[XML-SIG] DOM: Multiple proxy problem

Fred L. Drake Fred L. Drake, Jr." <fdrake@acm.org
Wed, 7 Oct 1998 10:21:28 -0400 (EDT)

Markus Fleck writes:
 > This must be what Tim Berners-Lee meant when he said (at WWW7):
 >    "You need to build a system that is futureproof; it's no good just
 >    making a modular system. You need to realize that your system is just
 >    going to be a module in some bigger system to come, and so you have
 >    to be part of something else, and it's a bit of a way of life."

  Nice quote.

 > I found this to be a very nice "feature" of XML, but I also think that
 > this implies that strict, formal DTDs shouldn't generally have the
 > importance that they have now. Requiring a DTD for each XML document

  Actually, I don't see how DTDs "solve" the "futureproofing" problem, 
and I'm not sure that "futureproofing" is well defined.  (If it is,
I'd be very disappointed!)  There's a lot we can predict about the
future, and a lot we can be wrong about.
  Given XML+Namespaces, it probably isn't important that any given
document has exactly one document type, and the form taken by document 
type definitions will most likely change.  However, I don't think the
need for some formal notion of document types will change anytime soon 
(am I setting myself up for a big fall? ;).  The need for specifying
"intended interpretation" for document types will remain.  DTDs
(including the non-formal part) are one way of doing this, and offer
particular benefits and limitations.  They're also the best shared
means of doing this at the time -- DCD and XSchema just aren't quite
here yet.

 > may solve the problem of "futureproof" file formats, mostly because
 > it can serve as some kind of documentation, but it still doesn't enable
 > a legacy XML application (using an older version of a DTD) to
 > non-destructively process data that has been created using a
 > more recent, but upwards-compatible version of that DTD.

  Another concern is with well-formed XML that doesn't conform to a
DTD.  It is entirely reasonable to handle instances which contain
completely ad-hoc tagging or elements affiliated with alternate
namespaces.  An application need not be "legacy" to have problems with 
input data when attempting to perform minimal transforms!

 > But then, maybe the future isn't really important.

  Maybe not.  But I'll let my kids decide that after I die.  ;-)


Fred L. Drake, Jr.	     <fdrake@acm.org>
Corporation for National Research Initiatives
1895 Preston White Dr.	    Reston, VA  20191