. I tried it.
I settled on a rule that the rightmost bulb on the double-grave is centered on the letter, as well as possible.
To my eye, this gives the best overall impression of centering for the glyph. From there, I have to make some adjustments before I'm satisfied, as over the A. The accent over the i is already as far left as it can go without colliding with other letters.
I tried other rules, such as centering the lower points of the accents on the letter. This had very bad results, and didn't look right at all. It is also difficult to arrange for a rule that doesn't make the small i look different and goofy.
It's a trade-off. Every approach has shortcomings.
These things are very subjective, too.
Well I don't like the double-grave accent either. I think it was a poor idea to begin with, typographically.
And it may be best to make its dimensions less extreme...
Of course it is now just two grave accents, as seems appropriate.
However this accent is heavily used in other languages. It may be their convention that it should look like two grave accents together.
So I would have to study the situation.
I don't know what you mean by the deb package. What exactly do you
hope to achieve with that?
I assure you, writing a shell script to copy the font files and clean your cache is far easier than packaging up a deb package, and then installing it. Vastly easier -- this isn't what the packaging system is for.
I don't think the Mono is substantially different from the last
release, for your purposes. Nothing we have done would have affected it. So just use the old versions.
I'll keep thinking about altering the accents though.
Cheers!
|
Sat 29 Oct 2011 07:54:08 PM UTC, comment #19:
Now that I have all the fonts I need I did some testing. Here are the results:
A Series: The Serif Regular DOUBLE GRAVE & +INVERTED BREVE is placed a bit too much on the right.
E Series: The Serif Regular DOUBLE GRAVE is placed a bit too much on the right.
I Series: The Serif Italic INVERTED BREVE is placed a bit too much on the right.
O Series: The Serif Regular DOUBLE GRAVE is placed a bit too much on the right.
ER Series: OK
U Series: OK
Conclusion: the DOUBLE GRAVE gives lots of trouble, the best thing would be to redraw it, its current shape is not satisfactory, as far as I'm concerned.
P.S.
As far as the deb package is concerned, I think I should for the Mono version, what do you think?
|
Fri 28 Oct 2011 12:23:18 PM UTC, comment #18:
Dear Steve,
If I'd known how to handle an html, I'd never have bothered you in the first place. But if you tell me how to do so, I'll be happy to do it by myself :)
|
Fri 28 Oct 2011 12:04:35 PM UTC, comment #17:
Alessandro,
I think you can easiy modify that HTML page to test all these letters. All the accents should be there. Let me know if I'm wrong!
Cheers!
|
Thu 27 Oct 2011 10:41:09 PM UTC, comment #16:
U+0410 CYRILLIC CAPITAL LETTER A
U+0415 CYRILLIC CAPITAL LETTER IE
U+0418 CYRILLIC CAPITAL LETTER I
U+041E CYRILLIC CAPITAL LETTER O
U+0420 CYRILLIC CAPITAL LETTER ER
U+0423 CYRILLIC CAPITAL LETTER U
U+0430 CYRILLIC SMALL LETTER A.
U+0435 CYRILLIC SMALL LETTER IE
U+0438 CYRILLIC SMALL LETTER I
U+043E CYRILLIC SMALL LETTER O
U+0440 CYRILLIC SMALL LETTER ER
U+0443 CYRILLIC SMALL LETTER U
I've only seen a sample of U+0438 CYRILLIC SMALL LETTER I with diacritics, what about the rest?
|
Thu 27 Oct 2011 10:31:04 PM UTC, comment #15:
Hi Alessandro,
I don't know what you mean by "posted".
We have only discussed (Serbian) Cyrillic. To the best of my understanding, the job is done. What other letters do you mean?
|
Thu 27 Oct 2011 10:18:13 PM UTC, comment #14:
The CYRILLIC I series is OK, as I've told you. But, if I don't get it wrong, you've not posted other series (i.e. letters+combining diacritics) yet.
|
Thu 27 Oct 2011 09:57:29 PM UTC, comment #13:
Hi Alessandro,
So are we tentatively finished?
Could you have a quick look through all the faces, and tell me if everything is ship-shape?
|
Thu 27 Oct 2011 06:23:48 PM UTC, comment #12:
:) Gosh, you're so right! Such a stupid question for me to ask!!! That's my area of expertise, not YOUR!
By the way, I'm asking around, but I confess I never paid attention to that. Actually those diacritics are never used in normal Serbian typography, but just in dictionaries and learners' literature, in order to display prosody. It's a bit like the IPA, so it's kind of strange if not actually improper to use them in Italic, did you ever see an IPA notation in Italic? I did not. I suppose the same should apply to combining diacritics, they'll practically never be utilized with Italic glyphs, but I'm going to tell you if there is a minimal, but still meaning difference between the Italic Serbian PE and Italic I + MACRON.
|
Thu 27 Oct 2011 04:52:57 PM UTC, comment #11:
Alessandro,
This is your area of expertise, not mine. You tell me.
|
Thu 27 Oct 2011 12:12:16 PM UTC, comment #10:
Of course, but what if the Letter I too has a macron on it? Remember, among the combining accents there is U+0304 COMBINING MACRON.
|
Thu 27 Oct 2011 11:42:37 AM UTC, comment #9:
Hi Alessandro,
As near as I can tell, it is standard in Cyrillic typesetting for the italic (cursive) form of 'ii' to look like a Latin 'u'. Many examples showing this can easily be found, even the Wikipedia page for Cyrillic Alphabet.
