[
Lists Home |
Date Index |
Thread Index
]
On Wed, 27 Oct 2004 11:41:47 +0100, Michael Kay
<michael.h.kay@ntlworld.com> wrote:
> All precedents suggest that once a technology is entrenched, it only gets
> displaced if (a) there are important jobs that it can't do, or (b) there is
> a trusted alternative that is 1000% better. With XML, there's not the
> slightest evidence that either of these events is imminent.
Maybe I'm warped by repeated exposure to the MBA crowd's favorite book
on the subject -- Clayton Christensen's THE INNOVATOR'S DILEMMA -- but
that doesn't sound right to me. Christensen talks about the
distinction between "sustaining innovations" (making things
incrementally better/faster/cheaper for existing customers) and
"disruptive innovations" (that are often qualitatively worse that
existing products, but dramatically cheaper/simpler enough to make
them attractive to new customers). To use one of the technologies
Christensen talks about, computer disk drive technology of the last
few generations has done the job well, gets cheaper all the time, BUT
keeps getting displaced in the next generation by smaller/cheaper
technologies that don't initially work even as well. Why? Because
the next generation enables new products that did not seem even
feasible, much less "imminent" to the people who understood the
previous generation. Think of a 10GB drive of 10 years ago vs the
ones in an iPod. The only seriously profitable market for disk drives
right now AFAIK is the one for the little things that fit in mobile
devices, but the better/faster/cheaper versions of desktop PC disk
drives are a low margin commodity.
Could XML be displaced like this? Christensen's analysis suggests
that it would be only if the replacement is so much cheaper (in
various senses) than XML so as to enable new classes of applications
for which XML isn't suited, while being almost good enough to do what
XML does.
So, I'm interested in determining the classes of applications for
which XML is plausibly useful but not very well suited in practice.
Hand authoring is one; that's supposed to be a strength of XML, but
the RSS/Atom experience has very clearly shown that it is far easier
to hack software to process tag soup than to motivate and educate even
fairly geeky people to write valid XML.
Is there a market opportunity here for a disruptor? Don't know ... I
could certainly see a market for Office plugins that would let people
leverage the power of its XML support but via RELAXNG rather than XSD
and/or a human-readable XML mapping along the lines of RNC for
situations where the WYSIWYG interface doesn't give the control you
need.
Wireless is another -- that industry wants standards, doesn't want to
have to invent their own, likes a lot of what XML offers, but chokes
on the bandwidth/processing requirements. Also, there's an incentive
there to strip down the XML APIs to the bare bones that are really
useful, and not support stuff that was put in DOM or whatever to keep
Netscape happy :-) Some mutant descendent of XML specs is almost sure
to evolve in the wireless industry, and could displace XML in the rest
of the industry. High-speed web services is another: SOAP already
implicitly subsets XML (no DTDs and all the cruft that goes with
them). People are throwing specialized hardware at the XML parsing
overhead in firewalls, etc. and it seems likely to me that this
industry will develop efficient infoset interchange standards in order
to create a "rising tide to float all their boats". Disruptive
innovations there could also infiltrate the rest of the industry.
I'm not sure that there are important jobs that XML can't do, but I
think there are important jobs that XML could do better with a bit of
profiling/refactoring/optimization ... and if this doesn't happen then
I think it very likely that "disruptive innovations" will come along
to do those jobs better than XML.
|
