[
Lists Home |
Date Index |
Thread Index
]
"Cavnar-Johnson, John" wrote:
> Fundamentally, I don't see how the XML documents I define and use within
> my systems have a negative effect on you. Do you somehow fear that bad
> XML documents will drive out the good, even if your systems never come
> into contact with mine?
Yes, it is exactly this analogue of Gresham's Law that is a problem. Earlier
today, bryan (http://lists.xml.org/archives/xml-dev/200301/msg00312.html)
posted some XML which appeared to carry very little information as a
document in its own right, precisely because it was intended for a narrow
use by a single application. Such documents are designed for ease of direct
access--by a particular application--and for that very reason are
well-positioned to drive out more information rich documents which require
greater manipulation to use.
One point of the internetwork topology is that because you as a publisher of
documents do not know who might be interested in them, you must expect at
least the possibility that any system might want to access them, and that it
is not necessarily a violation of your documents if they can be put to
useful service by someone you never expected. If we abandon an internetwork
of documents rich in information which may be usefully manipulated in unique
combinations by uniquely valuable services, we all lose. We need other
people's data, and we need it badly enough for it to be worth our effort to
find it, comprehend it and instantiate it for our own uses. Just because it
is inconvenient for programmers we must not refuse to take the most
appropriate data on its own terms, and instead accept only what conveniently
comports with our datamodel.
Respectfully,
Walter Perry
|
