RE: [xml-dev] "Introducing MicroXML, Part 1: Explore the basic principle

["Len Bullard" <Len.Bullard@ses-i.com>]

>I'm guessing that the difference may be that we're talking about
>information, not physical resources?  Not sure.  What do you think?

It makes no difference.  Property is property.  There is a politics or
power domain and once there is property involved, it can't be ignored.

>I don't know much about XHTML.  I always thought of it as HTML.
>I believe what's required is a consensual merger, not a takeover.  And
>after that, the consensus process should be managed more closely than
it >has been.

That's a key point:  consensus.  As we've tried to explain, systems such
as you proposed have been built successfully.  Few of them still exist.
Why?  Again, see the first item: it isn't our property to modify and a
vote on it doesn't change that.

>I note that the IETF is having their 84th meeting this month.  Maybe
they
>have got a good handle on the consensus-management approach.  

Let's see how well they handle the so-called Censorship Code.  Or how
well the W3C handles Do Not Track.  Again see the first point.

> As long as the web itself IS an application (which in ISO 
> layer architecture, it is), whoever governs the web sets 
> those rules.  This is not TimBL, the W3C or any of the others 
> you mention but the mostly anonymous browser designers at a 
> handful of companies and their moneyed customers. 

>I don't know.  Isn't the Web community bigger than any of those things?

No, it isn't.  The number of people who drive cars are considerably
bigger than the number of car drivers.  Remember, what you are proposing
is not simply a specification change but a feature that has to be
implemented in every application using MicroXML.  For that precise
reason much of what was Hytime was thrown away no matter how
tantalizing.  Even the Part II of XML for Linking has more or less died
on the vine.  The scale of implementation is roughly inversely
proportional to the number of features and the more one adds, the more
resources are required to develop and field it, so the justification
must scale up in power exponentially.

>And thankfully we have somebody to pay browser designers and developers

>to move the yardsticks. The web is a kind of federation, where anyone
>can participate.  That's a thing of beauty!

That was true twenty years ago or so some thought.  See first point.
 
>Acutally, I'm also glad it can't occur.  That would be the worst 
>possible dictatorship.  

Benevolent dictatorship is one of the most successful organizational
types known because it is fast and irrevocable while in place.  However,
once past the initial acceptance and fielding, its very successes
preclude its continuance.  Why?  Consensus is a stag hunt and the rabbit
is too attractive to bypass so defections grow steadily.  That is why I
gave you an economic instead of a technical answer.

>I just wanted some leadership to be exercised on matters of import to
the >community. 

Then at this point the person you need to convince is Liam.   If he
accepts it, your chances improve considerably.  Do remember it is easier
to change no to yes than yes to no, so he is rightly hard to convince. 

>Well it's a good thing that the architecture of the web was so well
>established prior to Web 2.0, IMHO.  

Indeed.
 
>I think that adding simple, usable hypermedia to XML is not too big a
>technical change, and, to me, it doesn't seem too risky for anyone,
>since we can always not use those semantics, if we don't want to...

Two points:

1.  XML is NOT a hypermedia language.  By design.  It's a way of marking
up resources.

2.  We did try.  See XML Part II: Linking.

Why has part II died on the vine?  Because most of what it enables can
be done by other means that are already fielded and proven.

Why do we need HTML5?

>However, I think it could benefit everyone
>if we start thinking about what REST is and how we can burn it into our
>DNA so that we don't have to think too much about it again; it just
works.

What Andrew said.

>Here's what TimBL says about links:

Tim has a model.  What if I told you the universe we know is not about
connections but motion through a field that imparts the illusion of
mass?

IOW, the model isn't the issue here.  The implementation is.  It is of
no value to add hyperlinking to XML if it won't be implemented.   The
nut to crack is will it be implemented and if so, what difference will
that make to current web products and processes.   

The Do Not Track and the Censorship Code examples are illuminating.  

A good idea either works outright because all of the machinery is there,
ready to use and easy to understand (see JSON) or it doesn't because it
takes too much development and is hard to understand why it will be
useful (See Hytime).  As I said, systems like what you describe do exist
and have existed since before the web was "invented".

>But I'm just guessing, we need some more insight, from the community.

They've been providing it to you based on decades of experience.

len