Topic on Talk:Article feedback

I started a redirect at w:tissue printer and was promptly presented with a feedback box on the redirect page I'd just created. Now I understand that exempting me from editing a page only I'd edited would involve a level of user-tailoring that I probably don't want. But why rate a redirect?

Likewise, I think rating a stub article is silly. There's only so much a few sentences can tell.

Therefore I suggest that articles shorter than some character limit should simply be exempt from this rating system. I'm thinking that 64 characters should be non-controversial, while 1024 might be too much. Wnt 23:58, 29 May 2011 (UTC)

I believe this is done and waiting to be deployed.

This was requested under bug 29164 and has been fixed in the development version. It will be live during the next deployment.

I read http://www.mediawiki.org/wiki/Special:Code/MediaWiki/88146 and http://www.mediawiki.org/wiki/Special:Code/MediaWiki/88151 - these do sound like they should abolish the rating of redirects. I still wonder if other short articles like stubs should really be rated. Wnt 19:34, 30 May 2011 (UTC)

While stubs don't contain a lot of information, they may still contain perceived misinformation or bias. Indeed, looking at stub ratings, you'll generally find them to get reasonable ratings except in the "Complete" category. Now, when a stub has low ratings across the board, that's when we might be looking at a highly dubious new page that slipped through new page patrol.

I agree on this. I think the feature can be useful even for small articles.

Support for sure.

Yes, also rating disambiguation pages is ... awkward.

They can be rated, but the four categories seem to be suited specifically to rate medium-length articles only. As the OP stated, they make no sense for really short articles and stubs. Yes you can rate them "Incomplete", but then does it make sense to rate them as not Trustworthy or not Objective as well? Does "Well-written" mean quality of writing even if it's short, or is an article that is too short per definition not well-written?

Similarly for very long articles that present a very wide range of opinions and perhaps quality of writing, does that make them Trustworthy? Well-written? I wouldn't like to summarize that into a 1-5 star rating.

IMO, the four categories are not very well chosen, as they mean very different things in differing contexts of articles.

Sure, these are definitely the four (most important) things you are interested in measuring. But that doesn't mean you must measure them by straight-out asking for them! :) You got a huge encyclopaedia here at your fingertips, how about starting to read up a bit on Social Research?

Basically, you forget there's an intermediate step. The four categories are called "axes" on which you'd like to score most proper articles. In that context, they are probably well-chosen. But in order to obtain the scores on these axes, you must formulate the questions wisely, which is most often NOT done by asking "How much would you rate this on the axis of X?" (which is what you're doing now). Entire textbooks have been written about this subject, which at least a century old (and depending on how you look at it, even much older).

Does "Well-written" mean quality of writing even if it's short, or is an article that is too short per definition not well-written?

I would choose the first option. But for consistency (an important aspect which was mentioned on other comment), this should be made clear to the reader, so that different people use the same criteria when rating a page.

I suggest that articles shorter than some character limit should simply be exempt from this rating system. I'm thinking that 64 characters should be non-controversial, while 1024 might be too much.

It should be noted, though, that this would need to be configurable on each wiki, and it should probably use some other measure instead of the size of the wikitext in bytes, since e.g. on Wikisources a page can have just a few characters and still be a long page (since most pages are transcluded from "Page:" namespace by the Extension:Proofread Page). Maybe it would be better to consider the size of the expanded wikitext of the page.