On Wed, Nov 27, 2013 at 11:00 PM, Tim Bray <tbray(_at_)textuality(_dot_)com>
wrote:
I listed some arguments against this in
http://lists.w3.org/Archives/Public/www-tag/2013Oct/0041.html and at the
moment I still believe them. Is there new information?
On top of that, I have no fear of anyone trying to change JSON in the
future; they would be resoundingly ignored by the community of
implementers. I speak as one who would love to add built-in date/time
literals but know that it won’t happen. -T
The JavaScript variant may stay the same but people will add features if
they find they are necessary to meet their requirements.
As was previously established, many implementations already support a list
of values so as to enable use in append only logs. I have added length
encoded blocks so as to avoid repeatedly base64 encoding binary blobs.
The Javascript world can ignore me, I try my best to ignore the javascript
world. But if people are implementing one of my protocols they will find
that while it will work perfectly OK with JSON, implementations will be
more efficient if they support JSON-A or JSON-B.
Suggesting that people will never change a spec is silly. Of course specs
will be changed.
--
Website: http://hallambaker.com/