Tom Morris

Over the last few days, I've switched over to Google Reader. I like NewsRiver, but I find that Google Reader gives me the same kind of functionality, but with a few less annoyances.

For the most part, I use it in almost exactly the same way I used NewsRiver - load it up with feeds, click "All Items" and surf through the river.

I do have a few little things I wish that the Gfolks would add.

Firstly, on the mobile version, when you 'star' something, it should load the next headline. When I'm reading headlines on my phone, I'm doing a simple thing - going through and saying "yea" or "nay". Star means "yea", skip means "nay". Don't show me the same headline again, okay.

Secondly, please, please, please do tagging properly. If you are going to have tags in Google Reader, don't have a controlled vocabulary. Go to del.icio.us, see how they do it, and replicate. The same is true with Gmail. We don't want labels, we want tags. They're different things, you know.

Similarly, the "share" function should tap in to external APIs like del.icio.us. If I click 'share', it should give me a little Ajax pop-in-front window so I can tag the stuff and have it saved automagically in to my del.icio.us.

Release the API. I want to have an offline version of Google Reader. However much we dream about universal availability of access, it's never going to happen. I've said it before and I'll say it again - tunnels. Nice as it would be to use the 'net while in the Channel Tunnel, it's not happening is it?

Have a blog copy-n-post system. Make it so I can replace that silly 'Email' link with a 'Blog' link. And, more importantly, let me set what happens when I click it. I just want some HTML fragments to copy and paste. I don't want to post it to Blogger or through the MetaWeblog API.

Give me my attention. All that attention data are not belong to you.

And make a linkable OPML file available.

Comments | TrackBack

I'm slowly getting in to the Semantic Web stuff. It's a slow process. for one main reason - RDF and SemWeb people have a tendency to multiply acronyms, abbreviations and technical terms like there's no tommorow - hence RDF, OWL, DAML, GRDDL, SPARQL and so on.

I validaed my first RDF file this morning - I wrote it in the super-excellent Oxygen XML editor (I bought a licence for it yesterday). Then I can go on to the W3C website and use the RDF Validator - which not only shows me errors, but also produces a list of all the triples in the document and shows me a PNG file with the data in visual form.

Next I want to figure out how to scale this up using a database. I'm thinking it would be amazingly cool to have a triple-enabled del.icio.us-type system. Imagine - you add a bookmark to your database, and it then prompts you for a whole load of data based on whatever ontologies you've told it about.

It also would show you all the resources in the database which make reference to the bookmark you are adding. Semantic social bookmarking. I'm not sure I've totally got my head around it, but it's worth thinking about.

It's something I've thought about with regard to OPML - basically, have a way to mark a specific bunch of data as belonging to an OPML file in, say, del.icio.us.

If this kind of thing sounds exciting, why don't you hire me to be your R&D department? I can sit deep in the caverns of your company and fiddle about with SemWeb stuff. The more I try to grok this technology, the further I get away from working out how I could turn it in to a business.

I'm writing my RDF in XML, rather than in Turtle or N3 or N-Triples. That's because I'm already used to working with XML, but converting to and from these other formats is something I haven't yet figured out. I know it's possible, but doing it on-the-fly is something I haven't yet worked out.

What would be really cool is if the people who make Oxygen would build in better support for RDF right in to the Oxygen environment. That'd make my day.

Comments | TrackBack

Whoosh. That sound was a whole room full of university administrators gleefully missing the point. Nobody ever said that "Mickey Mouse degrees" weren't economically valuable. That was never the point. What we said was that they weren't appropriate for academia in the way that, say, history or mathematics is.

I have nothing against developing skills relevant to the market. But academia isn't the place for that. I go to university to learn about philosophy. It does a pretty good job of giving me a good overview of the subject and guiding me to read the appropriate things. Meanwhile, I teach myself how to code - mostly because I'd lose it if I had to sit in a classroom learning it.

Academia is there for the provision of academic subjects. That doesn't mean that non-academic subjects shouldn't be taught. Just not in university. Funnily enough, we had this system called polytechnics before. If you wanted to learn academic things, you went to an academic institution - like Oxford or UCL. And if you wanted to learn technical or trade-related things, you went to a polytechnic.

Somebody who knows how to make surfboards provides a valuable skill. But it's not an academic skill - it derives from certain academic disciplines (physics, engineering) as well as certain aesthetic choices. Meanwhile, studying old books is an academic skill. Both are valuable - but in different ways. This new report doesn't seem to actually take much account of this though.

Comments | TrackBack