[ai_group] Re: Database format
- From: "Jeethu Rao" <jeethur@xxxxxxxxxxxxx>
- To: <ai_group@xxxxxxxxxxxxx>
- Date: Thu, 18 Apr 2002 15:42:16 +0530
David,
Why not create an individual file for every object ? Or we can
have many objects in a single file, with another file as the index for it ?
I guess the 2nd Idea is a lil bit better.Whatever it may be, I'd prefer
an OOP language for this kinda work. Alice is written in Java, but Java is
quite a bit slow for this kinda work (The new GNU GCJ native code compiler
for Java may be better, but I've never tried java on GCC). C++ would be a
good
option. The best feature is STL, which provides many flexible alternatives
to
C arrays. Anyway, it still is possible to do it all in BCX or C, but It'll
be
quite a bit harder than in C++.
The Language used is not very significant.The main things that matter are
the
underlying ideas and algs. And yes, speed.
If you're interested, I have written a C++ Vector like library for C.
Its a kind of storage class (An automatically resizing array). It will
surely
work fine with BCX. It might just come in handy in your AI parser.
Jeethu Rao
> -----Original Message-----
> From: ai_group-bounce@xxxxxxxxxxxxx
> [mailto:ai_group-bounce@xxxxxxxxxxxxx]On Behalf Of David Garner
> Sent: Thursday, April 18, 2002 4:57 AM
> To: ai_group@xxxxxxxxxxxxx
> Subject: [ai_group] Database format
>
>
>
> Hi All,
>
> As I've stated before, I've thought about the database formatting and I
> think an adaptation/adoption of Dr.X's proposal of a single character
> code at the start of each line and a single item per line would be the
> easiest and most flexible format to get started with (in any programming
> language I'm familiar with). XML would work, but with a lot of overhead
> which is not required and does not add anything. The suggested format
> could always be really easily re-packaged into XML by a simple program
> if there was a reason to do it.
>
> I've been reading my antique AI book during my morning stationary bike
> rides and they've been talking about O-A-V object/attribute/value
> constructs and rules. They've also been describing forward and backward
> chaining approaches to problem solving. It's all interesting, but I
> still don't see how to make a computer do that kind of thing in a
> general way.
>
> I keep hoping the reading or someone here will give me a place to start.
> I really think computers are to a point where they should be able to
> start helping with a lot more of the information overload than they
> currently do.
>
> I guess if it was trivial, we'd all have intelligent assistants instead
> of glorified typewriters.
>
> David Garner
>
>
Other related posts:
