[
Lists Home |
Date Index |
Thread Index
]
.. the crazy guy speaks again ..
Have you ever been in a meeting where everyone suddenly agrees that
something is correct, and then about five minutes later everyone agrees that
it's completely wrong?
That's XML. At first everyone thought representing everything in XML was a
great idea, and now it seems that everyone is discovering why humans have so
many symbol systems. Perhaps the past 50,000 years was not a big mistake
after all.
Adam Boswoth told me a story about "pattern" representation in the early XSL
implementation at Microsoft. This was the precursor to XPATH. An engineer
proposed a 100% XML description that looked much like a reverse-polish
calculator. Adam had the foresight (this was several years ago, mind you)
that this was NOT user-friendly. And so there's a non-XML language for
navigating to XML data.
As a programmer, I'd rather not have to reintroduce myself to the quirky
creatures known as sed and awk. XML is so enticing.. a generic parser for
everything.. ah, that sounds so nice like a stroll on the beach. I can hear
the waves crashing.
WAKE UP!! says those darn human factors. good for programmers != good for
people
As far as schemas go, XML is still in diapers. Real data processing
requires types, relations, and constraints. Hierarchical data was abandoned
at least 20 years ago. Lo and behold it's back. Same salesman, different
suit.
Today's relational databases are way ahead of whatever imposters we have
today in the XML schema realm. If only there was a way to 'export' a true
database 'snapshot' in a lightweight, open, and platform-independent manner.
Imagine if it supported queries and updates and multimaster replication!
XML does not address these needs in the slightest, except perhaps as a
message format for remote web services that may implement this
functionality. XML was simply not designed to be a database. XML is useful
for some things, but real data processing requires real databases.
Therefore, expect the various XML schema flavors to disappoint. XLink is a
joke. You're simply asking XML to do too much.
XML is a way to put data into a flat file. XML has nice extensibility
features that come with the price of much larger files compared to
comma-separated files or some other terse format. XML lulls developers into
a false sense of security until they have to process large amounts of data,
first faltering with the DOM (out of memory and horrible performance), and
then going insane while coding against SAX.
HTML is good for its domain. XML is a replacement for HTML. XML is not a
good technology when it comes to large-scale data processing (even
medium-scale, sorry for not being pecise), period, end of story.
And for those who think XML is the ideal format for EVERYTHING, please
return to the top of this email. Be careful not to loop endlessly because
I'm sure you have real work to do.
----- Original Message -----
From: "Simon St.Laurent" <simonstl@simonstl.com>
To: "The Deviants" <xml-dev@lists.xml.org>
Sent: Wednesday, January 30, 2002 8:33 PM
Subject: Re: [xml-dev] misprocessing namespaces (was Re: [xml-dev] There isa
meaning, but it's not in the data alone)
On Wed, 2002-01-30 at 21:51, John Cowan wrote:
> Paul Prescod scripsit:
>
> > One of the major arguments for moving away from DTDs was that an
> > XML-based syntax would be simpler!
>
> Simpler for programmers, who have bbls of XML parsers to use on them.
> Not necessarily simpler for schema authors.
>
> I wish XML Schema had a non-XML format...
Yeah. I've come a long way from my early enthusiasm for XSchema / DDML.
I think part of the vision that was somehow lost along the way was the
prospect of using XML tools to build, process, and store schemas. DDML
had high hopes for the (eternally-delayed) XLink, and was designed to be
simple enough to be processable that way.
At this point, I don't think schemas (such as they are) have proven
especially tractable this way. I'm moving in general to a notion of XML
as a canonical markup format, and think things like the non-XML format
for RELAX NG are a step forward at this point, not a step back.
(Examplotron's use of XML instance syntax is brilliant, but it's a
special case.)
--
Simon St.Laurent
Ring around the content, a pocket full of brackets
Errors, errors, all fall down!
http://simonstl.com
-----------------------------------------------------------------
The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
initiative of OASIS <http://www.oasis-open.org>
The list archives are at http://lists.xml.org/archives/xml-dev/
To subscribe or unsubscribe from this list use the subscription
manager: <http://lists.xml.org/ob/adm.pl>
|
