Well-said Eric :-)
I think you have hit several key points in your message below that I would
like to specifically highlight, since they are the "root of my
belligerence".
1) What makes things "catch on"?
2) What is the world going to look like in 2-5 years?
These are things that we simply cannot answer at the moment. We're
lacking the "killer app" that will take the Semantic Web by storm! In
fact, I suspect that we have not yet even *conceived* of it... It may be
(or maybe not?) that the Web browser was slightly more "obvious" for the
case of HTML... though it's easy to say these things in hindsight.
Certainly the >300,000% increase in Web traffic [1] that accompanied the
release of Mosaic is testament to it being "killer", and we have nothing
even close to that for the SW! I suspect, though (in fact, I'd be willing
to bet my career!) that the Browser is *not* going to be the Killer App
for the Semantic Web. As such, any arguments for semantic web
technologies that rely on being able to type a URL into a browser and see
something useful are simply blah blah to me. Frankly, if I have to use a
browser to navigate the semantic web, then I have already lost interest.
Now, having said that, we can look at the *amazing* work that Eric Jain
has done to support both the browsing-community as well as the agent
community. I have to tip my hat to him!! Yet when I read messages like
the one he wrote this morning, I truly pray that the kinds of problems he
describes ("prevents people from ending up with extension-less files after
doing a save-as, a big source of confusion, based on my observations",
"when I show such pages to our biologists, they still think it's some kind
of error page, with all the gobbledygook about 'commitment',
'representation' and 'URI'") simply don't happen in the 2-5 year
time-frame. If we're still interacting with the Semantic Web through a
browser rather than an agent in 2-5 years, if we're still displaying 303,
404, or ANY kind of error-page to our end-users, then we should all be
pretty red-faced... an agent shouldn't have the problem of File/Save_As
without an extension, since the content should have been explicitly
defined either in the HTTP headers, or preferably in the entity's RDF
metadata, and it should be handling errors with finesse and "browsing"
onwards without human intervention.
Frankly, I think at least a part of the problem is that "GET" is holding
us back. Having to be compliant with the Browser/LWP/wget or your
favorite HTTP retrieval tool is, I believe, preventing us from imagining
what the Killer App of the semantic web will look like, and then making
the brave step away from pure HTTP URIs in order to achieve that goal. I
wonder, sometimes, if TBL's statement "the Semantic Web is an extension of
the existing Web" is perhaps one of the more questionable statements he
has made in his career... I simply don't see it that way! The Semantic
Web is (or at least could be) a very different animal than the Web, and I
have the feeling that this different animal is going to need a new set of
protocols and paradigms that are not encumbered by a specification (HTTP),
and a world full of billions of legacy documents, that were never designed
to do what the Semantic Web should be capable of doing. It might just be
easier/cleaner to "start from scratch" when it come to building something
that is likely to have behaviours that we have barely begun to imagine...
I agree 100000% that the LSID spec needs to be altered, tweaked, polished,
and if nothing else, better documented :-) However, it does represent a
very different way of thinking about identification/resolution than HTTP
(no matter how much we twist the HTTP spec to suit our current
requirements) and was designed with the Semantic Web in mind; moreover, I
think that in the next 2-5 years the perceived "requirements" that are
tying us to the HTTP spec are going to become less significant and easier
to discard as we develop SW agents that can do cleverer things. LSID
isn't utopia... but I worry that HTTP is even less so! LSIDs also are't
domain-specific, but rather (as Phil Lord pointed-out) *problem* specific,
so we shouldn't be afraid of using them based on domain-specific arguments
(though, frankly, even TBL himself seems to agree that the life sciences
are the main community using the SW, so the phrase "domain specific" in
terms of a protocol that *we* designed is a bit of a red-herring... we are
*the* domain, and I assume that's why the HCLS group has been asked for
their recommendations...)
Anyway, just some thoughts... Thanks, Eric, for stepping in (again) as a
moderating voice :-)
M
[1] http://www.hixie.ch/commentary/web/history
On Sun, 26 Aug 2007 12:51:22 -0700, Eric Neumann <[EMAIL PROTECTED]>
wrote:
In an attempt to modulate the tone a bit, it's clear that with such a
large and complex group of people and communities, many who had not been
part of earlier OMG/I3C discussions are not aware of all the details of
what had been discussed, proposed, and recommended. Having been a
LSR-OMG chair many years ago, I know what it takes to put RFPs through
DTC, PTC, and AB mechanisms at OMG. A lot of careful technical
forethought and agreeing has to go in to it...
At the same time, many groups in biological data and identifier
discussions are still getting up to speed what is meant by web
uniqueness and resolution within the W3C world. It's always easier to
respond to messages than to review the massive amount of technical
papers on the subject (I think simple tech/usage summaries are often
lacking). But this seems to lead to a lot of earlier email discussions
coming up again and again, i.e., info equilibration. As well as the side
effect of evoking emotions when not intended...
My guess is all sides here can provide an 80-90% technical solution to
the main set of data issues raised. That is not the main point of our
discussions though. In going forwards we need to also think about
learning from past attempts (successes and partial successes), what
factors help things "catch on" more quickly and are easy to
implement/adopt, and where do data providers and consumers (including
the non-informatics people) want to be in 2-5 years? I think we will be
capturing most of these shortly, and I look forwards to lots of useable
contributions.
I am not weighing in on any specific side here, but do hope to see an
outcome that is acceptable by most people AND offers the largest
potential for success, i.e., improves the quality of science and
medicine at a global scale.
Remember, before the web took off in the mid-90's, many pointed to the
limitations of other hyptext systems to why a global network of
documents would never succeed... past does not imply the future!
Eric
-----Original Message-----
From: [EMAIL PROTECTED] on behalf of Mark Wilkinson
Sent: Sun 8/26/2007 2:46 PM
To: Hilmar Lapp; [EMAIL PROTECTED]
Cc: Miller, Michael D (Rosetta); Eric Jain; Ricardo Pereira;
public-semweb-lifesci; Sean Martin
Subject: Re: [Fwd: Re: identifier to use]
On Sun, 26 Aug 2007 08:40:26 -0700, Hilmar Lapp <[EMAIL PROTECTED]> wrote:
If cannot do it through OMG, maybe LSID should be moved out of OMG. No
matter what, there is one consensus that is LSID won't be supported as
is.
Consensus by whom? There are organizations that support it already, such
as TDWG, IPNI, uBio, to name a few.
I think "consensus" here means "me and the people who agree with me"
Mark
--
--
Mark Wilkinson
Assistant Professor, Dept. Medical Genetics
University of British Columbia
PI Bioinformatics
iCAPTURE Centre, St. Paul's Hospital
Tel: 604 682 2344 x62129
Fax: 604 806 9274
***CONFIDENTIALITY NOTICE***
This electronic message is intended only for the use of the addressee and
may contain information that is privileged and confidential. Any
dissemination, distribution or copying of this communication by
unauthorized individuals is strictly prohibited. If you have received this
communication in error, please notify the sender immediately by reply
e-mail and delete the original and all copies from your system.
