Well, stepping back a bit from the task at hand, when we looked back and
attempted to understand why such strong pre existing demand existed for the WWW
that led to wide adoption ("phases of e-commerce" '97 if memory serves), it was
in part ironically the need for individuality and expression of
My 2 cents: If something is important, it should have a name (a URI). See the
W3C WebArch sec 2.1 "Good practice: Identify with URIs":
http://www.w3.org/TR/webarch/#uri-benefits
David Booth, Ph.D.
HP Software
+1 617 629 8881 office | [EMAIL PROTECTED]
http://www.hp.com/go/software
> -O
> From: Alan Ruttenberg
> . . .
> 2) I think that URIs should function first as unique identifiers, and
> only if possible, as elements of user interface. . . . .
I basically agree with this, but I think it is possible to strike
balance, since humans *do* still need to look at these URIs sometime
On Apr 20, 2007, at 1:13 PM, Mark Montgomery wrote:
I can see the potential for harm in attempting to zoom in too far
on granularity in standardization efforts (understanding the
appeal) and would therefore vote for prudent equilibrium between
adaptability and fixed. A bit messier perhaps t
Appreciate all of your debates and discussions recently- it's helped me and
no doubt others think things through a bit. Suspect many of us are not
chiming in to keep noise level down with rare tidbit of value to add- for
example in my case we are still primarily on the drawing board in part un
Just as something else to think about - The web page you cite does,
kind of, name these classes, in that a table is give, with two labeled
axes. Each box is, in some way, realizes the class that I am
suggesting you create.
However, the nature of html make the display of this unobtrusive. I am
su
Yes, this is an issue. I actually see it as two separate issues:
1) Creating names that are stable
2) Creating names that are readable
1) For stable names, a uri based on the logical definition will do,
e.g.manchester syntax definition, or an md5 of that. This will be
unambiguous and determinis
We've not yet implemented the redirect, or are sure what the redirect should do.
One issue is what is the thing that
http://purl.org/commons/pmid/7451682 refers to.
Is it the pubmed record as it exists now? The class of pubmed records
over the years?
Also, there are different forms of the pubmed
Hi Alan et al,
Although are are pros and cons for named classes vs. unnamed classes, I
prefer the latter for ease of ontology maintenance. As Huajun pointed
out, it's usually not a trivial issues to come up with standard (long)
class names that will make both human and machine happy. If these
I'm uncertain about the status of comments on anonymous classes. At a
minimum, they present present a challenge for current display tools,
though I suppose that any solution does. For my display tools I was
considering filtering out these named classes in certain types of
displays. I was able to
We met similar problems when we tried to relate publications to neurons.
Check the research notes for CA3 pyramidal neuron:
http://senselab.med.yale.edu/senselab/NeuronDB/ndbEavSum.asp?id=259&mo=4&re=
All of the notes are about papers supporting the evidences of being
present or absent of a rec
ok.
Often overlooked, but no less illustrated by the demo, is the ability of
underscore people and non-underscore people to work together.
jb
Date: Thu, 19 Apr 2007 15:44:41 -0400
From: Alan Ruttenberg <[EMAIL PROTECTED]>
To: John Barkley <[EMAIL PROTECTED]>
Cc: Jonathan Rees <[EMAIL P
12 matches
Mail list logo
