Wikipedia 10K Redux

Reconstructed by Reagle from Starling archive; see blog post for context.

Free_Links

You can now force a link over arbitrary text by putting it in brackets, like this: Alabama.

Also, something I just now learned from reading the source code, you can link to ISBN's like this: ISBN: 0887308856


Excellent! Now, is there a way to automate the conversion of old-style links (and pages) to new-style links? I seem to remember Cliff saying something about doing this. For example, I would like to convert links that look like LarrysText to Larry's Text; then it would be cool also to convert the pages themselves, in this case to Larry's Text (since apostrophes are not allowed in page names). What are the chances of this happening anytime soon? -- Larry Sanger


I am all for consistency. We should henceforth use free linking. We should also get the above question answered...that could save us a lot of mind-numbing gruntwork. -- Larry Sanger


Why doesn't my Free Links work correctly. Is it because they contain non-ASCII characters? Look at Sweden for Göteborg and Malmö.

LinusTolke


Fixing the non-ASCII character problem should be relatively easy. A similar problem has already been solved in the world of IMAP in the i18n of mailbox names, and the solution was to use modified UTF-7 encoding. For a complete explanation, please see RFC 2060, section 5.1.3. The short version is that M-UTF-7 uses US-ASCII representation for US-ASCII characters, and "shifts" into modified BASE 64 for all other charsets (UNICODE 16-bit octets).

Although I don't have a library routine I can whip out for you, I'm sure you can find one somewhere to hack up for WikiPedia. - Flavor


Talk

See WikiPediaBugs for more.