[
Lists Home |
Date Index |
Thread Index
]
These are coupled functions that swap sides as we drill down:
someProtocolMorphToKickOffAFunction://somethingOneHopesIsWhereItIsSupposedTo
BeWhenFunctionFires
^ ^
| |
The Computer Science Part The Social Behavior Part
^ ^
| |
"and this should be HTTP" "And this should be in a
recognizable format"
^ ^
| |
The Social Behavior Part The Computer Science Part
and that is life among the mammals and the silicon. Local development
and local persuasion are hard coupled.
The question is what are the advantages of sharing application
semantics for linking. Again, we are back to a shared code
framework just as one shares the Hypertext function call in VB.
The Hytimers would point to a notation associated with a linktype
that is, before groves. We've been around this topic a number
of times and it comes down to picking software. The abstractions
barely matter because their effectiveness still comes down to
what an implementation enables for handling finding and displaying
locations within and amongst content types. Regardless of having
a switch, a select statement, a dialog, etc., it is still just
application semantics for picking out of a list and associating
an event to that selection. It isn't a big enough deal to
warrant a standard so even if the commonality is there, no one
has sufficient incentive to converge on it.
That doesn't mean SVG is wrong. It means it has its own reasons
for using XLink and other applications may or may not share them.
One doesn't junk XLink. It simply can't be mandated.
len
-----Original Message-----
From: Simon St.Laurent [mailto:simonstl@simonstl.com]
clbullar@ingr.com (Bullard, Claude L (Len)) writes:
>Or admit that there is actually only ONE way to link on the WWW:
>
>someProtocolMorphToKickOffAFunction://
somethingOneHopesIsWhereItIsSupposedTo
>BeWhenFunctionFires
>
> ^ ^
> | |
>
> The Computer Science Part The Social Behavior Part
Sort of. I'm not thrilled with URIs as the one true identifier for the
Web either, though I think we're in roughly the same place on that. I
don't think identification and linking are the same, but I agree with
your next sentence thoroughly.
>Everything else is application semantics. It is
>the application semantics that don't mesh although one
>can make that happen by the same acts that influence
>norms of social behavior just as the link design is
>imposed top down.
I'm not sure that using norms is going to be effective in this case.
There seems to be a trend lately where "norm" is defined by
"organization and vendor preference", not "quality of work". That's
caused enormous pain on the schema side and created messes I expect
we'll still be cleaning in twenty years (when we finally have to replace
the schema-based systems no one wanted to touch). Fortunately, that
hasn't worked for XLink.
>There is no reason XLink won't work. It is a matter
>of persuasion. So far, no persuasion has been effective.
>As I said earlier, that is because there are easier ways
>that only depend on different scales of local control.
I think you have something here, though again I think you're talking
about means of social control rather than technology. You're suggesting
local persuasion; I'm suggesting local development. In the absence of
effective persuasion for XLink I expect the latter will happen anyway.
|
