Re: [xml-dev] XML's greatest cultural advantage over JSON

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

On 4/28/13 8:26 AM, David Lee wrote:
> For all of us the world we live in is the things we need to
> accomplish.    And in MY world JSON absolutely does not hold up to
> its promise whatsoever.  But I am perfectly happy to admit it works
> in other worlds where the model fits.

Except that you don't admit that until pressed on the point because 
you've just been blasting away.  Well, okay, you did admit that it's 
fine for serializing JavaScript objects, the smallest possible use case.

> What I am trying to accomplish with this rant is simply a warning
> that one should not blindly assume that a model that fits in one
> world is necessarily appropriate to another, even another world which
> is claimed to be appropriately similar. And knowing better worlds
> where it does fit doesnt help at all with where it doesnt.

Actually, it does help.  It helps me ask hard questions about what "fit" 
means, and what kinds of applications are worth building.

> Its not
> JSON itself at fault but IMHO the "religion" around it that pretends
> it's a perfect fit for everything.   Buyer Beware.

Buyer beware has been the rule for XML since its arrival as well.  It's 
only inside the XML "religion" that that is difficult to see.

I'm just amused this morning by the spectacle of various XML notables 
manning the defenses against the impertinent invaders.  This is the 
conversation I had hoped would break out when Doug Crockford spoke at 
XML 2007, but I guess we didn't take the idea that XML might face real 
competition seriously then.

Returning to the XML world is very strange for me.  I've been to see 
strange wondrous lands where the rules I was taught for XML don't apply. 
  On my return, the core of the XML world seems not to have changed at 
all, except that it is far less prosperous and the border guards are 
much crankier than in the past.

(The SGML frontier has quieted, while the JSON frontier seems perilous.)

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