POV-Ray : Newsgroups : povray.off-topic : The trouble with XSLT : Re: The trouble with XML Server Time
29 Jul 2024 18:15:43 EDT (-0400)
  Re: The trouble with XML  
From: Orchid Win7 v1
Date: 3 Mar 2012 17:05:00
Message: <4f52958c$1@news.povray.org>
>> Why would you want three different standard names for the same character?
>
> What's the chinese word for "quotation mark"? For that matter, why would
> you even want a "name" for a character, unless your text editor is too
> lame to actually store the character? Leap-frog into the 21st century
> and start using Windows Notepad to edit your text or something.

It's not possible to type exotic characters. But it's trivial to type in 
their names.

> Right. So you're complaining that XML is broken because your text editor
> can't handle Unicode? Just type the unicode characters directly into the
> text document, without using any code numbers.

...aaaand then watch it break into a thousand pieces because raw text 
files have no way of specifying what actual character encoding is being 
used. :-P

>> Well, if it were possible for one DTD to import several others, then it
>> could be pretty easy... As it is, constructing even one DTD is so
>> insanely hard that only the professionals can do it.
>
> ... says the man who programs in Haskell. ;-)

The problem is, the only real documentation on how to write DTDs is the 
antiquated SGML reference documentation, which is hardly easy going. 
 From what I've seen from other written DTDs, it's just damned hard, 
that's all.

>> What, a combinatorial explosion of DTDs for every possible combination of
>> XML applications that you might ever want to combine? No, that's just
>> silly.
>
> Why not? You get a combinatorial explosion of types in your computer
> programs for every new program you write.

Actually... no, you don't.

In any half-decent language, you can implement sorting and searching. 
Once. In a language that's not so good, you have to reimplement this for 
every individual datatype. And that's what the DTD situation feels like; 
endlessly reimplementing the same language specs over and over again. 
And these aren't exactly trivial languages... It would be frighteningly 
easy to come up with an XHTML+MathML DTD where the XHTML part doesn't 
actually match the stand-alone XHTML DTD, for example.


Post a reply to this message

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.