"The Maturing Semantic Web: Lessons in Web-Scale Knowledge Representation" (Mark Greaves)
John Erickson
I know some of you have seen this due to my Twit about it earlier, but
for the rest of you: this is a great presentation by Mark Greaves, of
Vulcan Ventures (Mark Allen's VC et.al. company).
"The Maturing Semantic Web: Lessons in Web-Scale Knowledge Representation"
Mark Greaves
http://bit.ly/4NxgZM (pdf)
Mark's presentation gives a deeper, richer perspective on how semantic
technologies and the greater ecosystem have evolved. What it shows is
there isn't one single answer to Kingsley's "How come?" twit (such as,
it was hijacked by the AI people, etc) but rather that there are many
factors that have converged.
A very thoughtful, business-oriented perspective. Have a look and a think... ;)
John
John S. Erickson, Ph.D.
olyer...@gmail.com
http://bitwacker.wordpress.com
@ olyerickson
Kingsley Idehen
Enjoy old Jobs demos re. what we call Linked Data today (without the
ubiquity of HTTP and elegance of REST):
1. http://www.youtube.com/watch?gl=GB&hl=en-GB&v=odqojmB6C_Y -- 3/5
2. http://www.youtube.com/watch?v=j7WpcRReDlo -- 4/5
--
Regards,
Kingsley Idehen Weblog: http://www.openlinksw.com/blog/~kidehen
President & CEO
OpenLink Software Web: http://www.openlinksw.com
Brian Gryth
The url provided to the Greaves presentation seems to be broken.
Thanks, Brian
> olyerick...@gmail.comhttp://bitwacker.wordpress.com
> @ olyerickson
John Erickson
Thanks for checking! ;)
John
John Erickson
As one who did a start-up ca. 1995 based on distributing metadata via
"networked COM" (soon called "ActiveX...") , I can attest to how
radical these ideas were --- and how far ahead of *everyone* else NeXT
was at that time! It would have been interesting to have been able to
ride their coattails at that time...
It is informative to look at how Jobs packaged himself at that time:
he was totally focused on making NeXT's technology relevant to
enterprise customers. He's making the web-in-the-enterprise case very,
very gently, talking about how browsers will be like "free 3270
terminals" and referring over and over again to the "HP servers" on
the back end (emphasizing that NeXT had transitioned to being about
server software, not cool workstations). Even his attire is all about
"enterprise," playing the part of Enterprise/Corporate Steve...
Regarding the "ubiquity of HTTP and elegance of REST," the world
wasn't ready for anything but HTML-based UIs over HTTP at that time.
Any thing "cool" with data that you wanted to view in the browser had
to be server-side. You could do very cool things with ActiveX (which
we did, augmenting multimedia presentations with metadata pulled from
remote servers), but getting the client code to the user was a huge
issue (investors wanted you to get onto Microsoft's "golden master"
before talking money). And, as you say, ActiveX itself didn't talk
HTTP, most people assumed you had to do TCP/IP level operations to do
anything "interesting."
I should note that around 1997-1998 people were starting to get the
idea of URIs as service enpoints. For example, in the Handle
System/DOI community people were starting to see the value of having a
"halo" of metadata related to deployed objects; for example, at
Frankfurt in 1998 I demoed an approach where the client used a
standard URI syntax to index into different elements of an object's
Handle record and pull out URLs for getting bibliographic data,
copyright data, specific purchase endpoint, abstract, etc. Around this
time Tony Hammond (now with Nature Publishing) first used RDF/XML to
create a "manifest" of such compound objects and persisted the RDF as
the default response to a HS resolution, which was 'way ahead of it's
time. The purpose of this was to enable various calling clients or
other services to perform reflection on the object, to make decisions
based on what resources and services were available.
The W3C-standard content negotiation approach (esp. using 30x
responses) that the Linked Data community has adopted seems to make
more sense and is definitely more streamlined, but this was not wildly
socialized 10+ years ago... ;)
John
Kingsley Idehen
I am sure you can also see that failure to place Linked Data in an
appropriate location in the Distributed Object Technology genealogy tree
has cost it easy messaging and corporate communications wins re.
enterprise decision makers and developers alike.
