[microformats-discuss] Microformats and the Semantic Web

[Tantek Ç elik]

On 8/11/05 2:16 AM, "Danny Ayers" <danny.ayers at gmail.com> wrote:

> On 8/10/05, Ryan King <ryan at technorati.com> wrote:
> 
>>> My observation is that accessibility engenders adoption.
> 
> Assuming you mean ease of use, yep, I'm sure this is a big factor. But
> definitely not the only one.

It's perhaps the most important one.


> Things like fitness for purpose,

You never know what purposes people will use new technologies for, so this
is less important a factor than you might think.

What's more important is to model current behavior, whether you can
understand a "purpose" in that behavior or not.  The behavior exists, that's
all that matters for adoption.


> utility,

Some utility is needed, but too much utility is often an encumbrance.


> tool support, vendor support,

Those two often come directly from ease of use/adoption.


> timing,

Indeed.


> documentation,

That both helps and hurts.

There are plenty of W3C formats with *too much* documentation (e.g. lengthy
cryptic specs).  Too much documentation is a negative.


> evangelism
> (including hype and FUD about alternatives),

That's of limited utility.  It only works if people get some actual results
quickly.  Otherwise it engenders a backlash.  This is happening with RDF.

One of the reasons that "microformats" and "lowercase semantic web" are
resonating with so many people so quickly is ironically thanks to how many
years and years RDF and the Uppercase "Semantic Web" have been hyped with
very little of *additional* utility to show for it (additional = very little
marginal utility above and beyond existing or simpler solutions), at least
on the public Web.


> and maybe just plain luck
> play big roles. Barely a day goes by without my looking at a PDF doc,

I hardly ever look at PDF docs.


> Email message format is
> relatively simple, but the docs (RFC 2822) aren't exactly friendly.

Right.  Note the balance.  The format is easy to *publish*.


> Blog trackback format (RDF/XML hidden in XHTML comments) has to be one
> of the clunkiest designs ever (no fault to Ben Trott, there wasn't an
> obvious alternative at the time) yet seemingly within seconds of
> Movable Type supporting it, every blogging tool under the sun
> supported it.

Not true.  Few tools support TrackBack today, and the clunkiness of user
interface means it's not used except by a very select few.  Mainstream media
that have tried implementing TrackBack have universally found it to be a
failure (e.g. CNET).


> The same can be said about the blogging and pingback
> APIs - truly ugly formats, with much simple alternatives (doc-oriented
> XML over HTTP) yet with huge adoption.

WTF is "doc-oriented XML"?   Do you mean "generic XML"?  See recent blog
posts about that.


> RSS 2.0 is usually quoted as an
> example of a successful 'simple' format, yet Microsoft's Channel
> Definition Format was a simple predecessor complete with widespread
> impementation (in Win95) but it didn't take off.

Timing might be the difference there.

And you make a good point -- look how long it took RSS to take off.


> Whereas RSS 1.0, with
> its more complex syntax, is supported virtually everywhere RSS 2.0 is.

Also not true.  RSS 1.0 is largely a dead format.  We see a minor fraction
of RSS 1.0 feeds vs. the predominant RSS 2.0 format.


>>> The nice thing about microformats is that, while they sacrifice a
>>> lot of representational power, they maintain the valuable core
>>> proposition of representing data in familiar markup constructs.
> 
> Agreed, that is a huge plus.
>

Yes.

>>> People familiar with the standard html toolset look at that and say
>>> to themselves, "I can at least attempt it".  I'm not sure as many
>>> people say that when confronted with RDF.
> 
> I'd put that down to presentation - the fact that microformats can be
> viewed as RDF formats is significant.

To be frank Danny, it's significant to a certain community.  I'm not sure it
is significant to the overall adoption rate/potential of microformats (thus
I see it as in general off-topic for this list, though I wanted to let this
particular thread go on a little longer that last time to see if it revealed
any interesting points).  I mean, it's great that the compatibility is
there, but I can also see many more implementations developing which "just"
use microformats, take the data, and apply it directly to a solution,
without going through separate/more-abstract interim model.

Tantek