Web Ontology From 6.075 Feet: Modules And Metalanguages
Jeff Rubard
*Content* As A Guide For Gentle Readerships
Although the XML "hype" has not yet abated, those involved with
certain irrealities of the informatics present are invited to
consider another "angle" to new standards for Internet delivery
of meaningful. Growing pressure for the standardization of web
content *through* XHTML displays a tension at the heart of
CERN/W3's strategy for making markup both more robust and more
computationally tractable; by contrast this XHTML writer feels
his "tagging" to be very structural indeed, and the "Semantic Web"
as currently promised to be something less than savvy as regards
the formal-semantic prospects for "semantic" issues. The Rich
Description Framework provides you with a way to trick out all
sorts of data with "properties" that can then be relationally
"combined" in a style familiar to all -- but I invite the stern
of heart to consider another, more sensible, methodology for
*semantic* content such as this has traditionally been conceived.
The XHTML 1.1 standard, itself already of some longevity, is
based upon the recommendation that XHTML be "modularized" such
that the language permit of both content simple enough to be
delivered through sub-PC systems (XHTML Basic) and content
complex enough to require traditional HTML "keying". But for
those of us without a taste for proving Wirth to have been an
object-oriented programmer, I will point out that the 1.1
recommendation as it stands really contains a much more
interesting formal structure than the W3 recommendation suggests,
one which can allow "web designers" to develop "immanent"
or intrinsic models of the content they are preparing for delivery.
Why do this? Traditionally this is done by news outlets, for the
purpose of *typing* elements of the document (i.e., "workflow" as
a heuristic within an organization and a tangible without).
How could this be done with the reduced set of tags allowed by
XHTML? The <meta> tag, traditionally used for information not
intrinsically linked to the content of the webpage, actually
permits of systematic employment for the more-or-less Tarskian
"semantic definition of truth" with respect to body tags as a
"special theory" for the content delivered through a page: i.e.,
*predicates* rather than operators (considered in this light,
"x is <p>" tells you something informative about the "object" rather
than "systematically" altering its character; you already wrote the
paragraph, and it was paragraph enow for you, and HTTP tells no tales
unless "put on" to the trail -- real work is being done here by the
markup, and it is neither glossing nor hashing). What would this
involve? Simply, a set of meta-tags involving "homonymic"
explanations of the
"structural-descriptive" body-markup; including, well, as many
"regularities" of usage as your writing style permits of -- a
"metalanguage" *stricto sensu* for the page content.
What profiteth an organization to do this? Roughly, datamation
hygiene -- to transpose Tarski's "Criterium W" into a different key,
the content of a Web document is no more effectively manipulable than
it is, and if experiments in "face scanning" have gotten us all ready
for the thought that SQL-style manipulation of data has become more
expert in the absence of the "grammar of assent" then there's all
sorts
of lessons to be learned at almost-negligble cost. What can such a
methodology (involving no dodgy commitments to entities once banned
in Boston) not accomplish? Roughly, the "modernization" or
clarification
of a document's role in workflow: a task which modular extensions of
XHTML 1.1 could accomplish *almost* as handily as RDF or OWL: but here
we are again at play in the Mostowskian fields of generalized
structure,
and perhaps there is no really serious question beyond the ability
of a piece of information to travel from point X to point Y within
a decent interval (i.e., if Web structure becomes more unwieldy we
may begin to wonder why we weren't just "stealing a way" all this
time).
Jeff Rubard
> Get What You Want And Want What You Have?: Intrinsic Models Of
> *Content* As A Guide For Gentle Readerships
>
> Although the XML "hype" has not yet abated, those involved with
> certain irrealities of the informatics present are invited to
> consider another "angle" to new standards for Internet delivery
> of meaningful. Growing pressure for the standardization of web
And for those of you waiting around for me to define "content" in
some *other* inimitable style, here's an example of the metadata
arrangement I spoke of earlier: an "article" of mine outfitted with a
"Jukie scheme", that is to say a Tarskian theory of truth or "semantic
frame" for XHTML tags *and* special characters. The Jukie scheme
(and there's some history here) *not only* features no XHTML, there is
no "Real Soon Now": the page as written either "corresponds" to the
intended designation ("Oregon Electric: Back Again, Who Are They?" is a
an op-ed) by virtue of semantic/structural features indicated in the
"structural-descriptive" (tag^input) *homonym* for a particular semantic
move (i.e., quotation) or does not (i.e., the meaningfulness of the
content is, shall we say, unclear -- there is not a tight enough fit
between the markup choices being made and the uses to which such a
page could be put). What this *doesn't* do is teach you what you
don't know how: it flags "metacommunicative" or layout features which
are not really amenable to data mining *stricto sensu*, such that there
is little point in having such information stored as a profile. But,
y'know, considering the character of such an innovation I'd be curious
to hear opinions other than "not gonna work".
