User:Jukeboksi/Blog/June2004

< User:Jukeboksi‎ | Blog
Revision as of 20:00, 18 June 2004 by Jukeboksi (talk | contribs) (no. not getwiki)

15.6.2004

Today I downloaded MediaWiki 1.3.0beta3 and installed it on my iBook to try out the XML-import facility by commenting out the lines that disable the code from running. I have to say that the way it works by asking you to upload a downloaded xml-dump is really inadequate for our future needs.

Right. The GetWiki way is the correct way: direct to another web address and tranparently mirror the material as is, including all changes, until there is a need to fork just that one page. However there are ideological reasons and Wikimedia corruption reasons why you will not ever get the MediaWiki coders to do that.
No. The right way is to recruit PHP coders to make an import feature that works for us. I know very little PHP, I can hardly read the SpecialImport.php that contains the code for the Import feature.
  • What we would need is a myriad of wikis with radiobuttons to choose from which wiki to import and do it live, not manually saving and uploading the file.
Yes, absolutely. Being able to import from a different wiki per article is ideal. At least a different wiki per namespace!
  • Also importing histories failed
That's unacceptable given GFDL, but GetWiki doesn't do that right either. Also when Wikipedia exports it give you only the last edit comments not the whole list of editors you'd need to do GFDL attribution to the legally required standard.
  • Actually whole import thingy failed even though it reported success for uploading files without the history parts.
Disaster!
Yes, you see this issue all the time with redirects from Meta-Wikipedia to Wikipedia, and probably it's also required for Simple English to full English: rather than seeing NO article in Simple, it would make sense to see the Full one. Rather than seeing no article in full where there's one in simple, it would make sense to mirror it.