h a l f b a k e r y
I didn't say you were on to something, I said you were on something.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
or get an account
Please log in.
Before you can vote, you need to register.
Please log in or create an account.
I haven't gotten to the point where I bookmark useful searches I'd like to take a look at later, partially because I know I'd never get back to them. But for when I am doing a search I think something like that but more task-oriented would be neat.
One would keep adding keywords/searches to it, and
it could also automatically generate searches based on recombining several keywords. (This is definitely something which could be automated; when I have a handful of terms often enough choosing two at random and making that a search is good for results, or sometimes I just need to create sort of a multiplication table of results.) Then the searches (performed on various search engines) would become 'to do' items. The whole group of searches would also be one big 'to-do' (for the sake of checking it off all at once when 'done enough'). Individual searches would be checked for overlap; that way the user wouldn't have to manually check how much new was in each. But the user would individually check off each search as sufficiently checked out (though each goes on of course many pages, one usually closes the tab before the true end of potentially interesting content, and this would be just like that- which is what I mean by 'done enough').
Items in the search would individually be earmarked as interesting, though that's different than marking them as certainly relevant- they just need to be looked at. This is like opening the item in a tab. However, items could also be marked as 'certainly relevant', in which case a whole series of investigations might be automated; the program could do a Google 'related pages' search, tt could find statistically significant words within the document and offer to add those to the general pool, and maybe (maybe) it could do some sort of meta-searching where it uses the document to learn more what the whole search group is about, and tries to point the user to other good items faster.
I don't know if this would work best as a Firefox plugin or something standalone based on Google API or maybe a website. If it were a website it would be funny because it would get marketed as a new search engine, which is kind of silly.
||Does the "WorldBrain's Memex" browser extension do what