Shared publicly  - 
 
To be blunt, W3C, your allowance of Microdata is the reason why I will no longer have anything to do with RDFa. 

Your cowardice in not pushing back on this pure WhatWG invention just fragments an already fragile metadata annotation environment. 

Now, which should we use? RDFa? Or Microdata?

Or will a third or fourth option come along when some bully company demands it?

So, suck metadata and fade, for all I care. 
3
Manu Sporny's profile photoSteven Pemberton's profile photoDan Brickley's profile photoShelley Powers's profile photo
22 comments
 
Send in an objection. I agree with you entirely. Microdata came out of thin air, with no warning, no consultation, no coordination with interested parties, using (at the time) RDFa attributes. It was entirely invented to muddy the waters, and block work that had been going on for years, with consultation with the interested parties.
 
Steven, I objected a long, long time ago.

Nope, I figured the W3C basically told the world metadata annotation in HTML documents isn't important. OK by me.
 
Would RDFa 1.1 Lite exist in its current publisher-friendly simple form without the whole Microdata misadventure? I'm not so sure. I don't like how the whole thing happened, but RDFa is a lot simpler to adopt today than it was a year ago.

IMHO there may now be a case for mothballing Microdata, now that RDFa Lite has taken on a more convincing simple-enough-for-mainstream-webmasters role (but with better extensibility). 

However there are a few awkward leftovers. I keep hearing positive things about Microdata's 'itemref' mechanism; it might be worth another look. And then there's the question of APIs. Rather than simply asking for Microdata to be discontinued, it would be useful to investigate if there are real world customers for the Microdata API; and if so, what exactly they need from such a mechanism. 
 
I have to disagree with you, Dan.

There was never anything stopping the Microdata people from working with the RDFa folks into incorporating necessary changes. 

Now we have two 'standards' where there's really only effective room for one. And you know something? The W3C will just close its eyes and ears and pretend that it's OK to have both. 

Having two different metadata standards is no different than having two different web document standards, and two different stylesheet standards. The whole point of agreeing on one is to simplify the implementation. 

All the W3C is doing now is putting off a decision because it's basically afraid to challenge Google. 

So, the signal I'm getting is metadata doesn't matter. 
 
PS The RDF may have had some barriers to participation, but the RDFa folks have been everything that is open. 
 
One other thing:

The WhatWG is nothing more than Hixieland. Time to stop placating this monstrously egotistical individual. 
 
+Dan Brickley, it is true that RDFa Lite 1.1 was a direct response to Microdata and it probably would not have happened had it not been for Microdata's existence. However, I'm not saying "Microdata should never have existed." I'm saying, "Now that RDFa Lite 1.1 exists, do we really need Microdata to go to REC? Why not just a W3C Note?".

There are a few awkward leftovers - itemref is one of them, which we're discussing here:

http://www.w3.org/2010/02/rdfa/track/issues/144

We're also working on the API, as the data generated from the Microdata API isn't great from an extensibility and datatyping perspective. We think the solution is going to be RDFa -> JSON-LD, as it provides the best of both worlds. JSON-LD allows developers to work with JSON objects natively while being able to map those objects to a more robust data model (RDF) with relative ease (by referencing a JSON-LD Context). I'm available to chat more about this if you're interested in hearing more.
 
I'd be very interested to hear whether JSON-LD meets the needs that current implementors of the Microdata API feel they're addressing. 

To what extend could the Microdata API as-is, be used as an API for in-page RDFa Lite? What would break?
 
Although I object to microdata because its lack of process and consensus, it is not the first time W3C has done two of the same. CSS and XSLT-FO for instance. 
 
And there's XHTML and HTML. 

But these are also different things. Actually, quite a bit different. 

I've not followed especially closely to the RDFa Lite stuff, but from my understanding, it's basically identical to Microdata. 

More importantly, having both creates confusion. Should we use RDFa? Should we use Microdata? Should we use both? 

I'll be blunt: unlike style sheets or document formatting, metadata annotation is just not on the average web developers/designers high list of priorities. In case of any confusion, they'll as likely bag the whole thing. 
 
There was also XML Schema and RDF Schema, http://www.w3.org/TR/schema-arch (and a few related pairings, eg. XQuery 'vs' SPARQL). Or RDFS and OWL. Or OWL Full and OWL DL. We're fractious at every level of the org chart! 
 
Shelley, re "I've not followed especially closely to the RDFa Lite stuff, but from my understanding, it's basically identical to Microdata.".

That's not quite the case. From a publisher's perspective, the element/attribute structures look very similar; a trivial  search/replace regex can handle most conversions of microdata to rdfa lite. From a parser writer's perspective, they are pretty different: there is a lot more to do, to be an RDFa 1.1 parser (and there is no such thing as an RDFa Lite parser; you have to do the whole spec). Since there are many more publishers than parser writers, this is not a terrible state of affairs.
 
The point is, it makes no sense to have two competing specifications from the same group. 

Why should I waste my time annotating my web pages with metadata, when the W3C can't even make up its mind on the metadata spec I should use?

It's dumb.

Manu says it much more politely. 

http://manu.sporny.org/2012/microdata-cr/ 
 
+Dan Brickley re: "I'd be very interested to hear whether JSON-LD meets the needs that current implementors of the Microdata API feel they're addressing."

Who are the the people using the Microdata API? I don't know of any large use cases for it. I used to think it was a good idea to have a browser API for reading metadata from the current page (see our work on the RDFa API), but I don't think that anymore.

It's far too early to standardize on this. Our work on JSON-LD has shown that there are /many/ different ways to work with this data and it's not entirely clear how developers are going to want to work with it. The Microdata API isn't terrible, but it's lossy. You can't tell IRIs from text strings, no datatyping, etc. Standardizing it now is premature... especially since we can iterate a few times using Javascript instead of putting something into the browsers that makes a number of assumptions that we shouldn't be making.
 
Michael (tm) Smith is probably better connected to those implementors than me. I'd love to know more...
 
But Shelley, the "objection" I'm referring to is a W3C formal part of process. I don't mean send an objection to the people doing it, I mean submit a capital O Objection, so that the issue is forced at the next transition point.
 
Steven, I don't think it will make any difference. Regardless, I believe Manu will make a formal objection, and he's respected within the W3C. 
 
So are you, and two objections are better than one.
 
And a friendly, practical tutorial for Microdata users on how to migrate to RDFa Lite would be worth 100 objections
Add a comment...