Re: XML and Lout already work well

[Giovanni Zezza]

Il Thu, 20 Sep 2001 18:10:14 +0100, address@hidden scriveva:

>I'm not sure that syntax helps much. If @Gap and \gap mean 
>different things in different typesetting languages, some 
>confusion may remain if you use both despite the @ vs. \ 
>difference.

Exactly so: syntax (let's name it so) doesn't help much in mutual
understanding (or "interoperability"). Having <gap></gap> in Lout looking
the same as <p></p> in HTML doesn't help much in understanding what this
damn <gap> will do.

So much the less, if you have 11 vowels, it doesn't help writing them with
only 5 different symbols, just because so you will share the same roman
alphabet as every one. It would be better (from a learner point of view) to
have them written with 11 different symbols, no matters how much this may
divert you from others.

It doesn't help much forcing everything and all in one syntactical (let's
name it so) model, if another one would fit better. For there is something
syntax does help in: it helps in doing easier and better what you have to
do, and saying more simply and directly what you have to say.

>Of course, XML transformation tools don't do magic, they just 
>help doing a common task in a slightly easier way than if you had 
>to do everything yourself. That a bicycle can't be used to go to 
>the moon does not mean that a bicycle is not a useful way to 
>go somewhere that's closer.

The name of your "bicycle" is "parser". That's all. XML transformation tools
give you a parser. What to do with it is up to you. It, by itself, doesn't
go anywhere.

My personal opinion is that having a parser doesn't help so much in doing
whatever transformation you want to do. Writing the parser is usually the
lightest part of the job. By the way, you have already one, inside Lout
compiler code (I don't know how easily it can be isolated, but I'm inclined
to think it can).

>XML hype may have obscured for you the real benefits, in some 
>but of course not all circumstances, of a standardised way to 
>represent tree data structures in text files.

"Data structures" are data structures, languages are languages. I may like
(and I like in fact) to have my text written in some structural notation,
yet I see no need to have a language (a typesetting one, too) looking like
my text (on the other side, if you like literate programming, write some
LoutWeb; or maybe Lion could be suited to write Lout, I don't know).

Having all stages in the document production chain looking the same doesn't
add anything to productivity, or "interoperability", or communication, or
mutual comprehension between different departments or whatever. Put SGML (or
XML) where it's best suited for and where it can actually add some value to
your work: at the start of the chain. From here your XML transformation
tools may give you whatever you want: RTF, TeX, Lout and so on. If what you
get were another XML, why did you move from here to stay here?

Ciao.