RE: SIMILE demo storyboard feedback

Nick, thanks for taking the time to review and comment.

> First, I agree with MacKenzie that the Venn diagrams need 
> more work.  In particular, the first one is confusing. There 
> is no overlap between "images of artifacts designed by FLW" 
> and "courses about FLW".  Even if you use "images of 
> artifacts..." and "materials from courses..." there is no 
> real overlap, unless each community has the exact same image 
> cataloged.
> 
> I think instead what needs to be conveyed is that there are 
> two disjoint sets of information, those cataloged in VRA-Core 
> and those cataloged in IMS, but that the information need 
> spans both.  I've made a mockup of what I have in mind 
> http://www.ai.mit.edu/people/matsakis/simile-> venn.png

That is indeed what I was trying to convey through the image.  I think your diagram does a better job, and suggest that we adopt it.

"So the first requirement for our test corpus is that it must contain materials sourced and
described by different communities yet which intersect in a common domain of interest."

Also: the information need spans both disjoint sets of information, but the domain expertise and knowledge of the user is likley to be largely limited to one or the other.


> I 
> think the second Venn diagram is also a little 
> bit incorrect.  There are no elements that are both VRA-Core 
> elements and IMS elements.  Rather, there are elements that 
> are "expressible in VRA-Core" and "Expressible in IMS".  A 
> subtle distinction, but probably worth making. It is the 
> overlap between these two sets that is represented by the 
> vocabulary mappings.

Andy Seaborne has mentioned the importance of modelling the concepts that the elements represent.

So perhaps we have:

VRA Elements

    map

VRA Concepts

   (intersection and map)

IMS Concepts

    map

IMS Elements


I have observed debate among the team about whether mappings should be stated directly between the VRA and IMS elements, or through some set of concepts related to these elements.  Modelling "contributor" elements using concepts from vcard is an example of the latter approach.  What do you think?

> In the section on vocabulary mapping, I think it is perhaps 
> sweeping some things under the rug to say that in the free 
> text search the system will "know that 'Wright, Frank L.' is 
> equivalent to 'Frank Lloyd Wright'". Making this mapping can 
> be viewed as part of the demo.
>
> For the templated search, I'm 
> not sure I understand why the IMS records would not be 
> returned.  Is it that the lifecycle.contribute.author will be 
> the photographer, rather than FLW, or that there is not 
> lifecycle.contribute.author field in those records?

Mark, can you respond to the above two questions/comments?

- Mick

====
970.898.6788 office    240.536.0765 fax
617.899.3938 mobile    303.494.5202 residence
bass@alum.mit.edu      mick_bass@hp.com
====


> -----Original Message-----
> From: www-rdf-dspace-request@w3.org 
> [mailto:www-rdf-dspace-request@w3.org] On Behalf Of Nick Matsakis
> Sent: Wednesday, October 29, 2003 2:26 PM
> To: www-rdf-dspace@w3.org
> Subject: SIMILE demo storyboard feedback
> 
> 
> 
> 
> Here are some comments on the Demonstrator Storyboard, which 
> I had a chance to review more thoroughly.  These comments 
> refer to version 2.
> 
> First, I agree with MacKenzie that the Venn diagrams need 
> more work.  In particular, the first one is confusing. There 
> is no overlap between "images of artifacts designed by FLW" 
> and "courses about FLW".  Even if you use "images of 
> artifacts..." and "materials from courses..." there is no 
> real overlap, unless each community has the exact same image 
> cataloged.
> 
> I think instead what needs to be conveyed is that there are 
> two disjoint sets of information, those cataloged in VRA-Core 
> and those cataloged in IMS, but that the information need 
> spans both.  I've made a mockup of what I have in mind 
> http://www.ai.mit.edu/people/matsakis/simile-> venn.png
> 
> I 
> think the second Venn diagram is also a little 
> bit incorrect.  There are no elements that are both VRA-Core 
> elements and IMS elements.  Rather, there are elements that 
> are "expressible in VRA-Core" and "Expressible in IMS".  A 
> subtle distinction, but probably worth making. It is the 
> overlap between these two sets that is represented by the 
> vocabulary mappings.
> 
> In the section on vocabulary mapping, I think it is perhaps 
> sweeping some things under the rug to say that in the free 
> text search the system will "know that 'Wright, Frank L.' is 
> equivalent to 'Frank Lloyd Wright'". Making this mapping can 
> be viewed as part of the demo.  For the templated search, I'm 
> not sure I understand why the IMS records would not be 
> returned.  Is it that the lifecycle.contribute.author will be 
> the photographer, rather than FLW, or that there is not 
> lifecycle.contribute.author field in those records?
> 
> Nick Matsakis
>

Received on Wednesday, 29 October 2003 18:02:54 UTC