Maybe a simple solution would be to have a category and sourcing guidelines ("Category:Articles whose core content should primarily be from current peer reviewed sources"), for those topics where the topic and availability of references means that core topic material should be cited primarily from strict and limited sources.
Users can agree by consensus if an article should move to that category. Bots can also be used to identify and suggest articles where most cites are already from a list of peer reviewed or other highly reputed sources, for review.
This fits within normal handling. Then for such articles, a bot reviews these every while and highlights those cites which are more than 6(?) 12(?) months from publication so our academic experts can check these and update them or ensure they're still currently valid.
So... essentially isn't this something easily done already, without needing the weight of a taskforce and quality project behind it? Any small group of users could propose a mechanism of this kind and the low level of bot, feed, or other infrastructure needed to support it.
The twist is, we can then use the processes being discussed elsewhere to heavily "push" this along. We can ask readers to ("Help Wikipedia! This topic has 3 cites more than 6 months old. Please update them if they're out of date for the field!") or put these onto a feed or user filters, for users interested in the topics. Cites are tagged with the date of publication so a Citation Check Wizard can ask on editing, "I notice you're about to add a 18 month old cite to this topic. Are you sure this still represents latest knowledge in the field?". And so on. Think about it :)
It also shows how the seemingly simple recommendations we've been discussing can also be used to improve quality in other ways and places - including even specialisms like expert sourcing, which was never previously considered. That's the kind of pervasive effect that we can aim for. Notice we were considering wizards and notifications in a "basic quality" and "new user" context, but once added they can open whole new quality doorways across the wiki.
I just read your post in the other thread and this Citation Wizard, wow, I really like the idea! An independent Citation Check Wizard would be great too.
The problem I wanted to point out in my last post (08:20, 27 November 2009) wasn't the problem of outdated sources (Actually, 18 months seems little time for me, 10 to 20 years is already quite good for scientific papers). It's the problem of wrong citations due to a lack of understanding in the subject (see the list of examples I gave in my last post - they are unfortunately quite common). The citation wizard could help, for example by giving advice depending on expertise:
- You are:
- An expert for this subject, knowing well the consensus in the literature -> either use primary sources (scientific papers) or secondary sources (textbooks), depending on how general the information is.
- Just interested but knowing a lot about the subject -> use secondary sources (textbooks), only use primary sources (scientific papers) if you are sure what you're doing.
- Just searching for a source that can replace the stupid template! -> try to find a secondary source (textbook or neutral website) which has the same subject as this article or section.