[MM] Comments on Image Annotation Note (followup)

In response to my Boeing colleague's comments about interoperability
with other vocabularies for image markup, I sent this comment:

_____________________________________________ 
From: 	Uschold, Michael F  
Sent:	Tuesday, February 07, 2006 3:16 PM
To:	<Boeing Digital Asset Management Application Architect>
Subject:	RE: Semantic Image Annotation

Note that there is a separate note in the MM task force, specifically
addressing interoperability.
Semantic Web Image Annotation Interoperability
<http://www.w3.org/2001/sw/BestPractices/MM/interop.html> . 


There is a separate task force concerned with vocabularies: 

See: http://www.w3.org/2001/sw/BestPractices/VM/ for the overview.

Here are two documents about vocabularies.

*	Configuring Apache HTTP Server for RDFS/OWL Ontologies Cookbook
<http://www.w3.org/2001/sw/BestPractices/VM/http-examples/> , Editor's
Draft (latest version), Alistair Miles, editor. 
*	Managing a Vocabulary for the Semantic Web - Best Practice
<http://esw.w3.org/topic/VocabManagementNote> , Wiki draft (work
suspended to work on the cookbook
<http://www.w3.org/2001/sw/BestPractices/VM/http-examples/> ).

Mike
========

This was his response:

> ______________________________________________ 
> From: 	<Boeing Digital Asset Management Application Architect>
> Sent:	Thursday, February 09, 2006 4:17 PM
> To:	Uschold, Michael F
> Subject:	RE: Semantic Image Annotation
> 
> Mike -
> 
> I've taken a quick scan through all five links and
> they are either too incomplete or too detailed for
> me to find usable.
> 
> By "too detailed", I refer to the fact that they
> describe preferred implementation information
> out of context with the problems they are trying
> to solve. For example, "the cookbook" link is
> the best of all of these because it is the most
> complete but presumes knowledge of OWL and
> a controlled vocabulary built in something akin
> to XML (if not XML). I also get the sense that
> "the cookbook" would only be useful for a data
> migration or port to a semantic web environment
> so its utility appears to be one time only.
> 
> This leaves me scratching my head about how
> to use the information without getting down to
> the code level. I used to do that for a living but
> now spend my time on application architecture
> issues (meaning conceptual design, generating
> criteria for satisfying that design and picking
> components using that criteria to solve aspects
> of the design).
> 
> The closest a lot of this info comes to helping
> me is sneaking up on technical architecture
> standards for doing metamodel/metadata integration
> between a metadata repository and search services
> but there is too much physical design info and too
> little logical or conceptual design info for me to figure
> out how to link any possible problem to this potential
> solution.
> 
> Does this make sense? If not I'll try to invent some
> examples but that is kind of tough given my lack of
> detailed knowledge about semantic technology in
> general.
> 
<Boeing Digital Asset Management Application Architect>

============

I sent a further comment:

_____________________________________________ 
From: 	Uschold, Michael F  
Sent:	Thursday, February 09, 2006 4:42 PM
To:	<Boeing Digital Asset Management Application Architect>
Subject:	RE: Semantic Image Annotation

It does make sense, at least in broad terms. I will send this along as
further feedback. They will be able to tease apart more useful
information than I, which should result in updated version that may be
more useful to you.

Part of the intended audience is application developers, of course. But
you are also part of the intended audience, so they may have more work
to do to satisfy the needs of folk in positions like yours.

Mike

===

To which he responded:

Mike -

In regards to providing architecture info as
well as developer info, in theory all developers
are directed by architects, at least where
medium to large system development is
concerned.

So, providing dev info without architect info
is like writing a maintenance manual without
including an technical sales literature. If people
like me cannot figure out when to use a technology
then we won't specify it when doing COTS acquisitions,
working with current software suppliers to alter the
future direction of their products or even think to
design semantic technology standard support into
our own products.

<Boeing Digital Asset Management Application Architect>

Received on Friday, 10 February 2006 01:06:53 UTC