- From: Ian Hickson <ian@hixie.ch>
- Date: Wed, 31 Oct 2007 00:10:06 +0000 (UTC)
On Mon, 6 Dec 2004, George Lund wrote: > Ian Hickson <ian at hixie.ch> writes: > > > > Someone sent me a mail suggesting: > > > > | <breadcrumbs> > > | <a href="/">Main</a> >> <a href="/products/">Products</a> >> Dishwashers > > | </breadcrumbs> > > > > I think a better way of doing this would be: > > > > <navigation> > > <p> > > <a href="/" rel="top up up up">Main</a> > > > <a href="/products/" rel="up up">Products</a> > > > <a href="/products/dishwashers" rel="up">Dishwashers</a> > > > Second hand > > </p> > > </navigation> > > > > ...where we define rel="up" to mean "go up one level" (as now) and add the > > semantic that if the keyword is repeated, then it means up that many > > levels. I've noted this as something the spec will have to talk about. I've now added this to the spec. Let me know if you see any problems with it. > URLs already have these semantics built-in. <a href="../"> means > something special to web browsers without having to invent a new way of > doing that. What would these keywords do extra that can't already be > done if authors organise their URL-spaces sensibly? Well, some people can't organise their URL-spaces sensibly. Also, the URL space can only represent one hierarchy. > The idea of a <navigation> element would be very useful (like giving > speaking browsers the chance to skip their contents, for example). We now have <nav>. > But I don't see why that <p> mark-up should be added, as what we have > there isn't a normal paragraph in most human languages. It's more like > some kind of specially-ordered list, if anything. Yes, <ol> would also work here. On Mon, 6 Dec 2004, Olav Junker Kj?r wrote: > > By sensible URL-spaces, I assume you mean using slashes in URLs to > indicate the hierarachy? This is not really practical since it won't > work with many CMS'es, and it prevent you from moving rearranging the > hierarchy since this will break links. URL's should be allowed to be > opaque. Indeed. On Tue, 7 Dec 2004, George Lund wrote: > > Well, HTTP has well-defined ways of indicating that a resource has > moved. HTML shouldn't be trying to take over from these established (and > well-designed) divisions of labour within the system. > > What's proposed is a system where people (or processes) are forced to > edit the _contents_ of an object just because it has moved, whereas > conceptually moving something around in a hierarchy doesn't necessarily > require any modification of the object's contents at all. I think I'd > be correct to point out that such a situation wouldn't be very > 'RESTful'. Sure. But sometimes a resource is in two (or more) hierarchies at once, and then you can't use the URL space anymore. -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Received on Tuesday, 30 October 2007 17:10:06 UTC