Re: PiggyBank Usability Feedback

From: David Huynh <dfhuynh_at_csail.mit.edu>
Date: Tue, 01 Feb 2005 07:37:53 -0500

Stefano Mazzocchi wrote:

> Daniel O'Connor wrote:
>
>> Hi guys,
>> Some very nice work with Piggy-bank, I love having *something*
>> tangible to play with after authoring so much RDF :)
>
> People, thanks for the kind words. I have to agree with all of you
> that PB is simply a great idea. Kudos to David for such an exciting
> software! (and for having the great idea of using java *directly* in
> the browser and not just as a little applet, who would have thought :-)

Thank you all! But really, I've done nothing more than integrating great
pieces of software like Longwell, Jena, HSQLDB, Lucene, Jetty, Velocity,
etc. And of course, Firefox, which has shown itself to be a great
platform to build upon.

>> Things I'd like to see:
>> (1) Blank nodes appearing lower down in the sidebar. Reasoning:
>> Piggybank is aimed at end users, and a blank node is meaningless to
>> them. Something they can recognise, like a name ("Daniel O'Connor"),
>> helps them understand what piggy bank is actually doing - collecting
>> information.
>
> Yes, this was the first thing I told David myself. My suggestion was a
> little different, though. bNodes are, in fact, nodes that are not
> identified explicitly by a URI, therefore they are *part* of the
> explicitly identified node that they are connected to.
>
> Personally and practically, I don't see much difference between a
> bNode and a parsed literal, so I suggested David to treat any bNode as
> a parsed literal (in short, a literal that is an XML fragment instead
> of just a string). This would allow us not only to remove the bNodes
> entirely, but to be able to save them once saving the identified node
> that includes them.

Agreed totally.

>> (2) A loading bar / status bar. Not having visual feedback on how far
>> the snippet loading had gone I tend to forget to save it.
>
> Yep, I had the same exact suggestion... a little light at the bottom
> that turns on/off depending on the fact that the page has semantic
> tidbits or not and, if you go over it with your mouse, it shows a
> summary of the tidbits 'items' that it found... and if you click on
> it, the sidebar shows up.

Stefano and I did also discuss the idea of a small, translucent popup
window that automatically slides up from the status bar to show a
summary of RDF tidbits it picks up (e.g., "Found 3 Contacts, 27 News
Items. Click here to save all.").

>> (3) An easy to see 'auto save' and 'auto scutter' checkbox in the
>> sidebar pane.
>> It's bothersome to save everything and I would love to be able to
>> search *everything* I've browsed, no matter what it is.
>> Secondly, an auto scuttering option that chased down links in the
>> background would be handy. Not sure how possible however :)
>
> eheh, Ryan already integrated scutter in longwell, not sure if David
> picked that up but it's sure possible.
>
> The problem (as usual) is where to stop cuttering. The problem with
> automatic harvesting is that it might generate a ton of RDF that you
> might not really be interested in.... but it's true that we could
> follow and try to extract everything that looks like an ontology that
> might help the system describe what you are looking at... that
> shouldn't be too bad.

OK, maybe we'll have a little icon on the taskbar that has a stop sign
on it if auto-collect is off, and no stop sign if it's on.

Another idea: I was thinking of making the web server in Piggy-Bank
serve XUL instead of HTML. We'll get more interaction with XUL and we'll
show the full potential of Firefox as an application platform that way.
Any thought?

And yes, we are trying to find way to eliminate the web server without
sacrificing the features Piggy-Bank has offered so far (such as allowing
you to bookmark any URL it serves up).

David
Received on Tue Feb 01 2005 - 12:37:59 EST