At 13:20 +0100 UTC, on 2009-02-24, Helmut Tischer wrote: [...] > Sander wrote: > > To be nice to wary users. > > That's my point. So let's leave the technology now: > > Content Authors care about attracting readers, even if the readers are > not technically educated. That's what made the "http" to something > what everyone wants to use - and they even keep on calling it "www" ! Well, with "wary" I didn't necessarily mean "uneducated". (I'm quite wary myself, but I don't consider myself uneducated when it comes to the Web.) But I do get your point (up to a point) and agree of course with the idea of trying to make things no more complicated than necessary for end users. I'm just not convinced about the solution. In part because I don't think that users are truly helped by making them 'feel' that ".html" has any meaning on the Web. For some more arguments, see <http://webrepair.org/strategy/certification/requirements#req52> and the W3C Note it refers to: <http://www.w3.org/TR/chips/>. Also see TBL's unoffocial "Cool URIs don't change", at <http://www.w3.org/Provider/Style/URI>. If this is too much to read, consider the case where at some point an author changes a document from one format to another. What today is a HTML document, might tomorrow be a PDF, for instance. The content hasn't changed. Just the file format. If you were relying on a URL with a trailing ".html", it is now dead. If your URL were more generic, it would happily keep on working. For hyperlinks, in a Web context, a better solution would be something along these lines: <http://www.euronet.nl/~tekelenb/WWW/userfriendlierhyperlinks/> (which Dokuwiki basically already does). You are talking about a slightly different context though -- the situation where URLs are not hyperlinks but just text, like links in (plain text) email. So here's a thought: how about using content-negotiation for this? If, when requesting <http://www.dokuwiki.org/syntax.html>, the server would know to return the resource at <http://www.dokuwiki.org/syntax>, then one could add a trailing ".html" to links in plain text contexts, thus providing a hint that the URL points to a HTML document. This way nothing would (need to) be changed about Dokuwiki's URLs. It would merely allow someone to type an extra ".html" and, assuming the resource actually is a Web page, still get the intended document. If at some point the author changes the resource to some other format, Dokuwiki should probably return a 404. In that sense this does go against <http://www.w3.org/TR/chips/> and <http://www.w3.org/Provider/Style/URI>. But only in the sense of what sort of requests it accepts -- not in the sense of the URLs it actually generates itself. I'm inclined to consider this acceptable. It might seem to conflict with the general wiki concept of "requesting a non-existant page == creating a new page", but I believe Dokuwiki already provides the option to return a 404 in such situations, and in fact already does "pretty URLs", so I suppose this wouldn't be that much of a stretch. [...] > One further note: > For media files even DokuWiki uses the file extension to produce the > correct mime header Sure, but that's a *local* context. The reason that such headers are generated in the first place is because the file name extension has no meaning on the Internet. >, and DokuWiki behaves surprising (not for Dokuwiki > experts of course), if one mixes up > example.com/namespace/namespaceorfile with example.com/namespaceorfile/ I'm not sure which surprise you are referring to here. -- Sander Tekelenburg, <http://www.euronet.nl/~tekelenb/> -- DokuWiki mailing list - more info at http://wiki.splitbrain.org/wiki:mailinglist