On Tue, Nov 25, 2008 at 11:56 AM, Oskar Sigvardsson oskarsigvardsson@gmail.com wrote: [snip]
The moral argument says that we should make sure that people don't rely on only our information when it comes to serious decisions with serious consequences. I don't think it would be a bad thing at all if at the dosage section of an article on drugs we say "Consult your physician before taking medication" or on the article on nitroglycerin, have a small little disclaimer in the "Manufacturing" section saying "It is extremely dangerous to try this yourself if you are not a trained chemist".
The problem is that for almost any question of fact you can construct a scenario where having wrong information may be seriously harmful. For example, I've pulled formulas from articles on mathematics and electrical-engineering related subjects which could create significant fire hazards if they were wrong and I applied them uncritically, to things like a nuclear physicist creating another Chernobyl incident if he believes our figures on the neutron cross section of various materials, to someone losing a high dollar value bet based on some information we have on pokemon.
Even if you ignore the sillier possibilities we're still talking about a considerable chunk of the project which could present a risk of harm to people if they fail to apply appropriate scepticism and were the information incorrect or incomplete.
Once the warnings are more than something rare and highly targeted people will just suffer banner blindness and not even see them. Messages few people notice will not improve our moral position (and I agree in terms of pure-legal CYA we're already as covered as much as reasonable).
It's pretty easy to demonstrate that notices are almost totally ineffective. We could anon-notice "*Wikipedia is untrustworthy, do not take risks based on information included or potentially excluded from this page!*", and many people would simply ignore it.
Allow me to suggest these steps:
(1) Better public education about how Wikipedia works so the general public gives it the appropriate grain of salt. The "Ten things you didn't know about wikipedia" as a anon-notice was that kind of effort. We should do more things like that. (This could also include WMF organized press events, like a "Wikipedia Reliability day")
(2) For things like prescription drugs where there exists external notable and basically reliable materials on safety we should adopt a standard highly visable infobox field that link people to these resources.
(3) Adopt revision flagging so we have a tool to make efforts like (2) effective: Getting good information in is ineffective when ~2% of page views (millions of people) are to vandalized versions which may omit the good information.
None of these would require more disclaimers that people will just ignore, and I think they stand a better chance of improving people's safety when they use Wikipedia.