|
(comments
embedded below)
-John Bottoms
FirstStar Systems
Concord, MA USA
On 7/24/2013 7:32 AM, Kingsley Idehen wrote:
On 7/23/13 8:45 PM, John Bottoms
wrote:
Kingsley ,
I have quite a few tools in my work shop. When I start to make
something I get the tools that I need for that work. Size
matters when it comes to sawing. It helps if I had made the
object before, or something similar.
There are quite a few tools out there now that can be used for
ontological work. We wouldn't want to duplicate effort. But
clearly, as your link points out, the tool needed for a task
is entailed from the goal.
There are goals of the SW that have been discussed and some
tools exist to help accomplish some of those goals. It seems
that there are a number of subgoals, some of which have been
met and some of which have not been described succinctly. The
scope matters also. Which did you have in mind?
Tools include:
1. middleware for generating fine-grained structured data (endowed
with machine-readable entity relationship semantics) from
coarse-grained structured data
Granted, Big Data advocates are exploring the
developments for this issue. I believe IBM's Stream computer also
addresses this issue. Due to the need for adding meanings, the
most viable solutions I have seen include some sort of
human-in-the-loop approach. Short of that, tools can be developed
to collect and extract suggested meanings from Big Data, given
that you have sufficient data. Big Data itself raises a number of
questions. And truly Big Data projects are often expensive as they
typically require math, programming and Big Project management
expertise.
2. HTTP URIs are also a tool for entity denotation (naming) that
can optionally include resolution document content - this enables
the construction of web-like (or webby) structured data delivered
as document content
I believe URI's are better labeled as
encodings, correct me if I am in error. Tools may employ or attach
encodings, but the encoding itself does no real work without
additional functions or methods. If naming functions are added
then we are looking at a type of speech-act/language game that
DaveEddy has been grappling with for some time. And, with the
semantics behind the database, we can only hope to share encodings
for open meaning data, not just open linked data.
There may be silo solutions...IF the politics can be resolved.
But, users have typically resisted mandates from on high. Narrowly
focused silo's may be the best avenue for resolution. The
stakeholders for a community of interest cannot move forward on
semantics without solving this dilemmna.
The BioPortal approach seems to be the only viable approach for
sharing and resolving data and meta-data differences. Perhaps we
should reach out to groups that have started "automated reasoning"
systems designs. These include: Towntology (EU project for Urban
Planning), ForMaRE (reasoning tools for economics), OpenMath and
API4KB efforts.
These groups are moving ahead in reasoning systems, and therefore
also need tools. They are also in need of assistance in designing
ontologies. Perhaps we should suggest combining efforts on tools
as well as on data.
3. SPARQL query editors -- like their SQL counterparts, these
enable creation of globally accessible queries while also enabling
the sharing of query results and their definitions
I strongly agree on this with an additional
observation. SPARQL is at base a language. If you propose an
editor for SPARQL then it sounds as if you are proposing a tool
that converts something more user friendly such as a CNL or
simplified NL into a valid SPARQL query. Is this correct?
The focus of this issue is the need to move
from a syntax based construct to a semantic construct. Perhaps we
should add "extensibility" to Tim's "diversity, heterogeneity, and
interoperability".
4. Browser plugins and extensions -- which enhance existing
browsers such that they can be used to create, save, edit, and
share web-like structured data (endowed with machine-readable
entity relationship semantics).
JohnS:
Yes, I agree those three words, "diversity, heterogeneity, and
interoperability" at key but they strike me as features or
facets, not goals. I believe we need to state a goal for a
tool that has not already been addressed.
1-4 enable non-disruptive exploitation of the technologies that
are generally referred to as "the semantic web" stack.
I agree with John about the need to embrace and extend existing
solutions using the semantic web technology stack. Unfortunately,
this approach isn't the norm and its a large reason why we still
have a lot of confusion-driven-inertia swirling around the
"semantic web" meme and the technologies that comprise its stack.
To answer John's question precisely, major Web players such as
Amazon haven't embraced the stack (in overt ways) because their
developers either find the technology confusing or they find it
too disruptive to implement bearing in mind existing legacy
infrastructure. That said, the likes of Facebook, Google, and
Microsoft are increasing their use of these technologies. The same
applies to the U.S., UK, and many other countries.
In my eyes, the biggest irony around "the semantic web" is the
fact that at its core lies a power collection middlerware -- based
on the architecture of the World Wide Web -- that's artificially
obscured by poor narratives and overly provincial marketing.
