From: James Higgins (jameshiggins@earthlink.net)
Date: Fri Jul 20 2001 - 20:16:02 MDT
I think the amount of work required to implement the programmer's interface
is being heavily underestimated. Basically, I agree with Ben on the point
that the XML itself will be incomprehensible to programmers directly. Thus
the need for Flare Speak. However, since the XML representation of Flare
is eXtensible, how does this map into Flare Speak? You are going to have
all sorts of unrecognized annotations, extensions, data types, expressions,
etc. that the standard Flare Speak IDE doesn't recognize. If it doesn't
recognize it, how does it render it for the programmer? Further, it seems
to me that attempting to render all of this extra stuff will simply crowd
the Flare Speak, quite possibly making it nearly incomprehensible to the
programer as well.
The only way around this is to have one hell of an IDE that is incredibly
dynamic and extensible itself. But then you have ONE IDE. If someone goes
and writes another IDE (say a visual one as has been suggested many times),
then you very well may have conflicts between them. How does one describe
what every extension to the XML for Flare means, how to render it (in all
possible formats, text, visual, etc), its significance (when should the
programmer see it if it is an annotation), etc?
I think the amount of effort to develop Flare & Flare Speak may have been
greatly under estimated. Thus I also agree with Ben in that all this
effort would almost certainly be better used if focused on actually
creating a Real AI.
James Higgins
At 05:13 PM 7/20/2001 -0500, you wrote:
>----- Original Message -----
>From: "Ben Goertzel" <ben@webmind.com>
>To: <sl4@sysopmind.com>
>Sent: Friday, July 20, 2001 4:39 PM
>Subject: RE: Languages and AI
>
>
> > There WILL be a "real AI programming language" one day. It won't be LISP
> > or Prolog or Java or Flare or even Haskell, it will be something that we
>can
> > only design after we know for sure how to make a real AI...
>
>This is why a programming language needs to be _extensible_. That's the "X"
>in "XML", eXtensible (Markup Language).
>
>I have no experience with Smalltalk, but I think I've heard that it is
>extensible. In a previous post I ranted about the Pliant language, which
>looking at the design docs is also extensible.
>
>I haven't read up anything about annotated languages yet, but if Flare is
>actually extensible, then it may be a preferred human-programmer high level
>language than non-extensible languages such as Prolog, Java, etc. I think
>Flare is trying to be extensible by inheriting XML as its syntactic format.
>
>
>cheers,
> Simon
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:37 MDT