Lists Home |
Date Index |
I'm betting that a procedural language won't
absorb a data description language based on a
syntax description. It defeats one purpose
of having the syntax-based spec: to get
information out of an old system and into a
new one come replacement time. One might
ask why the InfoSet exists at all and that
might be more pertinent to her prediction.
I don't think performance is critical to
the success of XML. It is to Java and C#.
Are you predicting a new Lisp?
From: Mike Champion [mailto:firstname.lastname@example.org]
4/25/2002 11:56:51 AM, "Bullard, Claude L (Len)" <email@example.com> wrote:
The bit I found most interesting was from Daniela Florescu: "I don't think we will
have good performance as long as we have people marshalling data from XML to
Java and back," Florescu said. She went so far as to predict that eventually,
an extension of XQuery will replace both Java and SQL."
That sounded totally outrageous at first glance, but think about it a bit:
Java, SQL, and XML seemed like totally different beasts not too long ago,
but many developers find themselves having to use all three at once. Java
class libraries to a reasonable job of encapsulating the data behind OO
abstractions, but the OO model is looking less and less like the final word
on the subject: both RDBMS and XML encourage one to think of data liberated
from the programs or objects that produce and consume it, and approaches
such as "aspect oriented programming" and "Demeter" (neither of which
I understand well enough to have an opinion on) are at least enhancing, if
not challenging, the OOP paradigm.
So, I can kinda, sortof see a programming model that handles declarative
stuff easily, encourages a functional programming style, handles both
XML and relational data "natively", and allows data to be either encapsulated
in objects (or aspects?) or exposed for generic manipulation. Whether
XQuery is that model is quite another question.
So, have I been breathing bad air on airplanes too much lately that killed
a few bazillion brain cells, or might there be something to this idea
of convergence? Is Java more likely to absorb XML (XML InfoSets can be
serialized with curly braces more efficiently than with angle brackets)
than XML is to absorb Java? Come to think of it, isn't this common language
for code and data Curl's selling point?
Anyway, perhaps someone who actually went to the conference or knows what
the participants had in mind could help sort this out ...