Beat me to it again
Damn. I guess I shouldn’t be surprised. Someone beat me to it — this is what my first project with Ruby on Rails was supposed to be:
It’s is a tool that dynamically helps you choose among dozens (hundreds?) of content management systems.
(For those shopping for CMS’s see also opensourcecms.com, which lets you try out dozens and dozens of CMS’s and related things. (unfortunately this is limited to apps written in PHP4.))
This “decision maker” matrix is something that I have been wanting since like 2001, when I was sort of working on a version in Perl (or was it PHP?) to make such a thing for cell phones. (Choosing a cellphone plan can be daunting for the newcomer, if you remember that far back.) But then of course, I realized that such a tool would be useful for all kinds of purchasing decisions (Like a layer on top of pricewatch that does queries for you.) Surely, hundreds of other developers (entrepeneurs?) have thought of the value of something like this: Image a “push” ebay that stores your queries for desired items, and stores those items in a massive hyper-normalized (?) database. (A concept net? A knowledge-base?) It then notifies you when it finds items that fuzzy-match your query (within some tolerance).
Oh.. right. Amazon does this (among other online superstores). But does amazon have frontend that lets you construct queries with arbitrarily deep granularity with respect to the “problem domain”? What the hell am I talking about?
Surely, such a thing should be on its way to the surface by now, no? How come I haven’t heard of it? (It’s a rhetorical question — please don’t answer with an insult unless it’s clever ;))
I imagine the biggest challenges facing such a system would include but not be limited to:
- keeping data current (spiders? trust metrics?)
- keeping data accurate (”truthful”)
- keeping data [properties? aspects? attributes? features?] unbiased and subjective
- choosing aspect that can be compared to each other (aspect association)
- allowing new aspects to be added dynamically
As for “keeping data current”, you would want spiders, when possible. I have always wanted a system that recognizes a spider as a “user” or “contributor” along with its human counterparts. (To have both humans and bots in the same table makes me shudder with excitement and reminds me of a story.) As for trust metrics, imagine a slashdot where people contribute not articles but well-defined, structured data?  (If you could get this easy enough for end-users to do, and you could set up an online shop, people would probably come knocking.)
But surely there’s something I’m not thinking of.
Well I can’t wait to see it when it happens. Shouldn’t be long now.