[Making an old, private thread public.]

I need a browser tool. A user-agent. It needs to be able to allow me to create triples wherever I may be on the web. It should let me indicate the S-P-O. It should let me say if an entity is a person or software or a building. It should automatically link the entity type to a URI from schema that I set up. [Need Subject-Predicate-Object-Timestamp, actually.]

If the value is in the relationships between subjects and objects, then we can charge for access to this data. Rather than use an ad-supported model, charge companies to get access to the full blown ontology the system builds.

Needs to be integrated easily with all major tagging sites, like del.icio.us (or even Twine), so that users don't have to choose one over the other. (ie, one bookmarklet to rule the all) Users should always be able to see their own data, or look at other users data, but if companies want access to ALL the data, they need an API or whatever. After we've got enough data it should suggest what something is, like Adaptive Blue's organizer tool. Should also integrate easily with things like Calais. Should be pre-programmed with options for schema to use - Dublin Core, UMBEL, OpenCYC, etc.; the user should also be able to enter their own. That's one of the things that irked me about True Knowledge - I have to define each arc rather than point to an existing ontology. Eventually should be able to allow connection to licensed controlled vocabularies as well - eg D&B Company codes, S&P's GICS. URIs MUST be dereferenced.

Need a large triple store somewhere in the cloud. Should work on all browsers - YES, all. Do it right or don't do it at all, none of this "safari is only 5% of our traffic" bullshit. Need to balance comprehensiveness with a regular schedule of small releases to maintain interest and PR traffic.

The site itself should be simple, but rather than just a del.icio.us type interface, also provide a graphical representation of the graph each user is building over time. Gruff as an example? Good, but needs polish. Also should allow users to export their data in a variety of formats - doc, spreadsheet, RDF, OWL, SKOS, JSON, N3, Turtle, etc. User authentication model should include OpenID. OAuth? SIOC? (Maybe not SIOC.) Plugins for popular web cms/portals? WordPress, TypePad, Drupal, Joomla, Plone, OpenCMS, etc. Integration with library systems like Koha? Hm, no, those last couple might kill the commercialization capabilities. Maybe. Need to ponder.

I'm sure I can build on this, but I got out of bed to jot down these ideas, and I have to be bright-eyed and bushy-tailed for Media Training in the morning! What do you both think?