Thomas 'PointedEars' Lahn said:
Lasse Reichstein Nielsen wrote:
Utter nonsense. It is not that hard to write Valid markup, may it be HTML,
XML, or XHTML, if you put a little effort in it.
It's not *hard* in any way. Anybody who can close the parentheses
correctly on a Lisp program should have no problem writing XML.
You do have to stay attentive, making sure you don't misspell a tag or
attribute name, that you remember the "/>" on empty elements, that you
write only correct values into attributes, etc. That's not hard, just
demanding.
XML is also quite verbose, especially if you try to give tags
meaningfull (and thereby more easily readable) names. And you have to
write the end tag every time, even if the name is exactly the same as
the start tag. An assisting editor could help you here, filling in
the tag name when you start a closing tag.
If your editor doesn't understand the document definition (defined by,
e.g., DTD or XMLSchema), then it can't help you correct typos. If it
also assist you filling ind end tags, you will even create correctly
matched tags when you misspell a start tag.
All this means that writing XML soon becomes mindnumbingly boring.
The combination of requireing attention to detail *and* being boring
is the classic recipy for making stupid errors.
Possible? Absolutely. Hard? Hardly! Slow and errorprone? Check!
So, to make writing XML even remotely pleasent, the editor should
at least be able to check the syntax against a document definition.
Syntax highlighting is also important, since black-on-white XML
isn't very readable either.
XML is the lowest common denominator of formats. It is extremely well
suited for machine reading, and just as unsuitable for human
reading. Sure, it's easy to learn the basic structure, since
everything is really the same. It's also very hard to read any larger
document *because* it all looks the same. Having to pick out the tag
names, i.e., words, in a block of text, i.e., other words, begs for
syntax highlighting, at the least. HTML was actually better, since it
allowed you to write your tags with capital letters if you so
preferred.
But I guess the verbosity is really the killer for me. The boredom.
It's not like parsers are hard to build these days. Dumbing everything
down to XML makes sense for machine-to-machine communication, but for
humans writing information to computers, a language specific to the
application domain can make everything so much easier that it is
ridiculous.
(The development of syntax:
Math:
x |-> x * 2
Scheme:
(lambda (x) (* x 2))
MathML:
<lambda>
<bvar><ci> x </ci></bvar>
<apply>
<times/>
<ci> x </ci>
<cn> 2 </cn>
</apply>
</lambda>
/L