[
Lists Home |
Date Index |
Thread Index
]
"Simon St.Laurent" wrote:
> Elliotte Rusty Harold writes:
> >Does this have any use for XML? Is there any point to letting the
> >root shift from one node to another while still keeping everything in
> >the tree? Does this enable any processing models or solve any
> >problems?
So far this seems to me a promising premise for solving the concurrent
markup problems which are the specific focus of Patrick Durusau's work. Out
of the necessity to process overlapping markup I have used a solution based
on alias tables, where each table entry is a reference to a well-formed XML
structure of at least one element, and often a large and complex document.
In processing practice this may not be much different from Patrick's method
of proceeding over an aggregate tree and shifting the root to each new node,
though it is not nearly as clean and persuasive conceptually as Patrick's
premise that every node is (for processing) a root. In any case, the JITT
concept is promising, which I could never say for Patrick's previous
proposal for processing concurrent markup, the Bottom Up Virtual Hierarchies
about which I am afraid I have had unpleasant things to say on this list,
and elsewhere, and which I am now grateful that Patrick has withdrawn in
favor of a far better model.
> (LMNL is described at http://lmnl.org - I'm not sure that's come up
> previously in this thread.)
Further plug for the XML SIG: Gavin Nicol will be presenting LMNL to the
SIG in New York on December 17th. Taken together with Patrick Durusau's
presentation of JITT the following month, this should provide nourishment
for new avenues of thought about the processing of XML or XMLish structures.
As with the JITT presentation, the session on LMNL will be open to
non-members of the SIG who would like to reserve places, and information on
doing so will be posted here closer to the date.
|
