Re: [Json] Proposed minimal change for duplicate names in objects

[Nico Williams <[e-mail]>]

On Sat, Jul 6, 2013 at 11:58 PM, Tim Bray <[e-mail]> wrote:
> I’ll assume you’re right when you say dupe detection has been measured as
> expensive at run time, but I’m thinking that if I were writing a reader I'd
> implement a hash table with a test-and-set method, so I admit I'm surprised
> by the finding.  I think I’d need to see a little more research before I’d
> accept that as a given.  -T

Dunno about Tatu, but I think the vast majority of JSON usage must be
coupled either with generic hash tables (objects/dicts/maps/whatever)
or structs (with associated schema).  For the latter dup detection may
well be more expensive than not, but probably not that much more
expensive.  For the former dup detection is essentially free.

For other use cases, like Stephen Dolan's "launch missiles" example,
dup detection might well be expensive, but maybe JSON is the wrong
tool for those use cases.

The CPU cost of dup detection doesn't concern me.  It's the
architectural implication that does: minimal state streaming parsers
are out.

I have to ask myself: are there use cases where I would want a minimal
state streaming parser?  and do I mind losing the ability to have such
a thing?  My personal, tentative answers: "yes" and "probably not,
maybe in such a case I should not use JSON".  Others will no doubt
have different answers, which brings us to...

...more retreading.

I still think the best thing to do is note the divergent
interpretations and publish Internet JSON.

Nico
--