[ba-ohs-talk] Fwd: Re: [xml-dev] XLinks
This is a couple of emails bashed into one to show the thread, best read
bottom up. Near the top, Multivalent comes into play. This discussion
seems appropriate to HyperScope.
At 02:00 PM 3/16/2002 -0500, Thomas Passin wrote:
>[W. Hugh Chatfield I.S.P>
> > Ted Nelson (although he thinks XML is a mistake
> > http://www.xml.com/pub/a/w3j/s3.nelson.html)
>In this (pre-xml) reference, Nelson advocates having the equivalent of
>markup contained in a separate document rather than inline with the main
>document. He says that this would constrain the structure of the base
>document less than inline markup would. It's interesting that the
>Multivalent approach does just that for annotations and notes:
>Of course, you still have to connect parts of the overlay to their
>corresponding parts in the base document, which Nelson imagines you would
>do by counting bytes - the Multivalent approach uses its own document model
>combined with what they call "Robust Locations" to accomplish this.
>Tom P (02)
>From: "Christopher R. Maden" <firstname.lastname@example.org>
>Subject: Re: [xml-dev] XLinks
>At 06:22 AM 3/15/02, Len Bullard wrote:
> >Maybe I am just getting foggy in my dotage, but do we
> >need Xlinks if we use relational dbs and if so, for what
> >other than perhaps a convenient transport representation?
>"If it ain't broke, don't fix it." If you've got a linkbase that works
>fine now, go with it - there is no need for you to use XLink except, as you
>say, as "a convenient transport representation."
>Of course, in many senses, XML itself is "just" a "convenient transport
>representation," with data coming from databases, generated from the state
>of some process, or otherwise not existing natively as pointy brackets.
>Of course, I think you knew this - but thanks for the leading question. (-:
At 11:04 AM 3/16/2002 -0500, Hugh Chatfield wrote:
>I have access to the WWW..
>I have an SQL database...
>I am aware of the need to multiply link resources on the WWW.. perhaps first
>described in Vannevar Bush's "As We May Think" article in Atlantic Monthly
>1945 - where he envisioned the creation of an external "named trail" that
>could be shared (although he saw it as linking microfilm frames - but same
>difference - it was 1945 after all)
>Yet I still can't build and share these links...
>Building would take a bit of programming but it seems to me that the sharing
>part insists on some level of abstraction completely divorced from the
>underlying implementation technologies.
>A document I create in "Electric Pencil" can't easily be shared - but if I
>could save it as XML, then it can more easily.
>A "Vannevar Trail" I create using some local code using a relational DB
>system can't be easily shared - but if I could save it as XLink, then it
>can more easily.
>So I think you are right... XLink may only be a convenient transport layer
>(in the same way XML may only be a convenient transport layer for content).
>Ted Nelson (although he thinks XML is a mistake
>http://www.xml.com/pub/a/w3j/s3.nelson.html) considered the visible display
>of links to be of some importance (see http://xanadu.com/cosmicbook/ ).
>Methinks he is right as well.
>The real problem seems to be what to do with these things - how to share
>them - how to process them - how to display them .. not what particular
>underlying technology could be used for implementation.
>W. Hugh Chatfield I.S.P.
>CyberSpace Industries 2000 Inc.
>XML Consulting & Training
>From: Bullard, Claude L (Len) [mailto:email@example.com]
>Sent: Friday, March 15, 2002 9:23 AM
>To: 'Leigh Dodds'; xml-dev
>Subject: [xml-dev] XLinks
>On reading Bob DuCharme's article on "XLinks, Who Cares?" (05)
>I am struck that he leaves out the argument that is to me
>a primary reason XLink isn't taking off. Many of us use
>relational systems on the server that handle metadata
>relationships in tables. Many of the tasks that an XLink
>database is good for can be easily handled with a table
>that contains URLs as database types.
>Maybe I am just getting foggy in my dotage, but do we
>need Xlinks if we use relational dbs and if so, for what
>other than perhaps a convenient transport representation?