Ask HN: Feedback on an idea for a service

So this is an idea that's been kicking around in my head a while. I've sought feedback on it before, and had two different responses - an iPhone app dev said it sounded quite useful, while someone whose opinion I respect immensely was skeptical.

The idea is to create a "location wiki" - an attempt to describe every location with as much metadata as possible. This overlaps with what Chris Brogan wants Brightkite to become, "the annotated world", here http://www.chrisbrogan.com/if-i-owned-brightkite/ and I think an opportunity does exist to create this "annotated world" resource (though I don't think Brightkite is it).

This could have a couple of different interfaces: a website for users to visit, and an extensive API for developers.

The website would let users look up a location and see a description, photos, related locations, websites, whatever. For example looking up a train station would show you the line that station is part of, the suburb it's in, nearby businesses, etc., but the entry would also contain lower-level data about what type of place this is (eg. public transport), and could pull in items from other services that are geotagged.

But the main benefit I'm envisioning is having a service like this with a reasonably complete dataset ready to be plugged into third-party apps. As far as I'm aware, iPhone apps like nrme cobble together business listings and whatnot from a few different 3rd-party resources, but with this service, they could plug it in as their location-contextual backend and immediately have that data and functionality available. I am aware Google Maps also contains a lot of extra data about locations these days (business listings, photos, etc) but it's by no means comprehensive, nor does it aim to be I think.

There would be a few different ways to monetise, but I imagine selling ownership of a business' wiki entry would be the easiest - a business could buy their entry and restrict editing, publish updates, link to their website, whatever.

The main criticisms I've had so far are:

* Nobody needs it - I disagree, I think it's something that will only become more relevant as mobile/location-contextual internet use increases

* Not feasible, too large in scope - I agree in that I do think getting a reasonable amount of data would be the hardest task. But you have to start somewhere...

Would love to hear your feedback. If it's a stupid idea then please be gentle :) If not, who wants to start working on it with me?