[jdom-interest] Disruptive technologies

Simon St.Laurent simonstl at simonstl.com
Tue Jun 27 04:57:57 PDT 2000

At 09:13 AM 6/24/00 -0700, Josh Rehman wrote:
>I did not understand Simon's article. It appears to be a rather high level
>speculative article about the implications of XML. I believe that a
>restatement of his thesis, without jargon ('disruptive technology'?!), would
>be, "XML implies wider changes to internet technologies than most would
>believe." How is this thesis supported?

It is high-level - it's supposed to be high-level, and attempting to delve
into the details on all of the technologies involved would require a book
or three, and probably get the forest utterly lost in the trees.

>Maybe its my fault that I don't really understand what "XML demands
>"adaptive reuse" of existing infrastructure" means. Even with the examples
>that follow (which I assume are meant to illustrate what the term "adaptive
>reuse" means), it doesn't become clear. Again, its probably my fault since I
>am not up on many of the protocol initiatives that Simon mentions (XML-RPC,
>SOAP, new MIME identifiers!).

If you haven't participated in _any_ of those discussions, it's hard to see
what the big deal is.  If you've been attempting to read and keep up with
all of the messages on xml-uri at w3.org, xml-dist-app at w3.org, and
ietf-xml-mime at imc.org, it's pretty clear that there are some very large
problems at the intersections between XML and nearly every infrastructure
it uses or reuses.

Basically, XML as a generic data container has opened up possibilities for
transferring information that weren't anticipated when older specs are
written.  If that's not clear from the article, then something's wrong.

>For better or for worse, I have a hard time understanding solutions to
>problems that I haven't had, can't anticipate, and don't have a good way to
>visualize. And I'm a programmer! I can only imagine how well these kinds of
>arguments would go over with management staff, when talking about
>anticipating the reimplimention of core infrastructure (Scenario: 'OK Bob,
>XML is disruptive. Plan on replacing any software that even knows about the
>existance of HTTP, SMTP, MIME, or URIs. I know, its a big headache but you
>might as well bite the bullet. Why should you do this? Ha, I just told you!
>XML is a disruptive technology!')

I hardly expect managers to throw away infrastructure in favor of tools
that don't yet exist.  I would like programmers and protocol developers to
consider the changes that infrastructure may need to support large numbers
of applications that are only getting started.  I think it's pretty clear
from the end of the article that total replacements in the near-term are
not a likely solution.

>Don't get me wrong, I love abstraction as a tool to make concrete problems
>more tenable. But Simon's article was a bit too far in the clouds for me.
>But it just seemed a restatment of what every programmer knows: data
>structures are the core of any software system. Change these strucutures,
>and the system will drastically change. XML represents an oppurtunity to
>recast common data structures into a homogenous format, which improves data
>exchange mechanisms between disparate systems. Because of this, distributed
>systems will change, simplify, and become more common. Why all the drama?

Because the drama is real.  Visit the first few days of this email list's
archive, and consider the very real irritation that JDOM causes to folks
who created the DOM or use it regularly.  Visit the archives of the lists
mentioned in the article, and watch religious wars among people who claim
they're trying to arrive at consensus.

Yes, 'every programmer knows data structures are core of any software
system'. (Some, anyway!)  But every programmer has a different view of
those data structures, and a different view of how those data structures
should be integrated into the larger software system.

XML provides enough flexibility that these different views are clashing in
pretty much every arena of infrastructure development that XML is entering,
and rather violently at that.  To me, that signals 'disruptive technology',
and not necessarily in the pretty way marketers like to use it.

And kudos to Brett and Jason for keeping JDOM going through the initial
firestorm brought on by JDOM's marching in to DOM's turf.  They've managed
to make clear that different perspectives deserve different
implementations, making the need for change clear while staying out of the
key arena (browsers) that the DOM, well, dominates.  Bringing new
perspectives to existing (even recent!) technology is very difficult, and
not just for technical reasons.

Simon St.Laurent
XML Elements of Style / XML: A Primer, 2nd Ed.
http://www.simonstl.com - XML essays and books

More information about the jdom-interest mailing list