No matter how many ‘artificial intelligence’ approaches occur to help structure data, that data isnt very important (in fact, in some cases, devalued) unless it can be accessed in a controlled manner.

Thank you very much for visiting the blog.

Of course, BeFogg.com will be the first “vertical service” where we apply the principles we outline here in this blog.

Best regards,

Carlos

Offer any product/service to experienced travelers is so hard … because there are information but dispersed, isolated and no related… the Web3.0 techniques have to help us.
I think that beFogg is the best place for applying Web 3.0 principles, no?

I invite you pisco sour.. come on!

-roger

Thank you very much for your comment. :). We’ve just opened our blog, and we’re very happy to start receiving feedback.

Here goes some feedback.

“Pretty much everyone that uses a computer already *does* generate and maintain machine-processable data - after all, that’s all computers know. There’s material like enterprise databases, personal address and appointment books, music catalogues, project management material, etc etc. Even things which are content-oriented like blog posts and emails have a lot of data associated with them.”

Totally agree. When I said “generate” machine-processable data, I was thinking in the “generation” of structured formats in the process of an information exchange (i.e. people opening and publishing their internal databases). Sorry for not making this more clear (maybe I should have said “publishing” instead of “generating”).

The problem is that, when we’re speaking about the vast majority of the people in the Web, the motivations for publishing their meta-data in the Web are not very clear.

For end-users, doing it explicitly is not very attractive (at the end, it’s more work). In this case, the best strategy is to let our software do the job for us (i.e. a blog software generating hCard with our contact data or a social network generating FOAF with our contact list), but this is no great deal (you can mine this information now with a good accuracy).

For companies, new business models have to be defined in order to open and publish databases, specially when we’re speaking about making information freely available. We can think about Reuters or TeleAtlas, among others, that have been providing high quality meta-data for years (in exchange of a considerable sum), but I’m sure that’s not the “Semantic Web” you’re envisioning (Reuters and TeleAtlas operate in somewhat “closed” environments). We’ve got a long way to go here.

Governments are not constrained in the same way as companies, and many times provide publicly-available meta-data (i.e. geographical info provided by the US Government).

Summarizing, it’s not a problem of existence of structured data (as you point, it does exist) or technical feasibility of publishing it (it’s straightforward), it’s a problem of finding motivations for doing it (why?), and finding the appropriate scenario (market, business model, …) for doing it.

I feel obliged to comment on this point:

“Reaching the full potential of the semantic web would imply that almost everyone in the Web generates and maintains machine-processable data.”

Pretty much everyone that uses a computer already *does* generate and maintain machine-processable data - after all, that’s all computers know. There’s material like enterprise databases, personal address and appointment books, music catalogues, project management material, etc etc. Even things which are content-oriented like blog posts and emails have a lot of data associated with them.

“Clearly, this is a big hurdle for this vision to get real, mainly due to human and organizational behavior issues, as reported by Cory Doctorow’s Metacrap.”

There certainly is a big hurdle, and it is mainly due to human and organizational behavior issues, but it’s not primarily in creating data, it’s in exposing it on the Web in a useful fashion.

Cory Doctorow’s Metacrap was something of a red herring because it only really considered manually generated metadata. Not all data is metadata, and a lot of metadata is produced automatically (like the material in this blog’s feed, for example). People and organizations can and do create quality data (I hope my bank does, for starters) when sufficiently motivated and/or helped by tools.

Ok, there is a lot of human-readable content on the Web, and in that context metadata is significant. But even there, things like folksonomy tagging and Wikis suggest this isn’t an insurmountable obstacle.