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN"
"http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" >
<head>
<title>Oregon Electric: Back Again, Who Are They?</title>
<author>Jeff Rubard</author>
<meta scheme="Jukie:Op-Ed" content="x is a paragraph" name="p^x"></meta>
<meta scheme="Jukie:Op-Ed" content="x is bold" name="strong^x"></meta>
<meta scheme="Jukie:Op-Ed" content="x is an article title"
name="h3^x"></meta>
<meta scheme="Jukie:Op-Ed" content="x is a quote" name="q^x"></meta>
<meta scheme="Jukie:Op-Ed" content="x is emphasized" name="em^x"></meta>
<meta scheme="Jukie:Op-Ed" content="x and y" name="ampersand^xy"></meta>
</head>
<body>
<h3>Oregon Electric: Back Again, Who Are They?</h3>
<p>
The paper today announces that Enron intends to sell Portland
General Electric to a new utility company backed by Texas investors,
Oregon Electric, after a bidding period openly designed to defuse
objections. It is my expectation that this plan (backed by former
governor Neil Goldschmidt, among others) will put an end to talk of
making PGE a public utility along the lines of those which serve
smaller Northwest towns, but it is my feeling that the hoopla which
is clearly coming will be designed to put an end to talk of a
different kind <strong>before it starts</strong>. Namely, discussion of the
fact that the Oregon Electric name is well-known to longtime
residents as that of the company which operated various rail lines
from Portland to many places the northwestern quadrant of Oregon, and
offered regular interurban passenger service on those lines for
roughly twenty-five years (1908 to 1931).
</p>
<p>
Perhaps those with a newfound nostalgia for the days when Portland
had a widespread streetcar system can be expected to perk up here,
but Oregon Electric did not operate Columbia Crest and the myriad
other lines (which were going concerns for a somewhat longer period):
that was the Portland Traction Company, like Oregon Electric a
private concern. What Oregon Electric did was <em>build the metro
area</em>, in a fairly literal sense: many area cities were farm towns
with no direct connection to Portland before the interurban provided
them one, and Garden Home and Beaverton were nothing at all before
the railway built stops where their civic cores are today. Was this
a <q>disinterested service</q> to the area? Something like,
because where there is nothing there is nothing to carry; the Oregon
Electric was owned for many years by Jim Hill, operator of the Great
Northern Railway (today part of Burlington Northern & Santa Fe), and
this is not quite an irrelevancy, if you know what I mean. The Great
Northern employed a similar strategy all across the upper Midwest and
Rocky Mountain states, settling immigrants along the line; and they
are the reason Seattle, rather than Vancouver became the dominant
city in the Pacific Northwest.
</p>
<p>
So the Oregon Electric was a major force in shaping the character
of Northwest Oregon, but it is important (given the context) to
realize that it was really in no sense a <strong>predatory</strong>
corporation (this was no longer the era of robber barons, if you know
what I mean). And proof of this is that there was a competing interurban
line operated by the Southern Pacific, the Red Electrics (which ran
from Portland to points south along what are the routes ot the
present-day 99W and 99E). And faced with the prospect of this great
new locally-based company which provides most of the metro area with
its only source of electricity, I think we would be justified in
asking: <q>In what sense does this company resemble, or not
resemble, the company whose name it is taking over?</q> This is
not an otiose question, as during the Enron years PGE, and not only
the Enron affiliate, was frankly predatory: I, and perhaps others,
will remember them as the company that ate Civic Stadium (and to me,
that seems likely to have been <q>a matter of principle</q>).
And for all we know, we could be expected by OE to "feel the
tiger" on a regular basis; and frankly, I for one am long past
feeling a need to buck up corporations and other institutions as a
<q>matter of course</q>.
</p>
</body>
</html>
Karl Smith
Fortunately, those nice experts at MIT invented the Universal
Translator sometime in the early 1970s. Prior to the invention of the
Global Information Initiative. So I can just prepend:
<?UniversalTranslator
input="xx-buzzword"
output="en" ?>
to your message, reload it into my web browser, and... That's odd,
nothing happened. Perhaps the General Problem Solver can tell me what
went wrong?
XHTML is a fraud. RDF (it's "resource" not "rich" description format,
BTW) is probably a fraud too.
--
And there's no such colour as "LavenderBlush", either.
Philipp Lenssen
> Get What You Want And Want What You Have?: Intrinsic Models Of
> Content As A Guide For Gentle Readerships
>
Could you give a small example?
E.g. a working (or someday working) XHTML snippet.
--
Google Blogoscoped
http://blog.outer-court.com
Jeff Rubard
> opense...@graffiti.net (Jeff Rubard) wrote in message news:<49b5d62a.04031...@posting.google.com>...
