it's looking pretty good; certainly is generting the cat field
properly. A couple of notes:
* export generated html output is missing. That is necessary to create
googlable pages.
* it looks like some characters are being transofrmed incorrectly. Note
the question marks in potluck abstract, which i think should be dashes
* will the converter properly handle quotation marks inside a field? ie
if I say in bibtex
abstract={This is a "good" paper}
what will happen to the json?
* you've generated a proper accent mark on the e in Medard in the
content, but in the author facet it shows Medard with accent but also
M\'edard, and also shows M<gibberish>dard. but then you have proper
accents for laszlo lovasz. strange.
* I wonder if babel should generate a "bibtex:" attribute that stores
the raw bibtex associated with each the json record. this would give
you full-fidelity export of bibtex. Some work to deal with the bibtex
"crossref" attribute, but still, might be easiest overall. Because then
you could safely zap various tex things from the converted fields.
* I got surprised by the facet behavior, though i can now see a
justification for it: clicking on a value selected that value _only_.
to get multiple values I had to careful hit the checkboxes. Is that
intentional?
* I got tripped up by the "copy raw data button"---didn't expect it to
output everything, only the stuff I had currently selected.
* the converter messed up on the name "Douglass S. J. {De Couto}" (look
under C in the author facet). tex uses the braces to say "this person
has a two word last name". I have to ask Max Van Kleek if his name is
wrong too.
* You'll note i am both David Karger and David R. Karger (you have two
names too). It would be nice if there were a way to "merge" these two
things in the author facet. I guess one approach is to create a "name"
type, and the give each name a "person" property, such that two names
for the same person have the same person property, and then "group by
person", but this seems like overkill, doesn't it?
Daniel B. Giffin wrote:
> in the display of an item, i'd like some of the items that occur as
> its property values to be rendered "inline" (directly in the page)
> instead of as popups.
>
> (there was an earlier thread under the subject "Exhibit Handling
> Hierarchical JSON Data", and this could be one way to show some more
> of the graph's depth.)
>
> i think inline rendering could be specified in a lens by adding an
> attribute like "ex:inline" to an element that uses ex:content.
>
> for example:
>
> <div ex:role="lens" class="nobelist">
> ...
> <span ex:content=".co-winner" ex:inline="true">
>
> this would mean that the .co-winner property values should be rendered
> inline here.
>
> of course, there's some danger of writing recursive lenses that go on
> rendering dom nodes until your browser explodes. (i think my example
> would do that.) but i don't think that's a reason not to provide the
> facility.
>
> thoughts?
>
> (i might have a go at implementing this myself. i'm eyeing
> Exhibit.Lens.prototype._constructDefaultUI(). it seems like it could
> be straightforward, although it looks like there are some subtleties
> concerning rendering an item more than once, and perhaps other stuff i
> haven't thought of.)
> _______________________________________________
> General mailing list
> [email protected]
> http://simile.mit.edu/mailman/listinfo/general
>
_______________________________________________
General mailing list
[email protected]
http://simile.mit.edu/mailman/listinfo/general
