Autocoding project proposal.

Timothy Rue threeseas at
Fri Jan 25 17:31:53 CET 2002

On 24-Jan-02 23:22:35 Fernando Pérez <fperez at> wrote:
>I'm gonna regret posting here, I suspect ...
>> there is the stand alone version of one ofthe commands
>> IQ version 0.351 which does make use of some of the facets of the
>> other 8 commands.

>Ok, maybe it's just that we're all too dense here for your genius. So
>please bear with us simple minds. Nobody so far seems to be able to know
>exactly what you're after, so please show some summary of input and results
>of your code so we can understand what's going on.

You can try it out for yourself, instead of me trying to talk thru such a
thing. However, if you will read thru the other messages, you may find
what you are asking.

>Yes, I know you already gave a pointer to code, which is a good first step.
>It shows you care enough about this to have gotten it off the ground and
>that's great. Problem is: you talk about a lot of abstraction, yet the file
>you mention above is a *single 500 line chunk* without a single function
>definition for chrissakes! Is that your idea of abstraction? You've thrown
>in command line parsing with help messages with i/o with the core of
>whatever it is you're doing, in one big salad bowl. That code is utterly,
>absolutely incomprehensible to anyone but someone willing to spend quite
>some time to go through it in thorough detail. You don't define *one*
>single function, *one* single class. Nada. Zilch.

What is the main value of a function? I thought it was 1) to reduce
repetitive code and 2) to help format the source.

What would I need to define a class for, in that code?

Maybe you'd like to show me how that code can be broken into functions,
but for more than just the sake of breaking it down into functions.

Maybe just the loop section?

The top autocoder today has this as one of it's code generation features:

* No dynamic memory allocation, no unbound loops, no pointers, no
recursion, no dead code, no library calls.

*Readable and fully commented

Correct me if I'm wrong, but isn't a function call something of a pointer?
Of course this is the code that is generated by the SCADE autocoder but it
is still code. (point being is that the IQ code works and is commented
enough that it can be followed.)

The code is commented and if you run it with a file named "lnswitch.on"
existing (even and empty file - it only looks for the existance) then
it will give you alot of insight to what it's doing in the loop.
Expecially if you use it on a file that references other files that
referances each other.

See some of my other responses, maybe the last one to Kragen.

>So either you're going to seduce enough people with the results of that
>code that there will be interest in diving in and cleaning it up (very
>possible, I've done that in the past with messy code whose results were
>important enough to me) or you'll have to make that code digestible first.

>Sorry if this sounds harsh. I really hope your project (whatever it is)
>succeeds, if nothing else for your personal satisfaction. But I'm pointing
>to you some things which I suspect are a huge hurdle to its acceptance
>right now.

??????? Ok so you think it's a major hurdle this 537 line commented python

What about the command specifications, what could be the manual pages for
the different commands?

Have you seen those?

What seems to be the biggest problem right now is people are assuming to
much, rather than finding out weither or not their assumptions hold true.
Overall it seems that me being one person is having to take on such
assuming challenges by more in count, than the one I am.

Maybe instead of people telling me all these things that weight on the
side of no you cannot do it, people get on te positive side and start
considering all the reasons to do it and how.

Certainly reducing the need to re-invent will help generate the time
needed to do more and more complex and cool things.

*3 S.E.A.S - Virtual Interaction Configuration (VIC) - VISION OF VISIONS!*
   *~ ~ ~      Advancing How we Perceive and Use the Tool of Computers!*
Timothy Rue      What's *DONE* in all we do?  *AI PK OI IP OP SF IQ ID KE*
Email @ mailto:timrue at      >INPUT->(Processing)->OUTPUT>v
Web @  ^<--------<----9----<--------<

More information about the Python-list mailing list