James Clark on XML, JSON and lots of other stuff

Don Box's Spoutlet

Syndication

The brightest guy in XML-land (no offense Tim :-)) has just started blogging.

I love his analysis on both the state and role of schema and his followup on JSON.  A bunch of us are working hard to move Microsoft towards even more structural typing - wish us luck!

Unlike James, I'm less obsessed with making people use something other than their programming language to define schemas. That said, I wholeheartedly agree the world would be better off exchanging something closer to TEDI than XSD for describing structured data.

However, if the world really has the will to dump XSD, we might as well move to JSON at the same time :-)

 


Posted May 10 2007, 08:17 PM by don-box

Comments

H wrote re: James Clark on XML, JSON and lots of other stuff
on 05-11-2007 2:18 AM
I believe, and who cares, that he absolutely nails it in observation that XSD is overkill. Who knows what will happen to JSON and other new layering, abstract model, sequence, or other hacks.

Everyone seem to be playing around with new ideas, to make it easier or more profitable I hope (rather than for fame or herd following).

You probably have yours, making Silverlight and WPF accessible to target audience of another offering, hence integration, developer and JSON focus.

Other people have similar but more concrete business requirements and so on. We all look to target our own share of developers, business segment and make it all easy to transition, understand and/or adopt.

But I also wonder how many people stop and think: 'lets expose our internal schemas'.

And from that in-depth and fun post, I would only suggest both messaging infrastructure and payload need work, ideally together.

And talking of other areas where really bright guys are, shouldn't there be a storage engine team few corridors down from you?

They already dealt with similar problems and satisfied the world with a first data operating system if I ever saw one, it transitions different domains pretty well. And it has a huge research and development body of experience behind its work and achievements.

And I wish Just a Giant of man was still around, with all of us.
Dimitre Novatchev wrote re: James Clark on XML, JSON and lots of other stuff
on 05-11-2007 5:00 AM
>However, if the world really has
> the will to dump XSD, we might as
> well move to JSON at the same
> time :-)

No problem, in the XSLT land we are ready for this move ("Transforming JSON"):

http://dnovatchev.spaces.live.com/Blog/cns!44B0A32C2CCF7488!358.entry

:o)

Cheers,
Dimitre Novatchev
Jason Haley wrote Interesting Finds: May 11 2007
on 05-11-2007 9:04 AM
Bob Denny wrote re: James Clark on XML, JSON and lots of other stuff
on 05-13-2007 8:08 PM
XSD and friends -are- overkill. The typical result of European academics becoming involved, creating theie "design and decree" world. It doesn't work. At least not easily. All one has to do is recall the networking standards that came from this same gang back in the early 90's. Many hundreds of megabux of VC went down that rathole. It was all supposed to kill TCP/IP. Then there was STT, SEPP, and the combination, SET. All garbage from the same bunch. Look at the immards of Bluetooth and you'll die. I could go on and on.
Noah Mendelsohn wrote re: James Clark on XML, JSON and lots of other stuff
on 05-17-2007 4:27 PM
> The brightest guy in XML-land
> (no offense Tim :-)) has just started blogging.

Yes, indeed.

> I love his analysis on both the state
> and role of schema and his followup on JSON.

Two terrific posts, including much that I'd hoped to say myself someday if I could ever get my act in gear and start a blog.

> However, if the world really has the
> will to dump XSD, we might as well move
> to JSON at the same time

Ah, no. Not because I have anything against JSON if it makes certain low-end data oriented scenarios easier, and not because I'm here to defend XSD. Rather, XML and to some extend XSD (and RelaxNG and Schematron) bring the industry something absolutely crucial, that JSON loses, and that's a unified framework for data and documents. I'm talking about mixed content (rich strings), an architecture where order can matter when you want it to etc. Losing that would be a huge mistake, and I don't see the JSON camp "getting it" on that.

Cheers! How you been, Don?

Noah

dbox wrote 关于James Clark的XML, JSON和其它许多事情
on 06-05-2007 8:50 AM
?xml?????????(????Tim?)?????blog??

Add a Comment

(required)  
(optional)
(required)  
Remember Me?