As for why you seem to be unable to change the month terms in your citations, you'll need to give us some more information about what you've been doing thus far. Are you redefining the month terms in the CSL style, or in the CSL locale? Which style are you using? (if you modified it, please copy/paste it to http://gist.github.com/ and give us the link)

Oh, and one more thing, apparently either my changes to locale file or to Chicago, or to both finally worked and I forced Zotero to use DD.MM.YYYY format instead of full month name. It's probably a workaround rather than clean Chicago, but hey, it was not meant to cover languages other than English, so no wonder it doesn't always fit all the needs.

And the bottom line is that, apart from isolated cases, most professors here in Poland set one basic rule of thumb: in the end it doesn't really matter which citation system you use as long as you're consistent.
Cheers

And does Polish really not use short month names at all? If so, you could probably repurpose the short terms to cover the case "maj 2011" (month + year), making appropriate changes to the CSL. This is of course not a perfect solution, but it'd work.

@ Ajlyon, Czech and Polish grammars are really similar (except for Czech tenses being compound and Polish being simple; jsem byl vs. byłem), so in most cases the base is similar. What differs is the tradition and evolution of our languages.

Of course you are right that in case of pages the problem is obscured by abbreviation. Furthermore, I'm not sure about csl or xml in general, but I have had some experience with gettext files (translating .php themes using built-in gettext engine) and it seems that it handled Slavic numerals just fine, using some sort of matematical formula that defined when to use which form. A great example: a comment in Polish is "komentarz". In English the plurals are formed by adding s. As simple as it is: 1 comment, 2 comments, 3 comments and so on. In Polish (and most if not all Slavic languages), there are different endings for 1, 2 to 4 and 5 to infinity. Using gettext I was able to point php files to a correct form. Take a look at any Polish-language Wordpress-based blog: it features a counter saying 1 komentarz, but 2 komentarze and 10 komentarzy. Something similar could be achieved in Zotero if need be.

But is it really needed? IMO in the case of singular-plural for pages - no. However, in the case of, say columns - there is a problem. The abbreviation "kol." is already reserved in Polish for "kolega" (meaning colleague). So we can't create artificial abbreviations, we should use either "kolumna I" (column 1), kolumny II-III (columns 2-3) or "na kolumnie 15" (at column 15), depending on the context and citation style. Not a perfect example, but I can't think of any better right now, sorry.

As to genders, inflection and such, computers are English-centric, so short-cuts are acceptable in modern Polish publications (such as "zaprezentowano" I mentioned before: it could be translated as "was presented", without specifying the subject ("it") as it would require a gender-specific suffix). Another example of such short-cut used in citations is the term for access date (as in "this web-page was accessed on January 2nd, 2003"). It should be "udostępniono/udostępniona/udostępnione" (gender and numer-specific forms), but "dostęp" (literally: access) is often used. It is acceptable nowadays, which doesn't make it look less awkward though.

Oh, just noticed one more problem with English vs. ROTW: what has only one name in English could have many context-dependent names in other languages. Take the word "book" for instance. Normally (in the case of "book as a reference" instead of "chapter" or "web page"), it should be translated as "książka". However, when talking of, say, Book of Job from the Bible, or Book VII of Iliad, the correct term is "księga". So what is one string for the English translation, should actually be two strings for the Polish, both context-dependent.

As to short forms - do you mean "Jan, Feb, Mar, Apr," and so on? Yup, these are used, but I can't remember seeing such abbreviations anywhere outside of calendars. Certainly not in academic citations. When by "short form names" you mean names in nominative case, then Polish does use them. I believe I already mentioned (here or in the other thread), the reference could be:

Doe, John (1 stycznia 1945) "Blahblahblah" - the month name is in Accusative
but
Doe, John (styczeń 1945) "Blahblahblah" - the month name is in Nominative

The latter case is fine as long as the day is missing. So in theory the latter form is needed. In practice one rarely see such references in Polish publications. Partly because we have no tradition of citing daily publication date for anything except for daily newspapers and weeklies, usually year is enough. And parly because if it is a monthly, its' number is enough to tell which month it was.

@fbennett - sure, I figured I could hard-code Polish usage into the files. So my personal Chicago is no longer compatible with English usage (01.02.2003 date format instead of 1 February 2003, for instance), but it works fine for my needs. One has to bear in mind however that translating Zotero into different languages is a tad more complicated than "take one term and check its' meaning in Polish".
Cheers

It seems all (ALL!) of the problems the validator mentions are inherited from the original file as they refer to untranslated lines (ed., eds. and such). Didn't I mention that I marked untranslated lines with a hash? No wonder it doesn't validate nicely if you don't remove the hashes... :)

Also, some of the problems seem to be related to the fact that Zotero tries to be less English-centric than the CSL. Hence the document has fields for more ordinals than 4, while CSL document lists additional ordinals as errors. Just remove the lines after
"term name="ordinal-04" and it should validate nicely. As long as Zotero uses ordinal-04 the English way (adding th to anything above 4), it should work just as well.

Anyway, I'm not an expert on xml or coding or whatnot, just an educated guess. From my perspective the document is okay as it works, which is all that matters to me :)

@Halibutt: sure, but fixing up other people's XML isn't exactly my greatest hobby. If you can prepare a validating version, I'll upload it to the code repository so it ends up in the default installations of Zotero.