On Tue, November 13, 2012 15:34, Ed Barkmeyer wrote:
> Kingsley Idehen wrote:
>> ...
>> The problem always boils down to webby data object representation,
>> access, and relationship semantics that scales to the Web. Once in
>> place, we have a functional global data space that accommodates
>> intensional and extensional data interaction. Basically, we end up with
>> Data as a new kind of Electricity conducted via hyperlinks. (01)
I have a lot of problems with this analogy. To stretch it, we have DC
with a huge range of voltages (both positive and negative) and supporting
power supplies of unknown capacity and reliability, as well as AC with
additional varying parameters of frequency, and phase, as everyone
who puts "data" on the web finds their own sweet spot. However, the
dimensions of variation are far more than these. (02)
Until you know the context of specific data ("electricity") supply,
you combine two supply streams at your own risk. (03)
>> The above isn't specific to any format, as you know. It has everything
>> to do with the following, functioning at Web-scale: (04)
>> 1. Entity Relationship Model
>> 2. Entity Relationship Semantics
>> 3. Instances of Entity Relationship Graphs linked across a variety of
>> boundaries. (05)
>> The Web as it exists is evolving into what I've outlined above at
>> frenetic pace. (06)
> Yes. In general, however, we don't know what any of these relationships
> really is, because nearly none of them is formally defined in terms of
> any well-defined terms, and most of them are just links. Even the ones
> that are in RDF are primarily written in terms that are effectively
> "primitive", being linked to HTML articles that describe them in detail,
> and discuss the vagaries in definition and interpretation instead of
> defining them. What we are seeing is the Web version of a bunch of
> student papers referencing each other, and a few reliable major works
> that some of those papers are consistent with. A lot of the links are
> blissfully ignorant of what Pat Hayes calls the 'Horatio Principle' --
> the set of all X in my world and the set of all X in your world might be
> different enough to support different universal axioms. (07)
I totally agree. Unless you have a way to automatically specify a
context for each data source you access on the web, and use some
accepted mapping specifying the relationships among such contexts,
you will find fallacies in merging data from different sources. (08)
> Put another
> way, IF you had usable ontologies for all the stuff that is being linked
> together, the overall 'theory' would almost certainly be internally
> inconsistent. (09)
This would certainly be true if all the triples or graphs of triples
were absorbed into the same context. (010)
> So the question is whether this rapidly growing thing is
> creating knowledge or just further complicating the overflow of
> disorderly information. (011)
B, until contexts are handled in some reasonable way. (012)
> The Linked Web is going to be rather like
> telephone systems of 1910 -- essentially chaotic until the perception
> of the relationship between order and value overcomes the "first in
> your neighborhood" syndrome. (013)
Are you suggesting a government supported monopoly a la Ma Bell? 8)# (014)
Standards for defining context could come first -- and be easier to promote. (015)
>> Facebook has a Billion profiles and much more from its data space alone
>> (they serve up JSON or Turtle content). (016)
> And a great many utterly uninformed and uninformative postings, to say
> nothing of baseless attacks and deliberate frauds. The question is:
> what portion of that cacophony is signal? And how would you know? (017)
>> Then you add the Linked Open
>> Data cloud which is in excess of 55 Billion triples, and then close all
>> of this out with instance data for Schema.org and what I am claiming is
>> pretty obvious i.e., it's already happening as we reached the point of
>> critical mass pass a long time ago :-) (018)
> Oh, we have "critical mass", to be sure. The problem is to harness it
> to provide controlled useful energy instead of Hiroshima (and to protect
> the results from the next data tsunami, or the next "revolution" in web
> technology). (019)
Actually, an uncontrolled "critical mass" creates a meltdown or more limited
destruction of the system. A Hiroshima only is created from a critical mass
controlled in a different manner than for continuous energy production. (020)
I note that if critical infrastructure were to be based on support from a
data utility to which prosumers of any kind can contribute, the data
tsunamis that would result would be the result not only of "natural"
unpredictable events (data "earthquakes"), but of cyberwarfare --
which would be able to design tsunamis to breach the highest breakwaters.
[Arrrgh. I'm getting sick of this metaphor.] (021)
-- doug (022)
> -Ed
>
> --
> Edward J. Barkmeyer Email: edbark@xxxxxxxx
> National Institute of Standards & Technology
> Systems Integration Division, Engineering Laboratory
> 100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528
> Gaithersburg, MD 20899-8263 Cel: +1 240-672-5800
>
> "The opinions expressed above do not reflect consensus of NIST,
> and have not been reviewed by any Government authority." (023)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (024)
|