[
Lists Home |
Date Index |
Thread Index
]
8/3/2002 1:19:03 PM, "Didier PH Martin" <martind@netfolder.com> wrote:
> It's exactly what
>Cassandra, oops, Simon has been yelling about -- *programmers* can use
>proprietary APIs that -- deep down out of sight of anyone -- happen to
>use XML as an object serialization technique. Pardon me if I think
>the reference to "XML" in the headline was B.S.
>
>Didier replies:
>Precisely and two cause may be behind the future success of that:
>a) money
>b) ease of use
I'm sure you're right. My point is that this is a devil's bargain:
You've sold your soul to a proprietary technology that happens to
use XML to solve the vendor's problem, you haven't leveraged XML to
solve your problems. So, when your integration-via-proprietary-API
doesn't scale, or turns out to be insecure, or assumes a level of
network reliability that wireless can't provide, or doesn't work with the
next version of the other product, then your choices are to
a) wait for or buy into the vendor's update that solves the problem
with more proprietary voodoo, or b) start over (or rather, your
successor can start over after you get fired).
That's what I thought XML was supposed to take us away from.
Of course, I would agree that the the way forward is
to build cheap, cost-effective, easy to use development tools
that expose XML rather than hiding it. That's the challenge
we face as I see it, and that will require
- simple specs that can be implemented and understood at
a reasonable cost in time and money
- a clear understanding of both the "programmer" API/RPC/IDE
mindset and the document/stylesheet/authoring mindset.
- a whole lot of creativity, hard work, and willing to
work together.
|
