Re: [xml-dev] Not using mixed content? Then don't use XML

["Simon St.Laurent" <simonstl@simonstl.com>]

On 4/10/13 11:56 AM, James Fuller wrote:
> On Wed, Apr 10, 2013 at 4:13 PM, Simon St.Laurent <simonstl@simonstl.com> wrote:
>> Yes, absolutely.  Have you spent much time with folks using JSON?
>
> I am part of that crowd as much as I am XML folks ...I subscribe to
> the  'markup is for documents and json is for data' sentiment.

And so all the JSON folks you hang with are eagerly seeking schemas to 
supplement the one useful part they lost in moving away from XML?

We may be in different conversations, but I'm certainly not hearing that 
in mine.

> do you you think json is replacing markup ?

For many use cases, yes.  Even in the document-centric HTML world, JSON 
conversations are becoming central and the HTML document itself reduced 
to a template - or just constructed from JavaScript.

>> Data has to be constrained in code somewhere.  It's more about where -
>> locally, not globally - and when - on a conversation by conversation basis
>> rather than in advance.
>
> losing me again ... perhaps what you are advocating for is less rigid
> spec/constraint of author intent and more runtime introspection of
> data (that somehow communicates author intent) ?

I'm saying that we accept what Sean has called "probabilistic" 
interpretation, that we maybe understand what the other person's 
document is saying until told otherwise.  I'm also saying that we stop 
trying to lock down the ingredients allowed in any particular 
conversation/document.

Yes, that requires a change in our processing models and in our 
expectations.

> as many of my colleagues remind me ... what we do is more about
> engineering ... and all engineering is about is making tradeoffs ...
> for me personally identifying all the tradeoffs is quite a task (from
> which to balance things). Most of what you discuss about schemas seems
> to be a question about either not seeing the tradeoffs in their
> entirety or not balancing them ... apologies for the application of
> reductionism.

Actually, in many ways this is an attack on the engineering culture that 
brought us here.  It's not just the lawyers who dream of everyone 
marching lockstep along clearly defined pathways.

I don't believe that the current calculations about schema tradeoffs are 
the right set of calculations.

> I believe in our industry that hardware has the ability to change
> culture ... but not in 'straight arrow ways' e.g. consider the the
> efficiency of using bytes of RAM ... this approach was good enough to
> land a man on the moon ~40 years ago.  Now we have Gigs of RAM in a
> phone .... stunning to think but also stunning to see how sloth
> (versus efficiency) can also change mankind.

Moore's law certainly makes it easier for me to propose alternatives to 
schema models that might have made sense in the age of punchcards and 
seriously constrained resources, but I think everything I have in mind 
has been possible at least since the computers we had in 1998. In fact, 
I suspect that even the ancient Droid in my pocket could handle it.

Thanks,
-- 
Simon St.Laurent
http://simonstl.com/