>[signing in]: It's easy: just connect your Facebook or Twitter account [and no other means]
So "easy" is not allowing me to make a login not tied to everything else.
Anyone with an account care to inform me / us what you can do with an account? Just like / want it, or add thingds, or...? I'm interested, but the site contains no info about the site aside from the TOS and privacy policy.
I should really make a throw-away set of accounts on a handful of social sites, just to get around crap like this. If they allowed straight OAuth or OpenID instead of Twitter-hosted OAuth, maybe I'd use it. But nobody does that.
Joel Spolsky predicted this project 10 years ago :)
"When you go too far up, abstraction-wise, you run out of oxygen. Sometimes smart thinkers just don't know when to stop, and they create these absurd, all-encompassing, high-level pictures of the universe that are all good and fine, but don't actually mean anything at all."
Any website or web service with a database of "things" (with attributes) is a competitor to Thingd: WorldCat, IMDB, gdgt, Google Merchant Center, Zappos, CNET reviews, Etsy...
It's one of those ideas that seem so obvious and that has the potential to change the world, but it is highly dependent of having critical mass. Technologically, it is very simple to implement it. One doesn't need much more than something like CouchDB, and you could have users changing things even from the admin interface.
With job4dev (shameless plug), we have something very similar, applied to job listings and company info. We created a "wikipedia for jobs", but lots of the curating needs to be done by hand. We are automating as much as we can, using as many APIs as possible, but there's still the need for human involvement.
The problem is, you need to have people wanting to use this. It's the ultimate chicken-and-egg problem. I could see this working if there is some sort of incentive, or even if it ties itself to Mechanical Turk. Without that, it is no different than Freebase, dbpedia, or Aaron's Infogami.
freebase and dbpedia are tied to wikipedia wiki articles about concepts and they seem to be trying to do products.. different data domain I don't know which one is "better" but this seems to have commercial tie-ins where wikipedia might not.
Freebase was seeded with Wikipedia, but we're always loading data from new license compatible sources. We've got books from Stanford Library, music from MusicBrainz, films from Netflix, and a lot more.
We don't have good coverage of product data right now though.
Thinglink ( http://thinglink.com ) has tried several approaches in this space. Originally they offered this very abstract kind of "URL for things" service, then they pivoted to be more a social network around things, then they pivoted to be a social network for designers and people who like designer clothes and other designer objects.
Now, their final pivot was in my opinion marvelous, as it went to very basics and offered a super-simple service to tag things in photographs. It's still operating in a same space "to URLify real world objects", but it offers an immediately useful service, for example for fashion bloggers.
At the very heart of it, it could map every object to its BOM including carbon based lifeforms. Imagine if every object were a class, including humans. The atom could be a basic super class, and you could map out various dependencies and object superclasses (e.g, Oxygen < Atom < Base). And it might even work with non physical things (Capitalism < EconomicTheories).
3.75 inches
3.75 inches tall
3.75 x 3.75 x 5.75 inches
3.75 x 4.5 x 7 inches
The problem with human entered / non moderated data is that these lists will become longer and longer over time and the database will not understand that they all mean the same thing. Joins will be impossible and its relevance fades.
I still believed (or hope) that semantic web is the future of Internet.
If a very particular industry willing to open up their data in standard and extensible format (may be RDF or a simpler standard), like film theater listing their showtimes or governments release their statistical data in RDF rather than Excel only, many useful application could be developed on it and it is a big step towards semantic web already. And I think it is not too hard to do.
If they don't stop at just describing the semantic knowledge that a person would derive from the experience of objects, but instead also include the process knowledge—in a rigorous, possibly algorithmic form—then it would be the world's biggest MOO.
So "easy" is not allowing me to make a login not tied to everything else.
Anyone with an account care to inform me / us what you can do with an account? Just like / want it, or add thingds, or...? I'm interested, but the site contains no info about the site aside from the TOS and privacy policy.
I should really make a throw-away set of accounts on a handful of social sites, just to get around crap like this. If they allowed straight OAuth or OpenID instead of Twitter-hosted OAuth, maybe I'd use it. But nobody does that.