Lists Home |
Date Index |
"Champion, Mike" wrote:
> I'm not sure what to call this perspective (and no big company seems to
> advocate it), how about the "native XML dataflow/transformation"
> perspective??? (A better label is humbly solicited from anyone who sees the
> world this way).
John Cowan and I (not accustomed to beginning from a posture of agreement)
started a similar discussion at lunch on Tuesday which, as happens IRL and away
from the xml-dev arena, was interrupted and never resumed. The appreciation of a
syntactic dataflow/transformation perspective is not new, and in Homeric
studies--or, more broadly, studies of formulaic oral poetry--has been the
central question since the 1920's. Beginning with Wolf in 1798 the 'learned
Germans' meticulously constructed a poetics, and indeed an all-embracing and
self-consistent aesthetics, of Homeric epic based unfortunately on an
understanding of 'literature' which is alien to the compositional necessities
and therefore to the expressed nature of oral poetry. For the first century and
a quarter of modern scholarship it was presumed that the text was an author's
particularized expression of both widely-shared myth/legend/history (the Trojan
War) as well as of deeper, universally human impulses--Platonic forms, if you
will. These are, of course, key assumptions of Romanticism, which values most
the individual's expression of the universal and permanent. The vast edifice of
such scholarship was destroyed by Milman Parry's seminal discovery in the 1920's
that the formulaic phrases of oral poetry are the compositional mechanism for
providing the terms needed to narrate the story, in the combinatorial matrix of
the grammatical inflections required versus the metrical patterns of the verse.
Internalizing Parry's explanation of oral composition allows us to understand
the distinction between the 'underlying meaning' and the 'surface serialization'
in an entirely different way. In the first place, it is the expression, and not
the semantics behind it, which has a demonstrable and concrete permanence--these
phrases survive for centuries as the fundamental material of the poet's craft.
This means that in the practice of such poetry it is from these formulaic
phrases that the semantics are developed--not the other way around--and if new
meaning is to be expressed it must be done through old syntactic forms. In fact,
because this syntax consists of forms it is itself abstracted from any
particular instance of rendition. Instead of a duality of semantics and syntax
we have a tripartite distinction of semantics, syntactic form, and instance.
In our world of XML markup, the first time that I have seen consciousness and
appreciation that this is a three part distinction was last week at XML2001,
where that realization seemed to be popping up all over. I was particularly
struck by Rick Jelliffe's felicitous expression "rhetorical structure",
describing the schematics of the instance as distinct from either the semantic
or syntactic schemata which it might express. Rhetoric is about nothing but
rendition, so there is no room to argue that the rhetorical structure might
consist of anything more than can be derived from the empirical observation of
> I dunno, this may be simply my personal clustering into things that I find
> overwhelming and confusing and things that I find simple and natural. But
> it's not really about one perspective being "better" than the other; I will
> very happily admit that there are PLENTY of good use cases for the "XML for
> smart serialization of RDBMS or objects" perspective. Nevertheless, a clear
> alternative seems to be taking shape, and I think the distinction, and the use
> cases in which one or the other is more appropriate, deserve exploration.
I'll do some clustering of my own: as I see the process (and I am not a
disinterested observer) it is a unique instance operating upon, or processed by,
a unique observer or recipient on a unique occasion which is the only nexus for
transaction or conveyance of semantics. This is far, far from a tightly-coupled
system, and we would hardly expect specific semantic intent of an author to be
transmitted efficiently and intact through such a mechanism. Yet the assumption
that the semantics travel untouched in their fully-realized objects is at the
heart of the OO philosophy. By contrast, perhaps we can begin to think of XML's
very different assumptions--grounded in the primacy of syntax--as embodying a
source-code rather than object-code view of the world. In that world there is no
guarantee that I as recipient of your code will compile it--and therefore
execute it--as you expect. Such differences between us will not prevent our
executing transactions, but those transactions will consist of one of us taking
'as is' what the other offers, with no promise of how it will then be used.
Because of the asymmetry of our purposes, the transaction is not reversible,
though it might be offset with an additional transaction. Most important, the
substance of such a transaction is not our agreement (which would be an abstract
and semantic substance), but the stuff itself which is passed between us (which
is concrete and syntactic).
Agreement on semantics opens the door (through explicit congruence of purpose,
if nothing else) for the original author to seek to maintain those semantics,
and therefore to control far downstream the use of his 'property'. In a 'source
code' view of XML as the means of transmission and propagation, such claims are
absurd as well as unrealizable. If at the moment of transaction the goal of
whoever controls the receiving end of the process is not to maintain the
semantics as the transmitter intends, there is no way that the transmitter can
enforce downstream control, except by refusing the transaction itself, which
hardly accomplishes the goal.