David, (01)
Many very good people were working on the Semantic Web. For any
particular practical or theoretical subject, you can probably
find somebody in the SemWeb community who is knowledgeable about
that topic. But what the SemWeb lacks is somebody like Steve Jobs. (02)
I don't know Steve J. personally, so I can't claim any knowledge
about his personal habits. But the designs and products of his
companies (Apple, NextStep, and Apple) show: (03)
1. Good taste in recognizing something good. (04)
2. Practical experience plus the willingness to try
something at the cutting edge of what's possible. (05)
3. The expertise to see how #1 can be implemented by #2. (06)
4. A deep understanding of the user's problems and habits
and the imagination to see how a good design can help them
solve their problems in a way they find natural. (07)
Steve Jobs really understood the users' problems and how
innovative technology could help them. (08)
I won't go through the history, but compare those points
above to what the SemWeb has achieved: (09)
1. The initial response by anybody who looks at the
designs is somewhere between horror and vomit. (010)
2. Many SemWebbers have practical experience, but they
settled on the lowest common denominator: RDF, which
was an ugly notation that had no more functionality
than the Comma Separated Values from the 1950s. (011)
3. They chose the most conservative imaginable options
(triple stores from the 1960s and DLs from the 1970s). (012)
4. Their only vision was a laundry list of buzz words,
but nothing like an integrated solution with the
"WOW factor" that made Steve J's creations famous. (013)
Your examples show that nobody was thinking about the
users' problems and habits when they designed the SemWeb: (014)
DE> I have yet to have a substantive conversation with anyone
> there as to how all this SemWeb hand waving is going to
> address the data-owners problem of discovering & then
> unambiguously publishing the semantics of their data. (015)
Unlike Steve J, they don't know what the users want or need. (016)
DE> The data owner lives/works with their data on a daily
> basis & knows it. However that knowledge is — as we all
> know — tacit, not explicit. Dumping/extracting data out
> of systems is an adventure, since you never really know
> what you're going to get. (017)
The people who designed the SemWeb tools were technologists
who had no idea what problems the users faced. Their big
reason for XML was to provide a universal parser. But users
weren't asking for a universal parser. Even the programmers
didn't ask for it. They'd be much happier with CSV, which
only requires a loop that looks for commas. (018)
DE> My wife wanted to look up some information in town records.
> When she queried 103 Webster Avenue, there was nothing.
> Likewise 103 Webster Ave. nothing. But finally:
> 103 Webster Ave = victory. (019)
That's a good example. The SemWeb solution to ambiguity is to
provide a unique identifier for the term 'Avenue'. But they
don't address the question of what anyone would do with that
identifier or the definition it points to. (020)
Noah Webster published his dictionary in 1828, but native
English speakers rarely learn definitions from a dictionary.
Providing a perfect definition on the WWW won't make anyone
use terms consistently. Any software that requires a perfect
definition is guaranteed to fail. (021)
DE> I do not see the Semantic Web (now morphing to "Linked Data")
> addressing the basic issue of "what contents is in that package"? (022)
I agree. That linked data doesn't use any kind of definitions,
perfect or garbage. It has no semantics at all. (023)
Before throwing more solutions at people, we should try to
understand their problems. (024)
John (025)
_________________________________________________________________
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
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (026)
|