On 2/4/14 9:45 AM, John F Sowa wrote:
> Paul, Rich, and Kingsley,
>
> I want to emphasize that the entire family of *ML languages
> (GML, SGML, HTML, XML...) have definitely proved their value
> for word processing, formatting, and annotating texts.
>
> But a notation that is good for one purpose might not be good
> for other purposes. (01)
Yes, I agree. (02)
>
> PT
>> can you point out some specific deficiencies in particular
>> specifications produced by these poor benighted folk?
> Excellent question.
>
> But I certainly would not call them poor or benighted. Most of the
> deficiencies were caused by putting ideology ahead of pragmatism. (03)
Yes. (04)
>
> RC
>> XML is here to stay, IMHO, because it fills the void left by lack
>> of a truly semantic vehicle for transmission of messages.
> The *ML family of languages has won the competition for formatting
> and annotating texts. But note that XHTML lost out to HTML 5. (05)
Yes. (06)
>
> The second line indicates the problem: we need a good method
> for transmitting semantics. Note how rapidly hash tags became
> a de facto standard for communications among humans. That's
> a clue that hash tags are *superior* to IRIs for communicating
> meaning in certain kinds of applications. (07)
Not necessarily, I recently wrote a post about how hashtags and @handles
are basically end-user oriented shorthand for using HTTP URIs to denote
entities [1]. (08)
>
> Also note how rapidly Schema.org was adopted. That's because
> the terms *and* the definitions are readable without a tutorial
> or short course: English phrases and texts without IRIs. (09)
Hmm.. that's more to do with Google putting its considerable might to
commendable use. (010)
>
> We certainly need unique identifiers for terms that really do
> represent unique things. (011)
And if they are terms, they have to include references to description
documents (or descriptors) i.e., term duality characteristic must be
preserved. Basically, what you get on a platter via hash based HTTP URIs
when following Linked Data principles [2]. (012)
> Two centuries ago, Berzelius invented
> the notation NaCl for salt, and it has been enormously successful
> for inorganic chemistry. But things get bad in organic chemistry.
>
> JSON-LD is a "lightweight" attempt to mix IRIs with readable labels. (013)
Yes, and in the process Syntax obscures Semantics. Just like RDF/XML. (014)
> Simple examples aren't bad (http://json-ld.org ). It's a step in the
> right direction. But when you push it into more complex examples,
> it gets complex. C'est la vie. (015)
See my comment above :-) (016)
>
> RC
>> IMHO, English (or other natural language) is the best vehicle for
>> semantic transmission, but we have yet to perfect the mechanics of it.
> KI
>> Yes!
> I agree. And we need a pragmatic approach that promotes experiments
> with various implementations *before* edicting standards. (017)
Yes! There shouldn't be any edicts, we should simply put puzzle pieces
together i.e., let everything be loosely coupled so that evolution is
organic. (018)
>
> John
> (019)
Links: (020)
[1] http://selnd.com/1eHyRex -- Effects of Linked Open Data on the
future of SEO (covers @handles and hashtags) (021)
[2] http://bit.ly/WAJGCp -- Linked Data in a single slide . (022)
-- (023)
Regards, (024)
Kingsley Idehen
Founder & CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter Profile: https://twitter.com/kidehen
Google+ Profile: https://plus.google.com/+KingsleyIdehen/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen (025)
smime.p7s
Description: S/MIME Cryptographic Signature
_________________________________________________________________
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)
|