Getting Started with Scripting

[Mike Betts]

LBK,

I would love to give you a definitive answer be I'm stuck in the same quandary.

There are a few sources of good code, most notably IFI and Kevin's websites. You will have to decide for yourself.

My quandary is if my team should include the DDT (Dynamic Debugger Tool) support in the latest IFI default code into our system this year (it limits some other options).

And I am most definitely undecided if this scripting is such a good idea. C is the most popular embedded language because it allows very precise control of your hardware while remaining readable (vice assembler).

Higher language constructs, which were developed with appropriate disciplines (EBNF, et cetera), can remain useful for embedded applications. However, one starts to lose control of the hardware.

At a high level of abstraction, you lose all touch and feel with the hardware.

Where am I going with this digression? You may be comfortable with a desktop application locking up or giving you a Blue Screen of Death but would be at your lawyer's office in an instant if the airbag micro controller in your car malfunctioned.

The scripted language presented thus far seems to be very implementation specific and may difficult to maintain in the future. I'm not sure if this is the direction we should be following.

It is late and I am rambling... Excuse me please...

Bottom line, your team is responsible for the code in your robot and my team is responsible for ours. We cannot make anyone a scapegoat if our code does not function at the competition. Each of us may reach different conclusions.

Once again, sorry for not having an answer for you... Download it all and start dissecting...

Take what you need and leave the rest...

[Tom Bottiglieri]

Mike, I agree.

If the scripting was brought as far as a standalone GUI that would allow users to add function blocks or draw waypoints, then automatically write the code, it may be good.

But for this year, I think I am going to salvage the encoder loops, gyro code, and the PID from the FIRST supplied navigation code, and try to work it into the default code. If you havent noticed, the navigation codes' workspace is different from that of the usual default code, using robot_command() instead of default_routine(). I think it would be easier on my rookies to try and code in the default code, of which they know, rather than figuring out how the scripting works. It just seems like a waste of time with results that are very limited.

This is unless, of course, FIRST releases a new version of the code sometime soon.