On Sat, Aug 13, 2011 at 6:11 AM, TJ Frazier <[email protected]> wrote: > On 8/12/2011 08:47, Nóirín Plunkett wrote: >> >> 2011/8/12 Jürgen Schmidt<[email protected]>: >>> >>> On Fri, Aug 12, 2011 at 12:19 AM, Simon Phipps<[email protected]> wrote: >>> >>>> Maybe. But I see no reason why this list needs the protection of being >>>> on a >>>> controlled access page and would suggest doing so is what needs >>>> justifying. >>>> I have not seen a reasoned counter to my proposal for it to be on the >>>> community wiki, so will probably create such a page soon (unless someone >>>> else wants to). >>>> >>> >>> I think Rob has already pointed out why the list of tags besides the >>> mailing >>> list is a good idea and i support it. >>> >> >> Just so that silence isn't assumed to be assent, I think both Rob and >> Simon have good points, but I support the idea of putting the list of >> tags in a location that's more accessible to new contributions (ie, >> the wiki), even if that means it has to be one extra click for folk >> who visit the website. >> >> Noirin >> >> > The recent changes to the ML page provide an accidental illustration of why > most things belong on the wiki: web pages tend to get shabby and > out-of-date. > > The ML page[1] changes were very nicely done. Following the "these > guidelines" link like a good little newbie led me to the ASF page[2] on > email tips. In the "Other email guidelines" section, I also followed the > links; one[3] gave me a 404 error. Link rot. > > My hypothesis — I call it the Curse of the Web Page — is that such decay is > inevitable in a high-barrier-to-change environment, without a dedicated > corps of maintainers (a dull job). Old hands (who could fix things) are > unlikely to access the page at all (they already know this stuff), or > they're looking for something specific and wouldn't notice problems > elsewhere. Newbies, mousing around, /will/ notice, but can't fix it. The > Curse is not limited to links: typos, spelling, and other infelicities are > all evident on Apache pages; they need a little TLC. >
The way to deal with dead links is to run a link checker on the site occasionally. That does far better than a casual user will do to find these kinds of issues. Then you can fix them all in one batch. I assume OO.org already did something like this? Even major sites that depend on crowd sourcing editing, like Wikipedia, rely on bots to detect dead links. Think of it this way: You've heard of Linus's Law : "given enough eyesballs, all bugs shallow", right? But that doesn't mean that you ignore compiler warnings. Automated checking has an important role as well. > On a wiki, the technically adept newbie might fix the problem right there; > others would just leave a note on the Talk page. > How would you fix it in this case? Our page links to a valid page outside our project, right? And then that external page has a broken link. Even if our project used a wiki for our mailing list page, there is no way we could fix this problem from within our wiki. > [1] http://incubator.apache.org/openofficeorg/mailing-lists.html > [2] http://www.apache.org/dev/contrib-email-tips.html > [3] > http://cocoon.apache.org/community/contrib.html#Contribution+Notes+and+Tips > (busted link, 404) > -- > /tj/ > >
