Quote:
|
Originally Posted by sanddrag
As is having to ship the RC, when we can legally buy another one and program on that.
|
Or change out a few libraries and just program on the older (minus the PBASICs) controllers.
So now my views on R17, from a programmers perspective. With all due respect, I don't think that the people (person?) who wrote R17 have ever had to program during the six weeks. I know the FIRST standard for programmers is that a couple days is lucky. This year I was given about two full days and a couple hours worth of "Hey, mechanical team has to change this piece. You can have twenty minutes with the bot." This is definetly
not enough time to program anything more than a medium complexity robot.
I will bet money that if you ask the majority of programmers what they want during the six weeks, the answer will not be a new compiler. Nor will it be a new language, or a different IDE or a faster processor. It will be either more time with the bot or a second bot for testing. The former is not feasible on most teams (six weeks is already a strain) and the latter is not feasible due to financial complications. The only options that a programmer has is write all his code during the first five and a half weeks and testing at the last couple of days. This leads into my next thought...
Programming is
not merely writing code. To successfully program a FRC bot, you need to design the code framework, read up on the datasheets of the sensors you are using, write the actual code, debug the actual code (syntax and whatnot), load the program and finally... TEST. Testing is by far the most time consuming thing out of the entire process. Lets take a PID loop, for example. Someone who is familiar with PID can write the actual code in under half an hour. However, after it has been loaded onto the robot it will most certainly take them at least half an hour of tuning. Testing is not possible after the six weeks (lack of a robot, unless you can afford to build a second) so I don't think writing raw code should count towards "software development." If a programmer knows how to write a certain function, they should not have to take time out of their testing time at competitions to write it.
So if you've been following all that, congratulations. It's probably a tad hard to understand, I tend to jump from topic to topic. What I've said basically amounts to...
- I don't believe that R17 should apply to programming. There simply is not enough time given to programmers.
- The most beneficial thing to a programmer would be more time with the robot or a second robot purely for programming. The first option is hard due to time constraints, and the second due to financial constraints.
- Writing raw code is only a part of the software development process. Testing code is much harder on the complex systems that most programmers wish to design.
- Programmers shouldn't need to take precious testing time to write code.