BlogMatrix
 

Microsoft's webslices

edit David P. Janes 2008-03-05 19:50 UTC 7 comments  ·  ·

Microsoft is adding something to IE 8 called “webslices”. A webslice is a part of a webpage that updates and can be subscribed to – for example, a weather forecast.

A webslice is different that a syndication feed because:

  1. it’s part of a webpage, rather than an independent feed (though it can have an independent feed also)
  2. it’s a single thing that updates, as opposed to a pipe that new things (such as blog posts) are coming down one after the other

You can read the webslices spec on this page and brief commentary here.

Now, what makes this interesting to me is that it’s based on the hAtom microformat (which I authored and championed). It’s based but unfortunately it’s not the hAtom microformat, as they add their own top level concept, the “hslice”.

Now, I can understand to some degree why Microsoft did this:

  • a webslice is the concept of a feed with a single item that’s continually being updated
  • hAtom requires an “author” element
  • hAtom requires an “updated” or “created” element
  • people think that hAtom is about weblogs – but it’s not true.

These issues are addressable. hAtom 0.2 is going to drop the requirement for the “author” element – it will be assumed to be some value (what, we’re not sure yet) if it’s not physically present. hAtom isn’t about weblogs – it’s about semantically marking up microcontent – and so is entirely applicable to webslices.

The “updated/created” issue is a little tougher. Allow me to suggest that if not present, these default to now, and leave it up to parsers to figure out if things have changed. This will together with the time-to-live (ttl) element MS is proposing in the spec.

I humbly suggest then that MS change the webslice spec so that instead of looking like this:

<div class="hslice" id="1">
 <p class="entry-title">Item - $66.00</p>
 <div class="entry-content">high bidder: buyer1
 …
</div>
</div>

they make it look like this

<div class="hfeed hentry" id="1">
 <p class="entry-title">Item - $66.00</p>
 <div class="entry-content">high bidder: buyer1
 …
</div>
</div>

That is, change “hslice” to “hfeed hentry”. The beauty of this is that there’ll be no reason now why all the tools MS develops for working with webslices will also work with any page that contains hAtom.

If they’re interest in only seeing items that narrowly conform to the webslices spec (i.e. single item feeds), they can just look for single item feeds with no created/updated/author fields. But I think they should go all out and recognize all hAtom items and incidentally help put together the semantic web.

Comment #1Taylor

2008-03-05 21:38:56

I've been using hAtom to create trip wishlists

http://2007.xmlconference.org/public/schedule/detail/405

Actually I appreciated the requirement for an author, just wished it required a URL and not a name, ie, linking to their HP or openID.  I can however see the drive to create their own top level element.  During my talk, people kept getting confused about why I used hAtom, they were thinking synidcation of blog posts...while I'm thinking it's a great way to say 'I have a collection of links, each one editorialized and authored by the same person'...plust the date/and existing parser support made it ideal.

 To classify it I'm using the tag soup, an hAtom with tag travel:trip asserts that the content is a trip planned or taken, but I keep toying with the idea of just branching and creating hTrip, if nothing else but to avoid the mental block people get with it being hAtom.  The rest would stay exactly the same.

Comment #2David P. Janes

2008-03-05 22:02:01

Why can't you use a URL for an author? Ah, I see: because hCard requires a FN. I suspect we'll have to work that into the spec then.

<span class="author vcard"><a href="" class="url">..</a></span>

Making a new top level item would be an incredibly bad idea: it means all tools, such as FF extensions in 3.0, will not recognize your content. It's a bad idea too because they're misinformed about Atom is (The Atom Publishing Protocol (AtomPub) is an application-level protocol for publishing and editing Web resources) and hence they're misinformed about what hAtom does.

Comment #3David P. Janes

2008-03-05 22:11:15

Of course, another possibility is to introduce parameterization of "hentry/hfeed" -- i.e.

<div class="hentry trip">...</div>

or

<div class="hentry webslice">...</div>

Comment #4Stephen Paul Weber

2008-03-11 15:49:50

<div class="vcard author"> <a class="fn url" href="URL">URL</a> </div> Anyone? fn doesn't imply it's a human name, just that it's a valid label (formatted name) to call the person/org represented by the hCard

Comment #5Stephen Paul Weber

2008-03-11 15:51:42

Oh, also, re:updated - I'm not sure I get the obsession with being ATOM-compliant (but maybe that's my anti-ATOM bias) -- to me, some markup is always better.  I will happily encode hentries with only title and bookmark, I do it already -- I know it breaks the spec, but I'd rather have semantic invalid than tag soup.

Comment #6story

2009-04-05 20:13:17

Very useful files search engine. rapidshare-provider.com is a search engine designed to search files in various file sharing and uploading sites.

Comment #7Matt

2009-06-02 02:06:05

I'm not sure I get the obsession with being ATOM-compliant to me, some markup is always better.  I will happily encode hentries with only title and bookmark, I do it already.

club penguin

Add Comment