Fonts in PWG2013, forked from Issue #35

[funderburkjim]

At least two distinct issues were raised in Issue#35 regarding PWG2013 displays:

1. Web fonts
2. Size styling of Literary Sources

Since these two issues are nearly distinct (at least in implementation terms), this issue can focus on the first, Web fonts. Issue#35 can remain on the second item of size.

The purpose of this issue is to develop a consensus as to which choices of web fonts to use in PWG2013. Experimental versions of the basic display are made to allow us to see the effect of different fonts. Supposing a consensus develops as to which font choices are best, the aim is then to implement the consensus in four PWG displays (basic, list, advanced, mobile).

Marcis pointed out that the Old Indological fonts pertain to the non-devanagari text of PWG. The webtca display (http://www.sanskrit-lexicon.uni-koeln.de/scans/PWGScan/2013/web/webtca/indexcaller.php) works under the mistaken assumption that Old Indological Font pertains to Devanagari.

Version webtcb ( http://www.sanskrit-lexicon.uni-koeln.de/scans/PWGScan/2013/web/webtcb/indexcaller.php ) removes old indological from the choices of Devanagari fonts; it applies Old Indological font to all the non-devanagari text, including the text identified as literary sources.

Marcis also suggested that another font, CharterIndoCapital, be used within the literary sources. This font displays a string like 'Test' in all-capitals, but with 'est' in capital letters of a smaller-size.

Version webtcc http://www.sanskrit-lexicon.uni-koeln.de/scans/PWGScan/2013/web/webtcc/indexcaller.php ) does what webtcb does, and also applies CharterIndoCapital to the ls names.

At the moment, my favorite combination is (a) webtcc with (b) siddhanta font. In my tests, sanskrit2003 does not show svarita accent, and there is some odd vertical spacing in the non-Sanskrit at 'des Soma' under prARa (slp) headword.

The reason siddhanta is favored over praja font for devanagari is primarily since sidhhanta is free and praja is not. In some of the letter glyphs, praja is closer to the pwg text than siddhanta (intial 'a', retroflex 'n') , but praja is not identical (kz as in cakzus)

I am unsure whether Old Standard Indological and CharterIndoCapital fonts are free for our use.

Another detail regarding Old Standard is that webtcb,webtcc are using ony the 'Regular' ttf font; the italic one is not being used.

[gasyoun]

Even more amazing work done lately. I give you my thanks.

1) Old Standard Indological (based on http://www.fontsquirrel.com/fonts/Old-Standard-TT) & CharterIndoCapital are made by a colleague from Kiew of mine based on my request. Any license can be set. That should not be an issue.

2) "using ony the 'Regular' ttf font; the italic one is not being used" - we need to find out a solution, otherwise it's not similar to book, it's bad typography as well. So there must be a trick, but http://www.fontsquirrel.com/forum/discussion/82/using-bolditalic-properly-with-font-face/p1 seems not to be the one. So there should be 2 different CSS classes for normal and italics.

3) "sidhhanta is free and praja is not" - I surely agree. It's easy to read, free and has no flaws. If we speak about the digital version to copy as much of the original book layout, than we should use the French font https://www.dropbox.com/s/tl0mfxkmkjjz2zc/Gudakesa99.ttf , but as it's Sanskrit 1.2 encoding, so it would need some prior converting. I would propose we do it as a test. That way we could make the look and feel 80% close to the original. To make the font Unicode would take a few week of https://www.facebook.com/bayaryn life and he's the only person who might theoretically try to do it.

4) http://www.sanskrit-lexicon.uni-koeln.de/scans/PWGScan/2013/web/webtcc/indexcaller.php seems to be really nice, too bad the devanagari by default is Mangal. Because too many big letters make it hard to read, that's where Small Caps comes in handy.

[funderburkjim]

http://www.sanskrit-lexicon.uni-koeln.de/scans/PWGScan/2013/web/webtcd/indexcaller.php varies from the webtcc in two ways:

--1. It uses the "Old Standard Indologique-Italic.otf" font-family for italic text (css for 'i' tag is this font-family). There are indeed differences between the normal and italic forms of the Old Standard indologique font, and these differences closely agree with the differences between normal and italic text in the scan. Fantastic!

I saw that the sidhannta font is made by Mihail Bayaryn; Are these Old Standard Indologique fonts also made by him?

--2. A different way of styling the 'sizes' of text in literary sources was used; instead of font-sizes of 90%, 80%, 70% (as in webtcc), font-sizes of small, x-small, and xx-small are used in webtcd.

[gasyoun]

1) It's not fantastic, it's just the same font. Old Standard was based on Russian books on philology printed in late XIX century. PWK and PWK used it, as did all the other dictionaries of that time in Russia. The original website of the font is down for 2 years now, but the file is still at http://www.fontsquirrel.com/fonts/Old-Standard-TT - our file was not changed by Bayaryn. Bayaryn is from Minsk and works only on Devanagari fonts. He has made 3 different fonts himself and edited 1 more. 2) I'm no fan of the "small, x-small, and xx-small" idea. We should stick to the book. If I'm right, the smallest now is smaller than in the book. So absolute pt values might be even a better idea. 90%, 80%, 70% had no big flaws, did it? 5) Searched for "candra" Colebr. Misc. Ess. Ii, 163 (Xiv, 11). "Ii" -> "II" I guess, both should be equal size, second not smaller than the first one.

[funderburkjim]

1) Regarding 'It's not fantastic...' Just to clarify my use of 'fantastic': I was using this word as American slang, essentially equivalent to 'wonderful' or 'very good'. By 'fantastic' I did NOT mean that it was a 'product of fantasy', unrealistic, or any such pejorative sense.

Somewhere in the documentation for PWG (and PW) (a) mention should be made that web fonts are being used and (b) credit should be given to the creators of the web fonts used. From your link, and looking in the 'properties' of Old Standard TT, the author of the Old Standard TT fonts is Alexey Kryukov. Also, I see a 'SIL Open Font Licenses' file. How do you suggest the Cologne documentation be worded, for the Old Standard, the CharterIndoCaptial, and the Siddhanta fonts? (I have concluded that the best Devanagari web font to use is currently Siddhanta.)

Another detail: The names of the fonts previously downloaded were "Old Standard Indologique-Regular.otf" and "Old Standard Indologique-Italic.otf", which differ slightly from the names "OldStandard-Regular.ttf' etc from your above link; also file sizes are different, as well as file suffix.

2) Regarding 'small...' : Thanks for comment. Version 'webtce' same as 'webtcd' except for reverting to the 90%, 80%, 70% for sizes. The only 'flaw' in the 90%, 80% method was that, in one case which you mentioned and I investigated (see Issue 35, May20), the '80%' and '70%' font sizes resolved to 3.636px and 12.727px, and were visually indistinguishable.

You mentioned using abs. 'pt' or 'px' sizes as another possibility. If you would propose a specific pt size (or px size) for normal, and three smaller gradations, I can put the suggestions into another test display.

[gasyoun]

For Old Standard let it be as documented - Alexey Kryukov. SIL Open Font Licenses fit's the IAST part. CharterIndoCaptial - write Sergey Tkachenko. Siddhanta - Mihail Bayarin. Old Standard Indologique-Regular.otf contains diacritics that OldStandard-Regular.ttf does not, that is why the size is bigger as well. So it's a modified version. As per "visually indistinguishable" - oh, that solves the riddle. As per PT size - that would take some brainwork. Would it be enough for now?