On Lee's test site (http://www.piclab.com/wikitest/wiki.phtml), I have patched the stub detector so that /when the stub detector is active/, it will also show links to redirects with a yellowish background.
Should I use that on the live software 1. not at all 2. only when the stub detector is activated 3. for all logged-in users as an user option, off by default 4. for everybody
The most simple implementation is #2, and it doesn't require an additional database query. That's what I will do if I don't here good arguments against it ;-)
Magnus
Magnus Manske wrote:
On Lee's test site (http://www.piclab.com/wikitest/wiki.phtml), I have patched the stub detector so that /when the stub detector is active/, it will also show links to redirects with a yellowish background.
Should I use that on the live software
- not at all
- only when the stub detector is activated
- for all logged-in users as an user option, off by default
- for everybody
It's cute, and I can see it being handy at times. But, I think in general I'd find it very distracting; the only time it matters to see if a link goes via a redirect is when you're moving pages (or cleaning up after old botched moves).
The stub detector, on the other hand, tells me something very basic and always applicable: whether the link I'm following goes to a page that's so small it might be useless (if I'm looking for detailed info, maybe I'll skip it; if I'm looking to edit, I know it needs work).
They should be separately configurable; I would vote for #3 if it's kept.
Oh -- another thing; I notice the stub link marker doesn't do its magic in Recentchanges. I suppose it would add a teensy bit to the db lookup, but it might be nice.
-- brion vibber (brion @ pobox.com)
Brion VIBBER wrote:
They should be separately configurable; I would vote for #3 if it's kept.
Oh -- another thing; I notice the stub link marker doesn't do its magic in Recentchanges. I suppose it would add a teensy bit to the db lookup, but it might be nice.
That's why I didn't put it into the function for existing links (makeKnownLink) but rather in the "rendering" part. I suppose the database would be stressed too hard if many people would use it.
I patched Recent Changes at the test site so it displays stubs, without additional database queries. That should work.
Magnus
Magnus Manske wrote:
Brion VIBBER wrote:
They should be separately configurable; I would vote for #3 if it's kept.
Oh -- another thing; I notice the stub link marker doesn't do its magic in Recentchanges. I suppose it would add a teensy bit to the db lookup, but it might be nice.
That's why I didn't put it into the function for existing links (makeKnownLink) but rather in the "rendering" part. I suppose the database would be stressed too hard if many people would use it.
I patched Recent Changes at the test site so it displays stubs, without additional database queries. That should work.
Cool!
-- brion vibber (brion @ pobox.com)
Magnus Manske wrote:
On Lee's test site (http://www.piclab.com/wikitest/wiki.phtml), I have patched the stub detector so that /when the stub detector is active/, it will also show links to redirects with a yellowish background.
Should I use that on the live software
- not at all
- only when the stub detector is activated
- for all logged-in users as an user option, off by default
- for everybody
The most simple implementation is #2, and it doesn't require an additional database query. That's what I will do if I don't here good arguments against it ;-)
I like the stub detector, but I can't imagine ever wanting this. To avoid unnecessarily making Wikipedia look like a box of rainbow sprinkles, I would vote for 1 but settle for 3. But please no 2 (and 4 is just unspeakable).
-- Toby
wikipedia-l@lists.wikimedia.org