Re: [Json] Schemas & so on

[Erik Wilde <[e-mail]>]

hello martin.

On 2016-05-01 16:30, Martin J. Dürst wrote:
> Sorry for playing the devil's advocate, but asking the same for XML
> would give a lot of different opinions. I wouldn't expect the schema
> landscape (if it developed) to be much different, because there's a
> large span between simplicity and expressiveness that can be covered,
> and a lot of notational choices.

yes. and XML land was/is different because often, the document models 
are complex enough that some schema really is needed, and DTDs are just 
too primitive to be useful.

but seeing the popularity of JSON in APIs and how often these also 
specify non-trivial and non-rigid "schemas", part of me thinks that 
given a language that allows to better "document" the grammar part of an 
API, some not-so-small subset of API devs/users might find that useful.

> In addition to that, many people use JSON because they don't want to use
> XML (if they don't have any actual experience, they at least heard that
> it's somehow "problematic", and that often extends to schemata). So the
> JSON's users tendency seems to be to try and avoid schemata.

to some extent, that's certainly what's happening. but part of that may 
just be that so far, there simply isn't a good language to use.

so they add a bunch of examples. which of course are not painting the 
full picture. and consumers have to look at all of them and have to 
figure out the white spots. i think there is some sweet spot to make 
that a bit less tedious, without making it so complicated/complex that 
people will get XSD shock syndrome.

cheers,

dret.

-- 
erik wilde | mailto:[e-mail] |
            | http://dret.net/netdret    |
            | http://twitter.com/dret    |