Kingsley
-John Bottoms
FirstStar Systems
Concord, MA USA
On 7/23/2013 7:54 PM, Kingsley Idehen wrote:
On 7/23/13 1:00 PM, John F Sowa wrote:
Amazon began life as a bookseller, and
they extended their reach to
become a very large retail supplier of almost everything.
But their
service business has grown faster than their retail
business:
http://www.reuters.com/article/2013/07/21/net-us-amazon-cloud-idUSBRE96K04B20130721
Some excerpts:
After years of being dismissed as a
supplier of online computer
services to startups and small businesses, Amazon Web
Services (AWS)
beat out International Business Machines this year to snag
a $600
million contract with the Central Intelligence Agency.
Public cloud computing, which AWS pioneered in 2006, lets
companies
rent computing power, storage and other services from data
centers
shared with other customers - typically cheaper and more
flexible
than maintaining their own.
Five companies vied for the contract - AWS, IBM,
Microsoft, AT&T and
another unidentified firm, according to a report on the
bidding by
the U.S. Government Accountability Office.
My only knowledge of AWS comes from reading some of their
documentation
and some miscellaneous articles about it. They provide some
flexible,
high-speed methods for indexing, finding, and updating
anything in
their clouds.
But I noticed that 2006, when AWS started, is also the year
when the
DAML project finished its basic tools: RDF, OWL, and
SPARQL. Amazon
does not use any of those tools. But I noticed that some
people have
stored data that contains RDF links in AWS.
I also noticed that one of the Amazon tools, SimpleDB, is
implemented
in Erlang. That language was designed to support concurrent
processing
with multiple threads, especially for use by large telecoms.
AWS probably uses Erlang (or techniques inspired by Erlang)
for other
purposes, especially for their method of "autoscaling",
which is
"a feature that automatically adds or removes computing
power in
response to application use." For a brief overview of
Erlang,
see http://www.erlang.org/faq/introduction.html
.
"Auto-scaling is very complex and
there are not many cloud providers
that can do it well, but Amazon is great at it," said Kyle
Hilgendorf,
a cloud computing analyst at Gartner.
Erlang is an example of the kinds of tools that mainstream
developers
are willing to adopt and use for mission-critical
applications. One
more example: Facebook uses Erlang to support their chat
backend.
Why haven't developers found a way to build multi-billion
dollar
technology on top of the SW tools? They might provide some
support
for importing data from those tools, but they don't use them
as the
foundation for their technology. Why not?
John
John,
We really need to establish what 'Tool' means [1] to push this
discussion forward, coherently. Once the meaning of 'Tool' is
established we still have the thorny issue of what a Semantic
Web Tool is, bearing in mind the aforementioned buzz-phrase
is rife with confusion and controversy.
Personally, I believe the World Wide Web has always been a
Web of Semantically interlinked Data. The issue (in my eyes)
is that over time the fidelity and machine-readability of the
underlying entity relationship semantics are what continue to
evolve [2].
I am an extensive (an very early) user of AWS. I use this
platform to deploy very sophisticated solutions that leverage
various aspects of the Semantic Web technology stack [3]. AWS
itself will benefit immensely from Semantic Web technologies
once we find a way to reduce the confusion (and provincial
tendencies) swirling around this most important aspect of the
Web.
Today, when making AWS based EC2 AMIs you will notice that are
lacking on the data model front, and this makes automated
construction and management of AMI's more difficult than it
needs to be. Anyway, we are going to turn this data into
Linked Data and then present it back to the folks at Amazon
which could shed a lot of light on how these technology
provides immediate value to a thorny problem they are
grappling with etc..
Links:
[1] http://dbpedia.org/describe/?url="">
-- Description of a Tool
[2] http://bit.ly/10Y9FL1 --
Why I claim the World Wide Web was a Semantic Web
(coarse-grained fidelity, on the machine-readability front)
from inception (note: click on the links!)
[3] http://bit.ly/Y4aHx9 --
Amazon EC2 AMI for Virtuoso
[4] http://bit.ly/NzIm3t -- G+
note explaining AMI setup.
--
Regards,
Kingsley Idehen
Founder & CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca handle: @kidehen
Google+ Profile: https://plus.google.com/112399767740508618350/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen
|
_________________________________________________________________
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 (01)
|