ian at hixie.ch
Thu Aug 21 22:56:08 PDT 2008
On Fri, 22 Aug 2008, Silvia Pfeiffer wrote:
> On Fri, Aug 22, 2008 at 11:48 AM, Ian Hickson <ian at hixie.ch> wrote:
> > I honestly don't see significant interest in computer-readable
> > metadata. Just look at the average user's media library; most users
> > have terrible metadata hygene.
> It is true, we live in the middle ages of metadata hygene. But maybe we
> can help move out of it and into more modern times by making tools
> available that support metadata consistency better and that
> automatically take care of some of the hygiene. Any such efforts should
> be supported. We will not be able to get away completely from having to
> deal with dirty metadata, but if it has an advantage to people, they
> will care.
When the choice is between:
1. structured metadata with defined behaviour that has great benefits if
the metadata is kept maintained on the one hand, and
2. freeform data input which computers can make sense of using
heuristics even if the data is a mess on the other hand,
...then it appears, based on everything we've seen since the Web started,
and even before that, that users will pick the second one every time,
_even when they try to pick the first_.
In fact, this goes much further than just metadata. Given the choice
between two systems with characteristics as follows:
1. Completely predictable and reliable behaviour.
Easy to process data and make inferences.
Requires users to be disciplined and honest.
2. Unreliable behaviour.
Processing data is massively complicated.
Tolerant of users making huge mistakes and being dishonest.
...humanity seems to always pick option 2. Just look at the Web.
So I'm not convinced it's a given that we should automatically support any
system that attempts to use structured metadata and relies on users not
making mistakes or lying.
That's not to say that such systems are all doomed to fail. Maybe RDF and
RDFa will succeed where other systems haven't. But we (the HTML5
community) have to have at least some reason to think so before we support
them in HTML5, otherwise we'd be using different standards for this than
for the rest of the language.
> The whole SEO industry is built on that fact.
The SEO industry is built on _abusing_ legitimate metadata systems in
order to trick software into reacting in particular ways.
> > But in any case HTML5 already has extension mechanisms, so the
> > discussion should not be over whether RDFa is worth it or not, the
> > discussion should be over what extension mechanisms RDF needs that
> > HTML5 doesn't provide.
> I would like to understand exactly what changes to the existing HTML5
> spec would be required to support RDFa. Ben - can you clarify? Maybe the
> extension mechanism that Ian refers to already covers all the needs, but
> it has not been clarified.
Agreed. It would be useful to know what the requirements are here.
Ian Hickson U+1047E )\._.,--....,'``. fL
http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,.
Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
More information about the whatwg