Lists Home |
Date Index |
Simon St.Laurent wrote:
> It's an understatement to say that I'm sick and tired of programmers and
> database developers trying to cram what XML does into their particular
> views of the world - and worse, to inflict those views on XML itself
> through a standards process gone rotten.
It's a fact of life that most programmers think in terms of the
procedural logic they're trying to write and the user-visible effect
they're trying to achieve. The notion that programmers should put extra
work into trying to achieve data interoperability has *always* been a
There was a time when computers had "FORTRAN files" and "COBOL files"
and FORTRAN programs couldn't read COBOL files, and a lot of programmers
resisted moves towards a common filesystem format because they weren't
having a problem. Then there was a time when most OS filesystems came
with a built-in record structure that you really couldn't get around,
and if you wrote a sequence of bytes into a file it might come back
different because of the record mechanics. Selling Unix's world-view
that a file is a place you write bytes into and then read them back was
tough because the record view was less interoperable but handier for
many application programmers.
XML is just another chapter in the same long story. It usually turns
out to be a win to make programmmers' lives more difficult by forcing
them to deal with data as the data wants to be, not as their application
wants to see it. Usually they complain, and demand tools to make it
easier to get at the data their logic demands. Seems like a fair demand
I personally am right now in the process of writing a bunch of perl code
to read some XML and manipulate the data it represents and generate some
web output, and I also found it irritating and tedious to extract the
pieces I needed. Doesn't mean that XML isn't the right way to store the
upstream data, but it also doesn't mean that I'm not allowed to complain
and ask for better ways to do it.
So up to this point, I'm fine with what Adam is saying. -Tim