Re: Are you the right programmer?
On 2 Sep 2004, at 15:49, irv mullins wrote:
>
>
> posted by: irv mullins <irvm at ellijay.com>
>
> CoolStuff-USA wrote:
> >
> > I need a technology and/or someone who can help me implement my solution
> > vision:
> >
> > - Allow someone to identify their preferences
> > - Based on specific subject content preferences, serve up the best fit
> > - Assumes an AI or Matching engine
> > - .ASP web based solution
> >
> > For example, I want to buy or catch a pet butterfly. My problems are that I
> > do not know what all the butterfly's are in the world. I do know that I
> > like
> > certain
> >
> > colors and may also have certain constraints at my house and in my yard. I
> > want to be guided through an identification of my preferences, then be
> > presented
> >
> > only with a small sub-list of only the top matching butterflies.
>
> You are making a major assumption here: that a computer can or will
> somehow know more about butterflies and their environment than you do.
> That might be possible, given enough data from someone who DOES know a lot
> about
> butterflies, and geology, and meterology, etc. You are highly unlikely to
> obtain
> all this information and have it entered into your computer for free, of
> course.
>
> In order to help with decisions in even a limited field (selecting a
> small pet, for example) you are going to need a huge amount of data,
> checked for accuracy. If it isn't accurate, your program will be
> worthless (Google would be just as good) Achieving accuracy will be
> expensive.
>
> So you see, it's really not a programming problem.
This is quite true. A random enquery into Tiggr's *raw* db on another topic,
for info on a specific item (like your butterflies) where she has 50 or more
megabytes from *each* of different sources on the same topic, showed that
one source never heard of the item, two had heard of it but were missing very
important data, two had the data but some of it was wrong, and one source
had the correct complete data. That gave me a 16% chance of correctness if
relying totally on another's data, with no error detection and flagging. That's
not very good odds.
Lucky for me, i suppose, all 24 hours of every day are mine. And a fancy
difference engine. And i am looking for such a sponsor as this, but he just
gave you all the data he wouldn't give me about the project.
Kat
|
Not Categorized, Please Help
|
|