I thought this was why you put the macron over the Serbian 'pe', to distinguish it from the 'ii'.
What is the problem?
|
Wed 26 Oct 2011 03:17:15 PM UTC, comment #8:
1) OK
2) No, no need for that. Just drop it.
3) In the html the first two (grave and acute) are OK in Serif, the rest (other three accents) in Serif and the whole Sans line has the accents too much on the right. But that's the original version (stored on my PC), while the png looks very good.
4) No, wait, I checked the pdf, it's correct. Maybe you didn't understand why Latin letters were there in the first place.
It's the Cyrillic letter A plus various combining accents compared with Latin precomposed characters that should look exactly like Cyrillic+combining. I did it in order to help you positioning the accents above the characters.
Big trouble ahead: is it my impression or Italic CYRILLIC LETTER I+MACRON looks like Italik CYRILLIC LETTER PE?
|
Wed 26 Oct 2011 03:03:05 PM UTC, comment #7:
Alessandro
1) All the marks were there, but in the SansOblique faces something had gone wrong, and they were shifted.
This is now fixed, as well as a small inconsistency in SerifBold.
Attached is a picture of a Cyrillic letter with the marks, as rendered in Firefox in Ubuntu.
2) Am I to understand that there is no need for the mark U+030B
In Cyrillic? If so, I'll remove it -- these things get cluttered quickly.
3) Are you still seeing a problem? It sounds as if you are... Once again:
which face?
which letter?
which mark?
4) The sample you sent refers to latin letter a. I understood this was for Cyrillic.
I hope that's just an oversight.
(file #24207, file #24208)
|
Wed 26 Oct 2011 10:55:42 AM UTC, comment #6:
:) I didn't mean to accuse you of following Microsoft fonts (which apart from the new ClearType collection, is pure crap).
Actually Times New Roman doesn't even support combining accents, it just display the accent after the would-be accented character.
First of all, read U+0311 INVERTED BREVE instead of U+030B DOUBLE ACUTE. The double acute is a Hungarian accent which has nothing to do with Serbian.
Then, take a look at this pdf, you'll find a comparison between Cyrillic combinations and Latin precomposed glyphs that should look exactly the same, but somehow they don't because combining accents are badly placed over the Cyrillic letter.
I hope it'll help.
(file #24206)
|
Tue 25 Oct 2011 10:05:16 PM UTC, comment #5:
Which face, which letter, which accent, and in what sense not natural? Give me at least one example please.
It does not help to accuse me of following Microsoft fonts. I didn't even look at any other fonts when I placed the accents.
|
Tue 25 Oct 2011 08:30:44 PM UTC, comment #4:
Well, at the moment I can't specify anything since as I told you, TeXWorks refuses to process anything. But, I could clearly see that as in Times New Roman, the accents were badly placed over the letter, the feeling was not "natural". There was surely the need of some refinement.
|
Tue 25 Oct 2011 08:20:40 PM UTC, comment #3:
Sorry, it seems I only committed the Sans faces.
Now the Serif faces are committed as well.
If problems remain, please specify: which face, which letter, which mark.
Thanks!
|
Tue 25 Oct 2011 07:43:46 PM UTC, comment #2:
The question is: are they correctly placed over the letter? I did some tests when TeXWorks still worked and the answer seemed to be negative.
Palatalization mark is only used in Church Slavonic and not in Serbian.
|
Tue 25 Oct 2011 07:29:47 PM UTC, comment #1:
Hi Alessandro,
In Serif, mark anchors needed to be added only for macron and the inveted breve, In Sans, only the medium face had anchors. Fix in SVN.
The sans-serif faces need some work yet.
You know there are several other Cyrillic marks, including a paletization mark, U+0484.
|
Tue 25 Oct 2011 04:42:55 PM UTC, original submission:
Serbian uses five diacritics:
U+0301 COMBINING ACUTE ACCENT
U+0304 COMBINING MACRON
U+030B COMBINING DOUBLE ACUTE ACCENT
U+030F COMBINING DOUBLE GRAVE ACCENT
U+0311 COMBINING INVERTED BREVE
To be correctly positioned above:
U+0410 CYRILLIC CAPITAL LETTER A
U+0415 CYRILLIC CAPITAL LETTER IE
U+0418 CYRILLIC CAPITAL LETTER I
U+041E CYRILLIC CAPITAL LETTER O
U+0420 CYRILLIC CAPITAL LETTER ER
U+0423 CYRILLIC CAPITAL LETTER U
U+0430 CYRILLIC SMALL LETTER A.
U+0435 CYRILLIC SMALL LETTER IE
U+0438 CYRILLIC SMALL LETTER I
U+043E CYRILLIC SMALL LETTER O
U+0440 CYRILLIC SMALL LETTER ER
U+0443 CYRILLIC SMALL LETTER U
Sorry, but, at the moment I can't provide you with a pdf sample.
|