This page is about the development and administration of the Hebrew MoinMoin wiki.

See also and /Planning

Most of the issues related to a Hebrew wiki are discussed on MultipleLanguagesSupport. This discusses some specific Hebrew issues.

Unreadable Hebrew File Names

Problem:

Hebrew file names on the disk are unreadable. Its very hard to delete a page by hand using the file system.

Hebrew file names on disk use the quoted form. The page name "פתיחה" is saved on the disk as "(d7a4d7aad799d797d794)". This gibberish is a compact representation of the name in utf-8 encoding. This quoting is about 50% more efficient than the quoting used in 1.2, but unfortunately is not more readable.

Since moin does not support real sub pages yet, the complete name (e.g page/subpage/subsubpage) must be shorter then 254 characters. Current encoding (xxxx...) allows full page names of up to about 60 characters, in the worst case. Using common quoting like %xx%xx would allow much shorter names. When moin will support real sub pages, this limit will be removed and then we can use simpler encoding.

On certain platforms like Mac OS X, we don't need any encoding or need to encode only very few characters. But such wiki will be hard to move to another server that does not use same file system. Maybe in the future when utf-8 become the standard, this problem will disappear.

Solution:

All operations except for physically deleting a page directory can be done from the wiki. This operation is very rare and usually not needed. The simplest workaround is to rename them from the web interface to 'Trash/…' and then delete all pages starting with 'Trash/'.

Hebrew Page Links

Problem:

Certain browsers have transparent Hebrew support, and display Hebrew page URLs in plain Hebrew, like this: http://localhost/mywiki/עברית_קשה_שפה. But when pasting such link in into another site or an email message, the link may not work.

Other browsers, display the raw URL: http://localhost/mywiki/%D7%A2%D7%91...D7%A4%D7%94. These links can be safely pasted anywhere on the web or in email message.

Unfortunately, If you use the Hebrew link, it might not work and if you use the quoted link, nobody can read it. The joy of the web :)

There is a new standard for international URLs, but its not well supported currently. In a few years from now this problem will be gone.

Solution:

An effective solution is creating a redirect page named DifficultHebrew, in this form:

#redirect עברית קשה שפה

Now you can link to the Hebrew page using a standard readable URL: http://localhost/main/DifficultHebrew

TitleIndex

Problem:

Hebrew is sorted last, instead of first.

Solution:

None

Discussion:

We should sort the user language first, then other language.

Hebrew Links

Problem:

CamelCase will not work, because Hebrew does not have capital letters.

Solution:

Use extended links ["Extended Links"].

Discussion:

Perhaps חיבור_בקו_תחתון_של_מילים_עבריות can be made to work like CamelCase in English? Those automatic links are a really nice feature of wikis, and I see no other use for connecting hebrew words with underscores.

Extended Links are easier to type, specially for names with many words, and look better in the source. We don't plan to add this_kind_of_markup currently.

WordIndex

Problem:

WordIndex builds the index from WikiWords, since Hebrew does not uses them, Hebrew names are not in that index.

Solution:

None

Discussion:

Fix WordIndex to split by WikiWords or spaces, like the LikePages macro action does.

Hebrew and English Font Size

Problem:

This is a special problem with Hebrew, because Hebrew letters are usually smaller than English letters. If you want to use both Hebrew and English, you need to set the Hebrew font size to 110%-120% of the English font, or the latter will look much bigger.

Solution:

None

Discussion:

There should be a way to set all Hebrew text to be 110-120% larger then the English text - everywhere - including macros output.

This can be done using CSS 2 selectors like:

*[dir=ltr] p {font-size: 0.9em;}

This is a minor problem and does not happen with all fonts. Fix it sometime in the future. --NirSoffer

Mixing Hebrew and Non Hebrew Text

Problem:

Anywhere RTL and LTR text is mixed is usually broken visually. The problem exists in RecentChanges, in search results that contain both Hebrew page names and English page names, and possibly in other places.

Solution:

None

Discussion:

This is a major problem, and I don't know if we can create code that eliminates this problem automatically. In a word processor, you can fix this situation by setting characters direction. This can be done today by the ltr/rtl macros, but not always.

There is a partial solution using RLM and LRM characters. It works with other Hebrew wikis, but the characters are not?? invisible?? in Safari on Mac.

MoinMoin: HebrewSupport (last edited 2007-10-29 19:06:51 by localhost)