Re: fresnel thoughts

From: David R. Karger <>
Date: Tue, 22 Mar 2005 19:15:34 -0500

it is certainly true that i can choose a different intermediate representation
for data being passed to haystack than to longwell, and have rdf resources
referred to in that representation. But even if I am able to use the same
lenses as longwell, I will have to create a brand new set of styles to output
data in the different intermediate representation. It would be much less
work if I could reuse the existing lenses.

I suppose I could write a different _interpreter_ for the styles; one that
included references to the rdf in the intermediate representation it
produced, but that otherwise generated the same intermediate representation...
        From Tue Mar 22 18:59:19 2005
        Mailing-List: contact; run by ezmlm
        X-No-Archive: yes
        Reply-To: <>
        X-Authenticated: #1074481
        From: "Chris Bizer" <>
        To: <>
        Subject: Re: fresnel thoughts
        Date: Tue, 22 Mar 2005 18:59:07 -0500
        X-Y-GMX-Trusted: 0

        Hi David,

> I've been thinking a bit about how much of fresnel I can use with
> haystack. It is my hope that we can use both the "lens" and "style"
> vocabularies, and essentially to use the same lenses and styles that
> are being used in longwell.

        Yes, this would be cool.

> I did realize one interesting
> problem---one I created myself by insisting the hierarchical
> representation not contain rdf. Haystack is in general going to be
> pretty happy with the hierarchical layout generated by fresnel's
> lens/style combination. But items in haystack are supposed to be
> active---when the user right clicks, or drags one, haystack is going
> to have to know which _resource_ is being addressed by the user.

        We ruled actions out of scope for Fresnel at one of the meetings, so
        Haystack could extend Fresnel with a vocabulary for actions, if you guys
        want to model them using RDF.

> We
> can easily find out which _visual element_ of the hierarchical model
> is being addressed, but since we took any reference to the underlying
> rdf out of the intermediate representation, how can haystack map back
> to the underlying resource?
> It would be a shame for this to be the one issue that prevents
> haystack from adopting fresnel completely, so suggestions for fix
> would be appreciated!

        I don't see a major problem here. The intermediate tree I proposed yesterday
        hasn't contained this references. But there is no problem in extending this
        intermediate tree to contain the references you are needing.

        But again. I see the intermediate tree as an implementation issue. Fresnel
        lenses specify which data should be displayed. How this data is passed on to
        other parts of the display engine is complete up to the needs of these


Received on Wed Mar 23 2005 - 00:15:34 EST

This archive was generated by hypermail 2.3.0 : Thu Aug 09 2012 - 16:39:18 EDT