Yes, I also note the Steve 2.0 (NeXT) and the transition to Steve 3.0
(on his return to Apple).
We missed binding UDBC (Universal Database Connectivity the precursor to
iODBC which delivered ODBC to Unix and Linux) to Enterprise Object
Frameworks (EOF) by a whisker i.e. Apple acquired NeXT before we could
finalize a deal with NeXT. Sadly, post Apple acquisition EOF commenced
its gradual slide into the forgotten world.
Anyway, RDBMS to RDF middleware is going to be critical to an serious
enterprise exploitation of Linked Data, and basically, this why we made
RDF Views an integral part of Virtuoso (i.e. not leaving something so
vital to efforts like EOF ever again). We are also considering making a
Virtuoso custom Atomic store for Core Data (what's left of EOF re. Mac
OS X).
More than anything else, notice how Steve focuses on:
1. Productivity
2. Collaboration
3. Data Access.
Now he doesn't really hone into Integration and Context Halos, but I
think he did really good re. the items above bearing in mind the point
in time etc..
Links:
1. http://en.wikipedia.org/wiki/Enterprise_Objects_Framework
2. http://www.splicd.com/j02b8Fuz73A/1522/2108 -- Steve Demonstrating
EOF based DBKit UI (short excerpt variant)
3. http://www.youtube.com/watch?v=j02b8Fuz73A -- Full Demo
Kingsley
Peter Neubauer
interesting read!
Cheers,
/peter neubauer
COO and Sales, Neo Technology
GTalk: neubauer.peter
Skype peter.neubauer
Phone +46 704 106975
LinkedIn http://www.linkedin.com/in/neubauer
Twitter http://twitter.com/peterneubauer
http://www.neo4j.org - Relationships count.
http://gremlin.tinkerpop.com - PageRank in 2 lines of code.
http://www.linkedprocess.org - Computing at LinkedData scale.
Tom Morris
> Enjoy old Jobs demos re. what we call Linked Data today (without the
> ubiquity of HTTP and elegance of REST):
>
> 1. http://www.youtube.com/watch?gl=GB&hl=en-GB&v=odqojmB6C_Y -- 3/5
> 2. http://www.youtube.com/watch?v=j7WpcRReDlo -- 4/5
Is this 1995 demo intended to be an example to be emulated, an object
lesson in past failure to be avoided, or just some harmless end of the
decade reminiscing?
The 1968 Englebart demo is available on YouTube too, if you really
want to reminisce about how long some of these concepts have been
kicking around. :-)
Tom
Tom Morris
> I know some of you have seen this due to my Twit about it earlier, but
> for the rest of you: this is a great presentation by Mark Greaves, of
> Vulcan Ventures (Mark Allen's VC et.al. company).
>
> "The Maturing Semantic Web: Lessons in Web-Scale Knowledge Representation"
> Mark Greaves
> http://bit.ly/4NxgZM (pdf)
Thanks for the link. For anyone else who's curious, this was
apparently an invited presentation at the 17th International
Conference on Conceptual Structures, ICCS 2009 in Moscow, July 26-31,
2009. http://iccs09.hse.ru/
Tom
Kingsley Idehen
> On Tue, Dec 29, 2009 at 5:29 PM, Kingsley Idehen <kid...@openlinksw.com> wrote:
>
>
>> Enjoy old Jobs demos re. what we call Linked Data today (without the
>> ubiquity of HTTP and elegance of REST):
>>
>> 1. http://www.youtube.com/watch?gl=GB&hl=en-GB&v=odqojmB6C_Y -- 3/5
>> 2. http://www.youtube.com/watch?v=j7WpcRReDlo -- 4/5
>>
>
> Is this 1995 demo intended to be an example to be emulated, an object
> lesson in past failure to be avoided, or just some harmless end of the
> decade reminiscing?
>
>
the business of linked data. In short, we have more elegant and open
infrastructure in the form of RESTful interaction with Portable Data
Objects (resources in today's Web parlance) via HTTP.
> The 1968 Englebart demo is available on YouTube too, if you really
> want to reminisce about how long some of these concepts have been
> kicking around. :-)
>
Reminiscing is good, esp. when it enables to us to engage the best
teacher of all i.e. History :-)
Kingsley
> Tom