Re: [Phys-L] programming languages & controllers

[John Denker <jsd@av8n.com>]

On 03/17/2014 12:53 PM, Paul Lulai wrote:
> Does anyone have a recommendation for a language that might be
> useable at the h.s. level? We are currently using RobotC.  We are
> thinking about switching to something that is more scaleable.
> LabVIEW might be a candidate since the Lego NXTs are programmed in a
> version of labVIEW.

Wow, that's a good question.  

1) Beware that a good implementation of a mediocre
 language beats a bad implementation of a great language.
 So a lot of things that would make sense in the long run
 in theory don't necessarily work today in practice.


2) You actually have /two/ "market segments" that you need
 to satisfy:  The students and the teachers.  Do not assume
 that what works best for the students works best for the
 teachers.  Students are important ... but teachers are
 important too!

 In particular, students are happy to learn the latest most
 modern thing, because why not, they've got to learn something
 and the new thing is just as easy to learn as anything else.
 Meanwhile, the teachers have an investment in the old way
 of doing things.  It's not that they are dumb or stubborn
 or lazy, it's just that they don't have the time to learn
 something complicated and new every year.

 As a conspicuous example of what I'm talking about, young
 students have no trouble with highly parallel event-driven
 approaches.  It's just like a sports team, with a bunch
 of agents running around in parallel, mostly independent
 but loosely coordinated.  In contrast, somebody who learned
 "scientific" programming 20 years ago is going to run away 
 screaming, because he expects the computer to do one thing 
 at a time in a nice linear sequence.

 The closest I can come to a constructive suggestion here
 is
  a) Plan a gradual phase-in of new stuff, so that teachers
   don't have to deal with too much change all at once.
  b) Prepare /two/ sets of documentation, one with what the
   students need to know, and one with what the teachers
   need to know, which is generally wildly different.


3) Other things being equal, students are vastly happier 
 with drag-and-drop graphical interfaces such as you see 
 with gamemaker / labview / lego-robots ... in contrast to
 imperative programming languages such as C.


4) Other things being equal, work in the highest-level
 language available.  Young students can deal with the
 abstractions involved.  It's kinda cute to see one little
 kids explaining to another that you need to define a class 
 over there and then instantiate it five times over here.
 The grown-ups roll their eyes because they have no idea
 what's going on, but meanwhile the kids are turning out
 robots and games that actually work.

 To say the same thing the other way:  Do not get sucked
 into using low-level languages like javascript.  That is
 not a good use of anybody's time.

 If you learn a structured language first, you can then use
 a low-level language safely, because you will self-impose
 some discipline and structure.  The converse does not work.


5) The choice of language is more important than the choice
 of hardware platform.  Unless you're planning to make a
 million of them (e.g. washing machines) the cost of the
 software far outweighs the cost of the hardware.  Here
 the "cost" includes teacher time and student time learning
 the language and writing the code.  So pick the platform 
 that supports the best software.