> > Get What You Want And Want What You Have?: Intrinsic Models Of
> > *Content* As A Guide For Gentle Readerships
> >
> Translator sometime in the early 1970s. Prior to the invention of the
> Global Information Initiative. So I can just prepend:
They're nice experts, all right. And the fundamental message of this
potted categorial grammar is that W3 standards incorporate more
corporate buzz than logical form, such that a very selective hash
would have to be made of them to have fundamentally sound standards
for current informatic needs: but "messages" aside, there's no reason
to think that "simple recursion" (truth-functionality) could not serve
the useful function of notionally detaching "communicative intention"
with respect to a piece of content from its conditions of circulation
and/or analysis. In other words, sometimes an exploding cigar is just
an exploding cigar -- but "flattened aspects" are a reality as well.
> <?UniversalTranslator
> input="xx-buzzword"
> output="en" ?>
>
> to your message, reload it into my web browser, and... That's odd,
> nothing happened. Perhaps the General Problem Solver can tell me what
> went wrong?
I dunno, does Herb Simon dream of electric sheep for you? Perhaps in
a way he does: why would you think that the GPS and suchlike would be
any good at language-design considerations involving questions of
"order" not amenable to adjudication by automated theorem provers? No
reason at all, really -- but much of the technical lore you and I
know is in similar condition, as computer science to date has been
resolutely Promethean in ambition (more about which...).
> XHTML is a fraud. RDF (it's "resource" not "rich" description format,
> BTW) is probably a fraud too.
A fraud relative to what? HTML? Gopher? Netnews? RDF/OWL are
"frauds" relative to simple facts about first-order predicate logic:
so much so, in fact, that it should really be known from whence
"formal ontology" -- Church's lambda calculus, an (enlightening)
failure as foundation for mathematics and then an intensionality
toolkit. That is to say, much of the descriptive work W3 does with
its languages is willful: maybe you would need a functional Web
language, but it's more than a little SCA to define it in terms of
Aristotelian metaphysics (and perhaps a little "futurismic" to think
a more parsimonious markup language could not actually serve the role
of transposing layout issues into "intentional idiom").
Jeff Rubard
> Jeff Rubard wrote:
>
> > Get What You Want And Want What You Have?: Intrinsic Models Of
> > Content As A Guide For Gentle Readerships
> >
>
> Could you give a small example?
> E.g. a working (or someday working) XHTML snippet.
Well, no: "questions of detail" are all that this meta-tag methodology
permits, i.e. it's an exercise in pointedly makeshift semantics for
XHTML which does not permit invocation within the language itself (in
accordance with the standard scruples you may have learned from
Donald Davidson's work in college). What this metalanguage does
permit,
of course, is parsing by a scripting language -- and its "superiority"
to
XSL-FO as such is that I intend here to raise the issue of something
like the
*inverse* of that language, that is to say properly semantic markup
not
pertaining to physical layout rather than structural markup pertaining
to fine stylistic distinctions: the former would allow one to gauge
more
readily the extent to which the information contained within a
particular
Web document could be seriously incorporated into a search array, i.e.
we are talking about something like concretized "schemas" permitting a
rather great deal of fidelity to the authorial intention "associated"
with
the document. So far, so little: but if much of this information
seems
a little tired, it might have something to do with *at least* this
particular post not really conforming to any such strictures, which
(as you may not know) actually predate the (formalized) concepts
behind markup languages.
Jeff Rubard
> "Philipp Lenssen" <in...@outer-court.com> wrote in message news:<c3edod$264usl$1...@ID-203055.news.uni-berlin.de>...
> > Jeff Rubard wrote:
> >
> > > Get What You Want And Want What You Have?: Intrinsic Models Of
> > > Content As A Guide For Gentle Readerships
> > >
> >
> > Could you give a small example?
> > E.g. a working (or someday working) XHTML snippet.
>
> Well, no: "questions of detail" are all that this meta-tag methodology
> permits, i.e. it's an exercise in pointedly makeshift semantics for
> XHTML which does not permit invocation within the language itself (in
> accordance with the standard scruples you may have learned from
> Donald Davidson's work in college). What this metalanguage does
To get a sense of what I mean, consider the <def> tag. All credit
issues aside, the structural import of this tag as applied to a word
or phrase is that, by virtue of its "definitive" character, the
document as presented is not properly speaking "recursively
enumerable" -- that is to say, you're not *really* going to figure out
what that one part of the document means by considering how it "spins
out" in terms of evident textual connections (such as you could design
a recursive function to capture). It's a "rib", rather than a "shim",
and *encapsulates* the problem this one-shot content model is
intended to address (by itself being definitive or exemplary of
recursion a la the previous standard for formal semantics): thusly, it
should offset those elements in a document which are arguably
definitive.
