Re: SPAM-LOW: Re: Visible MetaData == "Visible to whom?" was Re: Design Principles

Maciej Stachowiak wrote:
> Browsers do show an icon for feeds in the UI, but, the XML icon on web 
> pages (which really should be an "RSS" or "Atom" or "Feed" icon) is an 
> <a href> link in the content. Wouldn't it have been better to just use 
> <a rel="feed"> instead of <link rel="alternate"> to discover feeds in 
> the first place? 
It is better when the site owner wants to include a visible link, but 
not when the site owner does not want to incorporate said link into the 
site owner otherwise feels is a nice clean visual design.

As an answer to an unasked but potentially implied question, my 
preference is for an author to include both a <link> and when applicable 
an <a>. This empowers both crawlers and user agents, and for dynamically 
generated content can potentially be maintained in one place.
> Then no one would have to wait for special UI in the browser to see 
> the feed links, 
Anyone with Greasemonkey and a script can expose that metadata. And 
that's my goal for T.oolicio.us; to make it possible for to see new 
metadata w/o requiring a browser revision or even people to have browser 
plug-ins.
> and there would be no chance of an explicitly author-added link in the 
> visible page content getting out of sync a feed specified in the 
> <head> section.
I don't think this a valid objection; i.e. don't allow it because 
authors might not maintain consistency. If consistency is a major 
concern, we really should deprecate a lot of existing HTML features and 
add in support for more indirection, i.e. the example to define macros 
for a collection of elements and/or multiple elements with the exact 
same attributes used (although, the macros *would* be nice...)

I would instead prefer to launch a subgroup that would address how we 
can increase the percentage of valid HTML pages across the web. I have 
many ideas for how to improve validity, some of which I brought up on 
the WHATWG but they shot down my ideas as not being of interest and/or 
out-of-scope.
> So this is actually a perfect example of why visible metadata is 
> better (and indeed HTML5 supports feed discovery on <a> elements, 
> belatedly solving htis problem).
It may be better when applicable, but that better should not eliminate 
the prospect of useful where not applicable.

-- 
-Mike Schinkel
http://www.mikeschinkel.com/blogs/
http://www.welldesignedurls.org
http://atlanta-web.org - http://t.oolicio.us

Received on Thursday, 29 March 2007 20:16:03 UTC