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

[Andrew Welch <andrew.j.welch@gmail.com>]

>> If only there was a simple api for xml... that was actually simple :)
>> I'm sure the list could come up with a new one.
>>
>
> I hate the term "impedance mismatch", but it captures well the idea that processing data is always going to be far easier if you use a programming language whose type system matches the data model of the data.
>
> So manipulating JSON in Javascript, and XML in XSLT, are both always going to be a lot easier than manipulating XML in Javascript or JSON in XSLT.

That sums it up perfectly.

> Although the 4th design goal of XML was "It shall be easy to write programs which process XML documents", it's not clear to me that this goal ever really influenced the design. If it had been taken seriously, the spec would have contained a data model and it would have been a much simpler one than the models (plural!) that had to be retrofitted after the event. Namespaces in particular were designed with total neglect of this objective.
>

yep... especially from a Java perspective.  It should really say "It
shall cause non-XML developers pain and resentment to do something
trivial with XML documents" :)

One of the reasons for the 'hackable xml' idea that eventually
prompted the MicroXML discussions was because of this.  If you
stripped away some of the complications, you could/should have much
more simplfied apis... the MicroXML push got quite far, but so far I
haven't seen any simplified apis.  I must admit I tried and failed to
come up with anything better that what's out there already... perhaps
it's just not possible in Java.

Scala has a built in type for xml - can anyone comment on that?


-- 
Andrew Welch
http://andrewjwelch.com