W3C home > Mailing lists > Public > www-html@w3.org > January 2011

Re: Link Relationship for Moving Down a Hierarchy

From: Ian Hickson <ian@hixie.ch>
Date: Tue, 18 Jan 2011 08:41:04 +0000 (UTC)
To: "Sean B. Palmer" <sean@miscoranda.com>, David Woolley <forums@david-woolley.me.uk>
cc: www-html <www-html@w3.org>
Message-ID: <Pine.LNX.4.64.1101171946411.28618@ps20323.dreamhostps.com>
On Sun, 16 Jan 2011, Sean B. Palmer wrote:
> On Sun, Jan 16, 2011 at 12:11 AM, Ian Hickson wrote:
> > 
> > what software are you hoping will act on this, and what are you hoping 
> > it will do with it?
> 
> Scripts which parse the @href and @src attributes in a site, compile a 
> graph from them, and use that graph to inform various things. Some 
> examples:
> 
> * Automatically generate tables of contents

For this, rel=chapter seems like what you want, if it's actually needed. 
Most tools of this nature seem to just grab all the documents in a 
directory, or something along those lines. (After all, if you're 
_generating_ a TOC, there's no guarantee that there's already a link to 
all the chapters.)


> * Give lists of file dependencies to make moving pages easier

That's already possible without anything new. For static documents, you 
can walk the document, grabbing resources mentioned (e.g. grabbing images 
mentioned from <img src>, grabbing stylesheets in <link rel=stylesheet> 
links, etc). The spec gives sufficient guidance for this already, I think, 
but we can add more if that would help. For dynamic apps, the appcache 
manifest feature seems ideal.


> * Show pages needing integration, what wikipedia calls "orphans"

Surely for this you just have to scan the documents to find out what 
documents aren't referenced. I don't see how a rel="" anything would help.


> This code is already written, and is using rev="up" as part of the first 
> example in the list above. Since @rev has been removed, I now need an 
> alternative. I do not mind which alternative, as long as it has 
> consensus and can be parsed.

Well, you can register a rel="down" value if it would help, but I'm not 
sure it's the best solution for the use cases you list. If you have code 
that's going to use it, though, it doesn't much matter what I think. :-)


On Sun, 16 Jan 2011, David Woolley wrote:
> Ian Hickson wrote:
> > 
> > Why? That is, what software are you hoping will act on this, and what 
> > are you hoping it will do with it?
> 
> Although my feeling was that "down" was too generic, I think that saying 
> the relation should not be there reflects the loss of the declarative 
> nature of HTML.

I wasn't at all trying to imply that. I'm not sure why you interpreted it
that way.


> HTML markup shouldn't just be there to instruct the browser, but also to 
> provide information on the author's intent.

What do browsers have to do with this?

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
Received on Tuesday, 18 January 2011 08:42:10 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 27 March 2012 18:16:18 GMT