[Sigia-l] XML alternatives
Listera
listera at rcn.com
Mon Jan 3 01:53:01 EST 2005
b0b d0n:
> Any stories/experiences/comments about any of these ML's implemented in your
> organization,people??
Comparing them most/all would be a monumental task.
I don't want to start 2005 with a rant on how unnecessarily complex XML has
become and how many variants it has spawn. :-)
There are two things to consider when choosing a ML:
First, taking any text and/or data and tagging it for export is generally
speaking quite easy in any language. Most popular scripting/dynamic
languages already come with pre-built libraries. Unless you are forced to
get into the arcana/complexity of XML, this isn't that difficult.
Second, choosing the 'right' ML should be done in the context which the
exchange will be happen. For example, if you have an internal system where
data is generated and consumed entirely within your organization, especially
by legacy artifacts, XML or even its variants may not be the best approach.
Tagging does add overhead, exporting, importing, parsing, converting, etc. A
very simple binary or even ascii format may suffice. Flash remoting used in
RIAs, for instance, is a lightweight client-server protocol that allows
passing objects/arrays back and forth without the overhead of XML. Even when
the XML consumption capabilities of Flash 7 has been greatly improved, many
Flash developers still prefer Flash remoting for speed and memory handling.
So when you're deciding on an architecture for data/info exchange with
potential partners you'll need to consider their current and future needs.
The ideal would be to totally abstract data I/O into replaceable modules so
you can use verbose XML when you have to or a lighterweight variation when
you can.
Ziya
Nullius in Verba
More information about the Sigia-l
mailing list