RE: [xml-dev] Fixing what's broke

[<vojtech.toman@emc.com>]

> > May I ask if the success of XML is attributed to its verbosity?
> >
> > Introducing terse but complex syntax like skipping/stripping end tags sound
> > like 're-engineering' a traffic junction to have no lights and assuming
> > everyone understands when to stop/move.
> > You save on traffic lights, but your traffic slows down. We can always point
> > people to a detailed manual instructing how to operate, but yeah, how many
> > humans do that diligently.
> 
> But hasn't HTML succeeded even more so than XML? And hasn't that partly
> been down to it making assumptions about what 'everyone understands'?
> So maybe XML would be even more successful - and, more on topic, get more
> take-up on the Web - if it became more like HTML which by and large would
> mean relaxing a lot of things.

In the web browser, it is great that the HTML parser can recover from the mistakes in the markup and still show me a more or less faithful rendition of the page. That's the obvious value of the relaxed parsing rules. The web page contains a lot of visual redundancy and I, as a human observer, typically don't care if two paragraphs are rendered as one, or if the table looks a little different in browser X. I can still make sense of it.

But in XML, if I look at it as a data exchange format, I am not sure if introducing more relaxed parsing rules would do any good. More options often lead to confusion and to more room for making mistakes, and the last thing you want is to get unexpected (maybe completely correct wrt the parsing/recovery rules, but unexpected to the data author nonetheless) results from the XML parser on the receiving side.

(Note: this wasn't about verbosity; that is a whole different subject.)

Vojtech

--
Vojtech Toman
Consultant Software Engineer
EMC | Information Intelligence Group
vojtech.toman@emc.com
http://developer.emc.com/xmltech