I've started this article to document the prevalent usage of Cyrillic in Wikipedia. Please correct or add as necessary. —Michael Z. 2005-12-8 08:46 Z

I just had a look at this discussion page in a few web browsers and found some problems, so I did some more quick testing.

On this discussion page:

• MSIE: k-acute and g-acute used for Macedonian, and the modifier-letter accents draw as boxes, the z-tie-bar used in ALA-LC transliteration is rendered as an uppercase A-grave
• Firefox: the centre dot used in BGN/PCGN transliteration for Ukrainian is rendered as a figure 6, the g-grave used in ISO 9 for Belarusian/Ukrainian ґ is rendered with the diacritic too far to the right
• Safari: g-grave is followed by the accent, when italicized followed by a box
• Safari with Lucida Grande font applied in style sheet: only italicized g-grave fails

In the table in ISO 9:

• MSIE: 28 Cyrillic letter pairs, 20 Latin letter pairs, two accents, and the palochka fail to draw correctly
• Firefox: 12-1/2 accented Latin letter pairs draw as letter followed by a box
• Safari: 12-1/2 accented Latin letters pairs draw as letter followed by the accent
• Safari with Lucida Grande font applied in style sheet: no problems

In the table in Romanization of Ukrainian:

• MSIE: ALA-LC tie bars fail to render, two as different accented letters, three as boxes
• Firefox: g-grave has the accent after the letter, the double-dagger reference link renders as a letter a
• Safari: ALA-LC tie bars render too wide, but readable, g-grave is followed by accent
• Safari with Lucida Grande font applied in style sheet: no problems

[Setup: MSIE 6 on vanilla WinXP; Safari 2.0.3 with a Wikipedia user style sheet specifying lucida grande font; Firefox 1.5.0.1 with default styl sheets; latter two on Mac OS X with lots of extra international fonts.] —Michael Z. 2006-02-07 04:14 Z

I can't find any guidelines in the MOS on how to cite foreign-language sources.

Virtually the whole English-speaking world uses Library of Congress transliteration in their catalogues. North American libraries do, of course, but also the British Library for acquisitions since 1975, and Libraries Australia. (Consequently, this is used in most bibliographies in English-language academic publications worldwide, too. Oxford once used the British Standard, but now mentions L.O.C. first in their style manual: “The transliteration system used by the US Library of Congress is in wide use, spurred on by the development of information technology and standardized cataloguing systems.”; Ritter 2002, Oxford Guide to Style, p 350) There's no reason to offer bibliographic transliteration in any other format.

Any objection to recommending this? —Michael Z. 2009-02-07 17:36 z

My concern is regarding the two-letter tie characters, which don't show correctly in most browser configurations I've seen and require quite a bit of shamanism to fix. Any simple way around that yet? I'd be hesitant to recommend the LOC system without diacritics and the ties.

Also, what bibliographic information would this proposal affect? The LOC system is used when it is easier to show the transliteration instead of original Cyrillic, but we don't have this problem in Wikipedia—we can simply show the original Cyrillic title and, if necessary, its translation. It seems to me it is going to be more trouble to make sure that proper LOC transliteration displays correctly instead of just showing the original Cyrillic. Thoughts?—Ëzhiki (Igels Hérissonovich Ïzhakoff-Amursky) • (yo?); 14:57, February 9, 2009 (UTC)

I'm mainly talking about titles in article “References” sections, but perhaps also when they appear in the main text.

Come on, we don't use transliteration because it is easier to type, we do it for the reader. Most library catalogues and bibliographies in English-language sources have these names in LOC transliteration only. Not only should we be adding transliteration to make the titles accessible, but the format we use should be compatible with what the rest of the anglophone world. —Michael Z. 2009-02-09 17:08 z

Have a look at Holodomor and Republic of Novgorod, and any other article with a lot of foreign-language references. The majority proportion of titles and authors names are in LOC transliteration already, without the tie bars.

How would we clean this up, ideally? Seems wrong to convert transliteration to Cyrillic, when this would make them less accessible to most editors, and more difficult to compare with bibliographic data elsewhere. Seems useless to put energy into converting LOC to BGN/PCGN, which would have the same effect.

I'm okay with leaving out the tie bars until this improves technically. Editors are unlikely to enter them anyway, but the spellings would remain the same. —Michael Z. 2009-02-09 18:02 z

Sorry, I don't quite buy the argument about the LOC transliteration being more accessible to readers. It was always my impression that the LOC transliteration was devised primarily to handle the library catalog cards because it was a major pain to work with non-Latin scripts directly (there was no Unicode nor even computers when the system was put in place). Come think of it—if you don't know, say, Russian, the LOC transliteration is going to look just as gibberish to you as original Russian. It may be easier for you to find the actual book using the LOC transliteration, but why would you want to do that if you can't read the language in the first place? And if you can read the language, what difference does it make to you whether our reference is in Russian or in LOC translit? All in all, I am not against using the LOC system in references in addition to original title (as you rightly noted, it sure makes more sense than BGN/PCGN); I just don't see the point. We are not a pre-1990s library with DOS PCs, and we don't have the same technical difficulties with non-Latin scripts. In my opinion, if someone wants to add LOC translit to references, let them go ahead and do it, but I just don't believe we should be explicitly recommending this practice.—Ëzhiki (Igels Hérissonovich Ïzhakoff-Amursky) • (yo?); 20:17, February 9, 2009 (UTC)

Why do English-language books have Cyrillic sources at all? The fact is that these will get cited, and the other fact is that most readers can't process Cyrillic text. I notice that some library systems are starting to add native-script search capability, but the hundreds of thousands of titles and authors catalogued in the last century, and possibly in the next, will remain available in LOC transliteration.

At the very least, editors shouldn't look at this guideline and interpret it as saying that references copied from their books should be converted to BGN/PCGN or whatever. I'd like to mention that most bibliographic information will be found in LOC translit, and it shouldn't be converted when editors enter it. —Michael Z. 2009-02-09 21:13 z

I'd still say that having an ISBN number is more important than LOC translit. If one really needs to find the book being cited, have them click out to the ever helpful Special:BookSources page to find whatever meta data they are looking for, including LOC translit if one is available. That's what the ISBN service page is for—to aggregate the meta data of all sorts so we wouldn't have to overload our references with them. Having author/title information in native script and possibly their translations is all that a reader needs for quick reference; and if they need more, shove them to the ISBN page and let them go wild there. The more metadata we add to the references, the more maintenance they require, and the more bloated they look. Why add extra work resulting in very marginal gains?

Now, for books that don't have ISBNs, that's another matter entirely...—Ëzhiki (Igels Hérissonovich Ïzhakoff-Amursky) • (yo?); 21:40, February 9, 2009 (UTC)

All good points, especially about avoiding bloat. Still, there's no time limit or budget for this little project of ours. —Michael Z. 2009-02-18 06:50 z

I'm stubborn. I'm going to add an advisory note about references. —Michael Z. 2009-02-18 06:52 z

Stubborness is not a virtue; I happen to know that firsthand :) Anyway, give it a try and see what happens, although I am willing to bet all the buttons on the shirt I am wearing that this advisory note will largely be ignored :) Cheers,—Ëzhiki (Igels Hérissonovich Ïzhakoff-Amursky) • (yo?); 15:50, February 18, 2009 (UTC)

I have proposed an update of the Ukrainian Romanization conventions at Wikipedia talk:Romanization of Ukrainian#Update to the National 2010 system. —Michael Z. 2013-05-02 02:25 z