Crispy on 23/5/2006 at 08:51
Gah. I'm sorry for duplicating work, but it's difficult to figure out what's been done already when it's all in different places. Perhaps I should just stay out of it then.
scumble on 23/5/2006 at 11:14
It's not the end of the world, Crispy - duplication is probably unavoidable and it can be sorted out fairly easily, so don't give up - just make use of the search function more. You may prefer to expand on some of the articles or perhaps provide some useful screenshots. The wiki will be much more useful and improve the sharing of knowledge if it's updated properly. I find it useful to just go over what I've done for later reference, as I often find I forget exactly how I did something.
veledan on 31/5/2006 at 18:49
I may be able to help with scripting the transfer of material if someone has a list of what needs doing so I don't blindly duplicate stuff. I have written bots for wikipedia (same software) using the (
http://meta.wikimedia.org/wiki/Pywikipediabot) pywikipediabot framework - I'll have a play with it to see if it works well on this wiki.
Duplication of pages is avoided on wikipedia by creating redirect pages by the way, and it would probably be a good idea to use them here. When you create a page on any topic, you can also create pages for all its other likely titles, which automatically redirect to your new article. So you might write an article [[Flare]] and then redirect [[Flares]] and [[Flarelight]] to it.
Redirect pages also help searching and they prevent broken links in other people's articles too: for example, if I were to write a different article and link to
Flares instead of
Flare by mistake, the redirect page would mean my erroneous link worked OK.
It's done by creating the page as normal and making its contents:
[INDENT]
[/INDENT]
substituting
Flare of course. The mediawiki software does the rest.
I'm (
http://en.wikipedia.org/w/index.php?title=User:Veledan) en:User:Veledan on wikipedia.
UPDATE: It works!
scumble on 1/6/2006 at 07:38
I hadn't come across that, veledan. That would be useful to move the remaining pages, but the bot would have to do the following:
* Grab the old page text from the Wikka wiki
* Put spaces between words in the titles
* Execute a regexp that could convert most of the markup
* Check for an existing page with the same name
* Create the new page
Not necessarily in precisely that order though. If you can come up with something, it would be a big help as I'm a bit short of time at the moment. Use your user page to make some notes so I know what you've done - at some point I'll think about using bots to clean up the content further.
It would be a good idea for us to register bots later on (if they're used for housekeeping tasks), but it would probably not matter if the recent changes page had all the newly created pages on it.
veledan on 1/6/2006 at 15:39
I'll take a look at the existing format and have a stab at it. I'm going to try to make an interactive bot to transfer pages one at a time. I agree the wiki isn't really suitable for fully automatic transfer. I have been away from this forum for a few weeks but I remember thinking there were quite a few old wiki pages that covered several topics and that could do with being split up.
As for other tasks, I think pywikipedia will come in most handy when lots of pages have been loaded. There are pywiki scripts for most common organisational tasks: for example there's one that will go through all uncategorised pages, and you can add each page to one or more categories with just two keypresses. Another performs bulk uploading of images.
I don't expect the wiki will ever be busy with bot writers :erg:, but I will certainly post details on my userpage of exactly what I'm up to before running any bot on the wiki. If you or anyone else wants to join in and write bots in python, we could create [[FleshWorks:Bots]] :D
UPDATE: I made a start on the regular expressions (
http://ttlg.com/wiki/index.php?title=User:Veledan/Sandbox/Regexes) here but given the fact that those regexes only cover 10% of the old wiki formatting, and given that regexes can be tough cookies to debug I'm beginning to agree that manual transfer might be quicker :erg:
scumble on 7/6/2006 at 06:22
The limited experiementation I did with some regexes in jEdit seemed to be working quite well. If you cover bold, italic and the lists that would be enough, don't worry about trying to add in all possible markup.
