Lists Home |
Date Index |
XSLT is often advertised as document to document for
some n of document. He picked the boundary cases we know are
there and are difficult. He made the complaint we hear
over and over again ("I need procedural constructs") and
that is a lot like what I hear when it is claimed XML
concepts don't match mainstream computing. Maybe
mainstream computing is *behind the times*. It is easy
to be elegant and out of step with the tools on the desk.
That's the dicey bet we make in this business. To an
OOPMan, XML is a throwback to a data object. To a
DataMan, XML is a step up from CSV. BTW: the XML SIG
did have these arguments more than once. Tim Bray
said it best with, it hits a good compromise for
enabling interoperability. To which I add, Thou
Mayest Not PROGRAM in XML, just with it.
What I am asking per the original posts is did any of that
happen as a result of choosing bottom-up design over
top-down design in XSLT? Hmmm. Did that choice ever really
get made? Or was it a case of "we have all these nice
notions from DSSSL and we'll simplify that and see what
is left, toss it out there, see how it is used and what
is complained about by whom, and wait for the next release"?
Is it realistic to retrofit complex systems concepts to
a human process (non-linear humans, remember) and then
inquire of the output if it is broken? I don't think
it is. It is completely legitimate to take a test case
and see if one can test a clearly stated requirement.
In too many cases, such tests cannot be made because the
requirements are stated at a level that make it easy for
the human process to declare a minimal victory at some
point, but not to determine if the feature set of the
product is complete. No one is right or wrong here;
he says it won't easily do what he has in mind, and
unless one cites chapter and verse of the design
requirement, no one can say he is wrong. We expect
requirements-as-controls to emerge from use. If
that is so, XSLT fails by his example. If we question
his example, we have to prove it is out of scope.
From: Elliotte Rusty Harold [mailto:email@example.com]
He is completely confused about what the intended purpose of XSLT
actually is. It was never intended to do what he wants it to do. It
shouldn't be a surprise he has trouble. Nor should this be considered
a knock on XSLT, since none of his use cases are something XSLT was
ever intended to handle.
This article is the rough equivalent of a review of giving a toaster
oven a bad review because it won't wash dishes.