Application Wishes?

I’m looking to find out whether anyone has any desire to see new applications based on RDF. What would you like to see that hasn’t been done yet? What do you think would be a cool demonstration?

This can be web apps, local apps, whatever. Most of my stuff these days are in Python, most of my RDF stuff is based around Redland (when it’s not dying because of something I broke) and most of my platform-specific knowledge is Linux based.

Given these constraints, what “cool thing” hasn’t been done to pieces yet? Would a redux of Plink be in order, perhaps with better tools for automatic information removal to avoid the situation of angry users? A web service allowing you to use julie’s database for queries? Mail filtering? Image searching by all kinds of fun stuff, a la MortenF’s application, only with a user-upload section?

I really am low on ideas at this point, so I’m interested in hearing what thoughts other people might have.

7 Responses to “Application Wishes?”

  1. bpt Says:

    I would like to see:
    * LocalNames+RDF integration
    * BBDB+FOAF integration
    * RDF-backed photo galleries
    * RDF recipes
    * An implementation of jwz’s Intertwingle concept in RDF
    * Wiktionary that doesn’t suck, using RDF

    From that last idea, I wonder if some kind of generic, non-sucking RdfWiki would be possible. the RdfWikis I’ve seen did not appear to be particularly useful, but the concept itself is still appealing to me.

    I may be working on an RDF-based recipes project sometime soon, because there’s been virtually no work done in that area (AFAICT) and it scratches an itch for me.

  2. Jimmy Cerra Says:

    Something I started working on (but couldn’t finish since Research/School takes up too much time) is an automatic annotation system that scans documents and performs Baysean classification, LSI, or some intelligent reasoning to infer RDF statments that can be accepted or modified by the user. Sort of an attempt to make adding metadata easier for document authors (inspired by the infamous “Metacrap” article).

  3. Danny Says:

    Oo, got loads of possibilities, most buried under piles of to-dos. One big one is a decent RDF-backed CMS, along the lines of those at opensourcecms. Doc management built around DC/RSS/PRISM, user management based on FOAF. But must support arbitrary RDF vocabs, maybe with pluggable GUI bits. Set it up to be configurable blog/Wiki /aggregator/whatever.
    A quick way to start/experiment might be to pick an existing CMS and replace its RDBMS backend with Redland.

  4. Danny Says:

    Heh, Jimmy – I’ve been working on something along those lines, you’ll hear about it if I get anything useful…[it’s a really wide open area, it would be good stuff for crschmidt too]

  5. Claes Says:

    I think a kind of local RSS source for “events” in desktop systems would be very useful. The idea would be to produce RDF summaries, similar to RSS, for things like recently opened files, recently started applications, most started applications, recently downloaded files, recently played audio files and so on. A background process would gather data, perhaps using DBUS events, process it and make summaries available for other applications to use. These summaries would periodically be written to disk in a “standard location”, somewhere below the home directory. These files could then be used as source for another application – an individual web application perhaps – that aggregates the most userful information from the outside world with the most useful information about the user’s data. For example, the most recent emails are shown next to the user’s most often used applications, next to recent news headlines from a site of the user’s choosing.

  6. Danny Ayers, Raw Blog Says:

    8217;s (rather cool) XML indexer for dotLucene on literals in RDF/XML. crsmidt, you still looking for work?

    […]

  7. Daniel O'Connor Says:

    RDF recipes – coming soon to a fridge near you.
    RDF-backed photo galleries – Gallery2 developers were very impressed by the idea of RDF, if a module was developed for that, you might have something similar – and deployed by default.