[
Lists Home |
Date Index |
Thread Index
]
Not oddly, some of the first Hytime tutorials that
Neill Kipp and Steve Newcomb presented (and which
I have old paper copies of) took up exactly the
problem of the notation: that is, when one is
addressing into a film or audio file, one is not
addressing into a markup so markup based addressing
even infoset is not sufficient. The structures
aren't the same. Each notational system needs
its own infoset and by some magic, these should
all be resolvable into a more abstract descriptive
information set, which is what a grove is as I
understand it: nodes and properties, you get to
name the names.
It isn't that complex (except the names chosen
for the grove sets themselves); it is... abstract.
Hytime got lost in ... time. I have some measure
of the fault there. We were all musicians and
time means a lot to us. :-)
The speed of light as a constant separates the universe
into gulfs of information where evolution depends not
only on temporal proximity to the origin (the big bang)
but the limits on interactions (exchanges of taxons)
among non-near neighbors. Intelligence only emerges
in a limited band of that space/time coordinate set.
The reason we haven't heard from the neighbors (alien
plots not withstanding) is that there are no
"ancient advanced civilazations". We are all in about
the same phase of our intellectual evolution and the
universe as a whole has not reached the point of enabling
the locals to understand how to fold space and time
to resolve the distances. We aren't mature and
we aren't capable. But we will be. What we will
use for money when we are will be fascinating to see.
len
-----Original Message-----
From: Didier PH Martin [mailto:martind@netfolder.com]
Len said:
We come back to groves for the same reasons
the original group ended up there. Until one has
a neutral means to describe properties of nodes
in a sharable descriptive language, it is tough to
decide when we are talking about the same "type"
of thing being addressed from multiple contexts.
Didier replies:
Yes indeed this is needed by any "software" interpreter. Moreover, by
experience, I discovered that Grove plans able to go beyond the document
model and move up (or down depending on the perspective) to the semantic
level brings more productivity and more readable code. For instance, If,
from a marked document I want to link to a particular frame contained in
a movie (1) the interpreter needs to resolve the impedance mismatch
between the types.
On the other hand I may rely on an addressing schema and let an other
interpreter resolve the issue. But in the last case, the movie object
interpreter will need to present the same addressing model (i.e. the
same name space convention). For instance to present a tree and a
addressing scheme like Xpath (i.e. a hierarchical name space). Is
something like Xpath sufficient to do that? I Don't know but what I
remember is that people got rebuffed by the complexity of function based
addressing ( a la lisp) as proposed by hytime (and also covered in DSSSL
specs).
In any cases, we will have to deal with different document types and I
am not talking here of markup document type but more of different
entities like HTML document vs. video or sound documents. The main
problem the web will have to face during this decade is to be able to
link text based documents to multimedia document. So to speak, to have a
link contained in a document to point to a particular movie scene or a
particular song part.
I guess the community has lost sight of the end goal and entered in
decadent Byzantine fights :-). I just hope Simon's effort will bring
back some common sense and help prepare the future and most of all, that
the group will use the lessons learned from the past.
(1) I know for several it is science fiction but, as you know, some try
to do that and documents are not all based on markup
|
