Subject:
|
Banging on the tasker
|
Newsgroups:
|
lugnet.robotics.rcx.pbforth
|
Date:
|
Wed, 3 May 2000 19:36:08 GMT
|
Reply-To:
|
SJM@JUDGEMENTspamless.COM
|
Viewed:
|
1629 times
|
| |
|
|
> Please keep banging on the tasker. It's a neat feature.
That's my plan.
The nice thing about the tasker is that people
can write code in a format they are more used
to. The same things can be done with event
driven single task code but it inverts the
control flow. It is more meaningful for most
people to write a PAUSE in a while loop than
to call a word that must return each invocation
which then gets called from some master top level
loop. The result is the same but tasks tend to be
easier to read for proceedurally oriented people.
State machines are an acquired taste.
As far as IR buffering is concerned that is only
important if you expect programs to communicate
while they are running. Once the robot starts
moving you don't want to block unnecessarily.
I don't mind polling in a task as long as it
is a poll and doesn't block. Before buffering I
would implement a ready_to_transmit word that
could be used in a well behaved task to output
one byte at a time.
Buffering would be nice, especially if it was
handled at interrupt level but I'm sure you
have limited time to invest in this and need to
set priorities. I personally would put things
like a Tcl/Tk GUI ahead of serial buffering.
This would reduce the barrier to entry and might
get more people involved. More users obviously
means more potential contributors.
Besides if I got desperate enough for serial
interrupts I might implement them myself. I've
been itching to get my hands on low level RCX
code anyway.
But not yet. First I have to make my current
project work. :-> I've rebuilt the mechanicals
many times. It's time to get the software to
do its part.
By the way I haven't said it explicitly but
thanks for this package. Thanks to you I'm having
lots of fun. This time around I might learn the
dictionary operators and become a REAL forth
programmer.
|
|
Message is in Reply To:
 | | RE: Multitasking questions
|
| (...) Ummm, I'm in the process of verifying all of the old scripts I package with pbForth. It looks like HAT is what actually works, and I've managed to change the documentation and not the code. HAT is in the origianl hForth implementation but I (...) (25 years ago, 3-May-00, to lugnet.robotics.rcx.pbforth)
|
6 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
