A real wow-project has gone into ver­sion 2: Exhibit. It is part of SIMILE, focussing on “Seman­tic Inter­op­er­abil­ity of Meta­data and Infor­ma­tion in unLike Envi­ron­ments”, which pro­vides a whole tool­box of prag­matic seman­tic web applications.

Exhibit itself presents a “a three-tier web appli­ca­tion frame­work writ­ten in Javascript, which you can include like you would include Google Maps.” An exhibit appli­ca­tion typ­i­cally con­sists of a con­tent pre­sen­ta­tion area and sev­eral wid­gets for fil­ter­ing, sort­ing and group­ing the con­tent pre­sen­ta­tion, fol­low­ing the faceted brows­ing par­a­digm also used in my elas­tic lists. Cur­rently, the avail­able wid­gets cover plain check­box lists, maps, and time­lines, and a live search.

From my first exper­i­ments with the tool, I can say it is really ridicu­lously easy to cre­ate cus­tom views on exist­ing data sets: Data can be imported “from a Google Spread­sheet, Excel spread­sheet, Edit­Grid spread­sheet, Bib­Tex files or any JSONP data source” with­out much effort.

Appli­ca­tions are built by load­ing the exhibit script and then assign­ing “Exhibit roles” to lay­ers in your java-script: For instance,

<div ex:role="facet" ex:expression=".discipline"
ex:facetLabel="Discipline" />

will cre­ate a fil­ter­ing wid­get look­ing for val­ues of the “dis­ci­pline” prop­erty in your data set and dis­play them as a list to be used for fil­ter­ing. It couldn’t pos­si­bly be eas­ier to build client-side facet brows­ing applications.

Other nifty are the inte­grated browser his­tory for fil­ter­ing steps and the abil­ity to export fil­tered views as e.g. HTML, but also tab-separated text files, RDF, etc.

Check out the Get­ting started page to get an impression.

The down side, of course is, that all data has to be loaded on the client before the appli­ca­tion can start. So we are rather speak­ing about hun­dreds than thou­sands of data items. Also, I would of course be inter­ested in build­ing cus­tom wid­gets, how­ever, this looks a wee bit more com­pli­cated from first looks into the code. And of course UI-wise, I would have some sug­ges­tions as well, but any­ways, I find the project quite impres­sive already as it is!

Kudos to David Huynh and the rest of the team at MIT. More info can also be found in David Huynh’s PhD the­sis.