Re: [xml-dev] Towards XML 2.0

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

On 9 December 2010 11:12, David Carlisle <davidc@nag.co.uk> wrote:
> On 08/12/2010 10:03, Andrew Welch wrote:
>
>> The typical problems that I see that frustrate people to point of
>> hating XML are just because they've treated it as a string - encoding
>> issues, broken namespaces, lone amperstands and left angle brackets
>> etc.  What appears to be a simple change to the file (say using a
>> regex) breaks it completely, and for encoding issues this can be a
>> real pain to track down.
>
>> ...
>
>> The 'next xml' needs to be treatable as a string... JSON is just a
>> string, right?
>
> Clearly I'm missing something,

No I'm just not making my point very well... :)

> because I can't see how  lone < and & messing
> up an XML document is really any different from mismatched { or [ messing up
> json?  I can see a qualitative difference between xml and html parsing,
> where html (especially html5 flavoured html) is designed to silently recover
> from more or less any arbitrary rubbish thrown at it, but json, like
> javascript (and xml) has stricter parsing rules doesn't it? and so
> consequently requires some more care in construction.

If it was just a string, you could construct the xml by concatenating
strings and then write it to disk, job done, no problems.

Currently you have to use a proper serialiser (because of namespaces,
encoding, escaping etc), and doing that requires a fair amount of
effort - you have to use (and learn) a library of some kind and write
lines and lines of code... all of which baffles and frustrates a
non-xml dev when all they want to do is create some simple xml.




-- 
Andrew Welch
http://andrewjwelch.com