> Date: Fri, 2 Nov 2007 00:41:14 -0400
> From: [EMAIL PROTECTED]
> To: [email protected]
> Subject: Re: Exhibit large data sets...
>
> Stefano Mazzocchi wrote:
>> Neil Ireson wrote:
>>
>>> Hi all,
>>>
>>> My major sticking point with the using Exhibit is the limited number of
>>> items it can handle. I know this is not it's primary purpose, I think David
>>> has said somewhere that it is suited to display between 50-500 items,
>>> whilst I (and I believe others) would be interested in applying the
>>> software to 1000s or even 10000s of items. I know there is Longwell but it
>>> is not as simple to prototype and the interface is just not as intuitive as
>>> that in Exhibit.
>>>
>> [snip]
>>
>> It's worth noting that David wrote both Longwell-CSI and Exhibit, in
>> that order.
>>
> We are closing in on hopefully the right solution. First, we started
> with Longwell, which scales relatively well but is difficult to
> configure. Then we built Longwell-CSI, which is easier to integrate with
> an arbitrary web site but is still not sufficiently configurable. Then
> we built Exhibit, easy to configure but does not scale. The next thing
> will incorporate all of the lessons we've learned, and we're targeting
> both scalability and configurability. That is not trivial and will take
> us some time.
Firstly it seems churlish to make anything which could be seen as a criticism
of your work. It is absolutely fantastic that you are providing the Simile code
and and the fact you respond so well to the request of users is even more
amazing.
I was having a look a the Longwell code which is, as you say, difficult to
configure, and saw the Longwell-CSI existed but unfortunately couldn't really
work out how to use it. The CSI examples use proprietary data, which means I
couldn't get anything displayed to work out what's happening. I could try to
use the Dwell example but that would involve a considerable faff to get DSpace
up and running and populated with my data just to extract it and then convert
it back into RDF and the example server mentioned on the page
http://simile.mit.edu/wiki/Dwell is not recognised.
Anyway in the end I am returning back to Exhibit as the idea was to do some
quick prototyping for some user experiments and sampling my data down to a
manageable size.
> Onto the more technical details, if you want your exhibit to render only
> a few items, then you need to set both
> ex:grouped="false"
> ex:showAll="false"
> This is because if grouping is on, then the showAll setting is overridden.
>
> You can also configure the number of items you want to render
> ex:abbreviatedCount="50"
I am using this approach already. I have tried to remove the grouping option
widget from the display to prevent user selecting this and having to wait a
while for the rendering of all the items but without success (I am not at all
familiar with Javascript). Also this method does not work for tabular views.
> We currently don't support pagination, but we would welcome patches to
> our code!
I would love to make some meaningful contribution if time and ability allow.
> By the way, Safari 3 is quite a bit faster than Firefox and IE.
Thanks for the pointer but when I tried Safari versus Firefox my completely
unscientific tests showed they both took exactly the same amount of time to
load and render a large data sets.
N
PS One thing I did like about the Longwell approach is that it used RDF (RDFS,
OWL, N3) data, especially the fact it uses the Sesame code (another fine
OpenSource project). I was wondering why the move from RDF to JSON in Exhibit?
Also in the future super Exhibit-Server approach is there any vague intention
of making this available as a front end to Sesame which would, I think, be a
killer application.
> David
>
> _______________________________________________
> General mailing list
> [email protected]
> http://simile.mit.edu/mailman/listinfo/general
_________________________________________________________________
Feel like a local wherever you go.
http://www.backofmyhand.com
_______________________________________________
General mailing list
[email protected]
http://simile.mit.edu/mailman/listinfo/general
