Re: [xml-dev] A question of necessity

[Alexander Johannesen <alexander.johannesen@gmail.com>]

Arjun Ray <arjun.ray@verizon.net> wrote:
> I don't claim to understand this, but I see a prima facie example of
> "neurosis", as per a celebrated post to this list a while back:

So, you don't understand it, yet feel qualified to make a diagnosis? :)

> Let's say
> you had only non-XML-namespace aware tools available.  Would the
> information representation problem have been difficult?  Impossible?

Hard to tell. Some of these ontologies are third-party, who knows what
format and structure they'd come up with in lieu of namespaces. Maybe?
Probably?

> But I won't even try to
> guess the difference between sanity and insanity.

Why all this mental health stigma? :)

> I need a stylesheet to parse?

No, but I do.

> | Again, can I do it with plain names? Sure, but very, very differently.
>
> Okay.  Colonified names were not necessary.

But my initial point is that you can say that about any component of
XML. *Can* you do without entities? Sure. *Can* you do it without
attributes? Sure. *Can* you do it without character-encoding? Sure.
There's no problem out there that you can't do without some portion of
XML, just like there's no portion of XML that you can claim to be
truly necessary, it all comes down to context and need. You might
argue that most people don't need the complexities that come with
namespaces, and you might be right in that, but that doesn't render it
useless or even wrong.

I feel you ask and move the goalposts at the same time, and it's
turning into a semantic quibble around what you think is 'necessary'
or not. I never said namespaces in XML  (hence; the ruleset in the
spec) is perfect, but I certainly appreciate them a lot even if you
think I'm neurotic and insane.

How about we get rid of that default namespace can change? I reckon
most hate would dissipate.

> Note that the common-sense notion of a namespace as a controlled set
> of names doesn't apply.  (The phrase made an appearance in some of the
> trial balloons, I mean drafts, but it quickly vanished.)

Even in its current form there's useful notions, even if the
complexity and sometimes gotchas are annoying. I happen to like the
mechanism for trying out different vocabularies by simply changing the
namespace identifier; I've got hundreds of fairly complex XSLT files
working on a few dozen XML input files, and, well, it's easy to work
in. But that's just me. Clearly I'm in a minority, insane, neurotic,
and hence, probably wrong.


Cheers,

Alex
-- 
 Information Alchemist / UX badass / GUI wrangler for hire
 http://thinkplot.org     |  http://www.linkedin.com/in/shelterit