[ExI] mit's answer to the stanford ai class

spike spike66 at att.net
Thu Dec 22 15:47:54 UTC 2011


>... On Behalf Of Kelly Anderson
...

>...There is a serious point here though, and that is that even very serious
computer scientists find it EXTREMELY difficult to produce anything like
good code using Excel. It just isn't a good tool for producing serious
software. There is no testing harness or framework, for one thing... there
is no control of execution, nor really an explanation of how it works at any
level of detail... Excel is a disaster from a software engineering
perspective...That being said, it is the most popular programming
environment in the history of the world. So it go the user interface right,
but didn't support the "real" programmers of the world...-Kelly


Ja.  Excel is really good for what excel is designed to do, like a golf
cart.  You can't convert it to a race car, it isn't a good highway cruiser.
But it has its uses, and in that narrow scope it is better than a race car
or a Lincoln Towncar.  Yes we know there are people who have no drivers'
license, who can operate a golf cart and get the job done with it, even if
not as well as the alternatives.

Excel interfaces well with the mind, or rather some minds.  I hope we
eventually figure out a way to create something analogous to a
spreadsheet/macro programming environment with some kind of software
meta-tool that somehow reads one's spreadsheet and macro code, then figures
out what the silly prole wanted to do, then generates the code to do it.

spike




More information about the extropy-chat mailing list