Re: XML

[thi <ttn@mingle.glug.org>]

Micah Yoder writes:

 > I assume that data created with one toolkit can be read/modified by
 > others.  I just don't *quite* understand how this is possible when
 > everything is stored in text files, which really would be inefficient
 > for a large school.

a text file, especially w/ all these tags, has regular structure that a
program can call a function to internalize.  after processing, another
function can be called to generate (and write to disk) the external
representation.  if many programs use the same two (library) functions,
they need only worry about their internal algorithms and confidently
relegate translation to and from the external representation to those
functions.  the more these libraries are used by programs, the more they
will be exercised with respect to their interoperability.  this normally
leads to higher quality code (yay).

it could be that these translation functions have hooks that allow them
to call other functions that can translate (say) gzip'ed files.  at the
minimum, one could do:

	#!/bin/sh
	temp=$0.$$
	gunzip -c $1 > $temp
	myprog $temp
	gzip -c $temp > $1
	rm $temp

wrapping like this is standard (if ugly).

thi