Subject:
|
Re: Mindstorms (more than just robotics?)
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Sun, 8 Nov 1998 11:35:11 GMT
|
Original-From:
|
Geoff Seel <geoffs@gseel.demon.co.uk/stopspam/>
|
Viewed:
|
2928 times
|
| |
|
|
Several replies in one go to avoid multiple messages
In message <jCKx2LBVjJR2EwcJ@phesk.demon.co.uk>, Peter Hesketh
<pbh@phesk.demon.co.uk> writes
> In article <RP0jfKAw2BR2Ew7K@gseel.demon.co.uk>, Geoff Seel
> <geoffs@gseel.demon.co.uk> writes
> > I personally like the programming, but, there is a MISTAKE! The MISTAKE
> > is, when it has a "stack" it does this: When it starts a new stack, it
> > does not carry on with the old one.
>
> Is that while building the stacks on screen, or when executing them in
> the 'bot? In other words, are you saying the 'bot won't multitask?
It appears that if a sensor gets reactivated before its stack has
completed the current execution of the stack is just abandoned and a new
one started. This can leave the system in an 'undefined' state. (N.B. I
see Kekoa Proudfoot suggests that this only happens when both states of
a sensor have stacks and switching from one to the other occurs)
This is a pain to an experienced programmer, to a child trying to sort
out a program it can be even more discouraging.
---------------------------------------------------------
In message <364472D4.EE37AEE8@acm.org>, stephen p spackman
<stephen@acm.org> writes
<some snipping>
> Geoff Seel wrote:
>
> > The handling of stacks and the lack of re-entrancy make it a difficult
> > environment for teaching programming in a logical manner (hard to debug
> > multi-sensor systems sensibly).
>
> ...so you think it's easier to let people walk right into
> synchronisation bugs and learn the hard way?
No, I think people should have room to learn programming and debugging
with basic issues before being presented with systems that display
unpredictable behaviour as Mindstorms can do in some circumstances.
>
> > We have done some work with Logo since
> > he was about 8. It was fine as a way of learning to manipulate a turtle
> > but the handling of variables is horribly confusing (definition in a
> > different format to use).
>
> Well, defintion can't look the *same* as use or you can't assign a scope
> to a variable; making a definition and changing it are inherently
> different ideas. So how much "different" is a problem?
My point was potentially trivial and maybe only applies to the LOGO we
were using but when I define a variable it needs a " before it, when I
use it sometimes it does sometimes it doesn't so there are statements
like:
MAKE "QUESTIONS SUM :QUESTIONS 1
In my opinion that is a lot less clear than
QUESTIONS = QUESTIONS + 1
or
MAKE QUESTIONS SUM :QUESTIONS 1
or perhaps even
QUESTIONS++
--------------------------------------------
In message <199811071909.LAA07458@pixel.Stanford.EDU>, Kekoa Proudfoot
<kekoa@Graphics.Stanford.EDU> writes
<large amount snipped>
> Certainly, the process of figuring out how tasks really work and why is
> valuable, assuming that you don't simply give up because things don't work
> as you expect them to.
I totally agree with this, what I find worrying is that the problems
with stacks can make the behaviour unpredictable. In the early stages of
learning programming, and the fact that programs need debugging,
repeatable behaviour is important.
If your first program sometimes printed "Hello W" and sometimes "Hell"
but always "Hello World" when debugging (because debugging changed the
timing) that would be discouraging.
--
Geoff Seel Voice: +44 (0)181-390-0109
Financial Object Resources Ltd. Fax: +44 (0)870-054-7931
54, Berrylands Road, e-mail: geoffs@gseel.demon.co.uk
Surbiton, WWW: www.gseel.demon.co.uk
Surrey KT5 8PD
United Kingdom
|
|
Message is in Reply To:
 | | Re: Mindstorms (more than just robotics?)
|
| In article <RP0jfKAw2BR2Ew7K@gs...on.co.uk>, Geoff Seel <geoffs@gseel.demon.co.uk> writes (...) Is that while building the stacks on screen, or when executing them in the 'bot? In other words, are you saying the 'bot won't multitask? (26 years ago, 7-Nov-98, to lugnet.robotics)
|
10 Messages in This Thread:
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
