Re: Publishing a new draft (HTML5+RDFa)

On Fri, 31 Jul 2009, Manu Sporny wrote:
> >
> > The reason microdata is in HTML5 and RDFa is not is that RDFa has 
> > fundamental technical problems
> 
> As has been stated numerous times, what you are claiming to be 
> "technical problems" are, in fact, "personal preferences based on design 
> philosophy".

By "technical problems" I mean problems with the design, as opposed to 
editorial problems. They're primarily usability issues, which are to some 
extent subjective. I make no apology for having an opinion on what makes a 
usable language; it's my job to have such an opinion.

Generally speaking, my position on this topic is a straightforward one: 
simpler is better.

RDFa fails at being "simple" in a number of ways which I detailed in my 
last e-mail. I consider these problems to be serious.


> There was a very long set of discussions that we've had that resulted in 
> every demonstrable technical issue that HTML5+RDFa has, being recorded 
> here:
> 
> http://rdfa.info/wiki/rdfa-in-html-issues
> 
> We asked that others, including you, modify the wiki page to document 
> real technical problems.

It's not my job to maintain your issues list.


> > For example, the use of prefixes
> 
> If I understand you correctly, the use of prefixes is a "personal 
> preferences based on design philosophy" disagreement that you have as 
> you have yet to produce a technical problem with the use of prefixes. 
> How does that feature break UAs?

I don't understand what you mean by "break UAs".

Prefixes are widely documented to be an antipattern in language design. 
For example, see this e-mail:

   http://lists.w3.org/Archives/Member/w3c-archive/2009Jul/0125.html

...where I give a quite detailed analysis of why prefixes are a feature of 
poor language design.


> > the exclusive use of URIs for identifiers
> 
> Again, if I understand you correctly, the use of "URIs-as-identifiers" 
> is a "personal preferences based on design philosophy" disagreement that 
> you have as you have been unable to produce a technical problem with the 
> use of "URIs-as-identifiers". How does that feature break UAs?

Exclusively using URIs for identifiers is bad language design because URIs 
are a poor form of identifier for many things. This is why few languages 
use them as such -- for example, programming language usually use simple 
tokens as identifiers. URIs are good for identifying resources with a 
scheme, host, and path, but not for verbs or predicates. Using URIs for 
features which they aren't suitable for doesn't break UAs, it hurts 
_authors_, which is a far worse problem.


> > the use of rev="", the use of rel="" in a manner incompatible 
> > with the rest of HTML
> 
> What is the "technical problem" here? How is this breaking UAs?

rev="" has been shown to be poorly understood by authors.

The problems with rel="" have been discussed to death by Julian and 
others, and I won't go into them here.


> > the overly complicated processing model
> 
> Again, does this break UAs?

It makes the language bad. That's a much bigger problem.


> We have more than 7 RDFa processor implementations now, so the issue 
> isn't developer-oriented.

Did you see the number of problems Google had with its RDFa documentation? 
That's a sign.


> If you mean that it's overly-complicated for authors, then it might be 
> -- but the same amorphous claim could be made for large swaths of HTML5, 
> SVG and Javascript.

It _has_ been made of SVG, yes. If you have specific part of HTML5 that 
are too complex for authors, let me know, I'll fix them. I can't speak for 
JavaScript, I'm not directly involved in that work.


> So, what exactly is the "technical problem" with the processing model?

It's incredibly complicated.


> > the presence of features that aren't necessary (such as per-value data 
> > typing)
> 
> Again, your personal preference. There are some that would like to be 
> able to assign types to data. How is this a "technical problem"?

What is the use case for per-value data types?


> > the ability to include multiple name-value pairs per element.
> 
> Personal preference - what "technical problem" does this create in UAs?

It makes the language incomprehensible for authors.


Complexity for authors is one of my most important concerns in HTML5's 
development. You can dismiss this concern as being a "personal preference" 
if you like. Meanwhile, microdata is solving the same problems with orders 
of magnitude less complexity.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Friday, 31 July 2009 08:24:10 UTC