Coders' Wishlists for 2006

[Astronouth7303]

Seeing as it's getting closer to the kick-off, and everyone's gearing up for the next season, I'd thought I'd ask what everyone would like to see change as far as programming goes.

Here's mine:

• More memory
• Trig
• Use Eclipse right out of the box
• Open source dashboard, loader, & dev reader

[Andrew Blair]

Hey, even as a an amatuer programmer, more speed! It would be nice to be able to to more intensive math without crashing the control code. And maybe hardware enabled floating point.

[scitobor 617]

Only thing I want is a longer autonomous period and more sensors in the KOP. I like the current controllers though it would be nice if the compiler worked natively on Linux.

[CJO]

I really like the open source dashboard.

One can always use small, cheap Parallax or Microchip controllers for sensor input/computation, and not worry about using the main RC for that.

[Mike]

Originally Posted by Astronouth7303

• Trig

Version 3.9, I believe, of the mcc18 compiler has that, just include <math.h>


EDIT: I'm going to wish for more program space. I have a feeling the project I'm working on is going to take up a good amount between autonomous code, user-driven code, serial communication code, eeprom code, etc.

[Rickertsen2]

Its been a while since have though about FRC code but:

I doubt they are going to make any hardware changes but i am going to wish for a 16bit PIC anyway.

Access to at least 1 whole set of GPIO pins would be great


In addition, i think the base code could use some radical changes in the way of abstraction and modulalarity. We probably rewrote 60% of it last year.

More memory. FLASH, RAM. There are a number of pics that would be almost drop in replacements that would meet this criteria.

[Kevin Watson]

Originally Posted by Astronouth7303

Seeing as it's getting closer to the kick-off, and everyone's gearing up for the next season, I'd thought I'd ask what everyone would like to see change as far as programming goes.

Here's mine:

• More memory
• Trig
• Use Eclipse right out of the box
• Open source dashboard, loader, & dev reader

What kind of trigonometric functions are you looking for?

-Kevin

[Alan Anderson]

I'd like to see more of an event-driven framework, or at least something that doesn't look like autonomous mode is an afterthought.

And I'd really like to see the limit_mix() function die.

[Matt Leese]

Originally Posted by Kevin Watson

What kind of trigonometric functions are you looking for?

-Kevin

Maybe there can be an implementation of a fixed-point set of basic trig functions that can be used. I know there are fixed-point approximations for some trig fucntions that work fairly well and don't require a large lookup table.

Matt

[Redneck]

No more MPLAB.
Give us support for something else.... anything else... just no more MPLAB.

[Kevin Watson]

Originally Posted by Matt Leese

Maybe there can be an implementation of a fixed-point set of basic trig functions that can be used. I know there are fixed-point approximations for some trig fucntions that work fairly well and don't require a large lookup table.

Matt

I've written some code that that will do fixed-point sin() and cos() using an EEPROM look-up table. One piece of code is executed once to create the table, another body of code is used to do fast look-up. Once I write the documentation, I'll release to code for folks to try out.

-Kevin

[jazzsaxman1988]

Originally Posted by Redneck

No more MPLAB.
Give us support for something else.... anything else... just no more MPLAB.

Amen... MPLAB should die. It'd be nice if we could set up a project in Visual Studio to handle it

[Ryan M.]

Well, there is support for other things thanks to the work of others. Search for "using eclipse."

What I would like to see:

• Much more organized default code. Sure, you can take the stuff they give you and fix it, but then when they change a few things next year you have to start all over. One easy thing which would be especially nice for newer programmers would be just move all the utility include files into a separate directory. (Leave things that people commonly want to edit in the main area.)
• Stripped down version of the default code. A version of the default code which contained only the essential, basic structure would be handy for the more advanced teams. Continue to have a version with examples, but also make one which gets rid of things like the default mappings of the joysticks to the motors.
• USB programming capabilities. Yeah, I know that USB to serial converts usually work, but driver for the one I got caused horrible stability issues on my Windows box (not to mention took 10 minutes to download...) and it doesn't work at all with my Linux laptop. (As a side note: if anyone knows of one that works with Linux, tell me please!) I realize that the hardware isn't likely to change for a few more years, but is a wishlist.
• Going off of that, native support for Linux would be nice. Not a big priority thanks the work of many people on this forum, but what if the next version of the compiler breaks all their work?

[Matt Leese]

Originally Posted by Kevin Watson

I've written some code that that will do fixed-point sin() and cos() using an EEPROM look-up table. One piece of code is executed once to create the table, another body of code is used to do fast look-up. Once I write the documentation, I'll release to code for folks to try out.

-Kevin

I had some similiar code last year but it got lost in a laptop crash. There's also a pretty good atan2 approximation (or at least it seemed to be) out there if you'd want to include that too. This is what I based my implementation on last year: http://www.dspguru.com/comp.dsp/tricks/alg/fxdatan2.htm

Matt

[JJG13]

Originally Posted by Redneck

No more MPLAB.
Give us support for something else.... anything else... just no more MPLAB.

Couldn't agree more.

I'd also like a USB programming port for the FRC. Not many laptops have serial ports anymore and in any case USB would be much faster. Oh yeah, we need to have the IFI loader have a large (infinate) buffer for the printfs.

I don't really care about the code FIRST gives us as producing code is my job. Other than that I want more code/variable space so that we can have better, more overly complicated autonomous modes.