Website Design

For this website for collaborative generation of a linguistic corpus and ontology, I'm thinking about a drag and drop based UI, where users can work with sentences and, using gestures, enter the knowledge obtained from reading them.

I downloaded phalanger for PHP integration into Visual Studio and .NET (they also have one for Mono). I'm looking at prototype and mootools for the javascript library. Also found this extension to mediawiki that adds web services to wiki (SOAP, WSDL) — I'll probably code some web services into it.

I'm thinking that the site will show articles with entities (nouns) as draggable objects that can be manipulated in the environment to allow users to easily provide the knowledge they obtain from reading (sentence by sentence). This will include anaphoric resolution and semantic relations between entities. By processing entire articles, context will be obtainable from the dataset. Hopefully the interface will be intuitive enough to encourage users to provide an ever-improving downloadable resource comparable to penn treebank and redwoods.

Code Storage, Collaboration

If anybody visiting knows a place where I can upload code to and link to here, please feel free to comment (on any blog entry for that matter). I might end up using an open source repository, the only problem is I'm not sure if some of the jar's or IKVM'd dll's (etc) EULA's allow that. As for the gb's of Wikipedia data after I run some tools on it, I might have to torrent that. Feel free to contact me or comment here if any of these projects sound interesting.

Also, if you know of any downloadable SRL projects or tools please let me know. [Update: found swirl and salsa]