PROGRAMMERS: WIND RIVER C++ vs LABVIEW vs JAVA

[The Lucas]

Originally Posted by davidthefat

Personally I wouldn't call LabView coding at all...

When I saw this thread I thought "Over/Under 5 posts before some throws out this particular flame" Looks like it a push and I was right on. This argument is getting old and predictable.

My team uses C++ because that is what works best for us. All 3 languages are coding. I suggest everyone try to learn all 3 of them. Trying to argue that any one language is better for programming FIRST robots is a meaningless pursuit, it is all based on personal opinion. They certainly have their own pros and cons (which are fine to discuss) but all are equally capable. Try to focus on what you would tell a rookie team trying to choose between the 3.

[davidthefat]

Originally Posted by The Lucas

When I saw this thread I thought "Over/Under 5 posts before some throws out this particular flame" Looks like it a push and I was right on. This argument is getting old and predictable.

My team uses C++ because that is what works best for us. All 3 languages are coding. I suggest everyone try to learn all 3 of them. Trying to argue that any one language is better for programming FIRST robots is a meaningless pursuit, it is all based on personal opinion. They certainly have their own pros and cons (which are fine to discuss) but all are equally capable. Try to focus on what you would tell a rookie team trying to choose between the 3.

I personally will tell the said rookie team to pick up a C++ book and read it and study and meditate on it, IMHO its way more useful in the real world out of the 3. If you learn C++, you pretty much know Java (except alot of the java specific stuff), won't take you very long to pick up on java. Plus most real world applications are in C/C++, like Operating Systems, Games, and tons of stuff.

[virtuald]

Originally Posted by davidthefat

I personally will tell the said rookie team to pick up a C++ book and read it and study and meditate on it, IMHO its way more useful in the real world out of the 3. If you learn C++, you pretty much know Java (except alot of the java specific stuff), won't take you very long to pick up on java. Plus most real world applications are in C/C++, like Operating Systems, Games, and tons of stuff.

Right... the 'real' world... and how many years have you been working in the software industry now?

It just depends on where you end up in the 'real' world. Different languages are better for different tasks. There is a lot to be said for domain-specific languages (such as Matlab) and scripting languages (such as python). Just depends on what you're trying to accomplish and what you're used to.

[Egg 3141592654]

It really depends on the task. I myself know Labview and Java like the back of my hand, and I know the very basic things about C++ (not enough for a robot). I have chosen Labview the past two years only because it is much easier to debug and understand core concepts for beginners. Next year, our team is moving onto Java, only because, as a long time text coder, Labview was very confusing to trace.

In terms of performance, our Labview takes FOREVER to deploy and build, but there is no noticeable lag difference between the two when they are in the robot and running. C++ could potentially be a bit more frugal with memory to decrease junk code, but that depends who is sitting behind the computer monitor. I say it depends on the task to determine which is the stronger language to code in.

[mathking]

I will offer another plea to be open minded about languages when you make decisions, and listen to your mentors, many of whom have a significant amount of experience in programming.

As for Lab View not being real programming, put that right out of your mind. (I have always been impressed with LabView, and the first time I saw was more than 20 years ago.) I have been teaching CS for a while now, and I know that one of the constant "discussions" in computer science education circles is whether to start students on Assembly programming or not. I count myself firmly in the "not" category here. Learning to program is about learning how to break a problem down into smaller problems, finding algorithms to solve those problems, translating those algorithms in instructions the computer can understand and then debugging. The last step takes an order of magnitude more time than all of the others. Too much detail at the start tends to make you worse, not better, at programming. For learning the ins and outs of programming a robot LabView is a fine choice. It is really good for modeling and controlling embedded systems. If you are of the "I would rather not use that particular front end, I would rather use the 'real stuff' do my own C programming" consider that someone has to write the C compiler (a task you will likely have to undertake if you go far enough in studying CS). So why are you looking down your nose at using something that "was probably coded in C anyway" when you are just using something that was ultimately coded in Assembly, which is not itself really how the computer works by the way.

As for language choice, what you are familiar with will tend to be what you favor. But one job of a mentor (and a main job of a teacher) is to help you grow intellectually. One good way of doing that is making you learn something new. Even better is writing code on a couple of platforms and evaluating to see which is better for what you need. This year we developed code for both LabView and Java. Our decision this year to ultimately choose Java over LabView was due to our desire to use CAN, which was much easier in Java than LabView. But this doesn't mean that Java is superior, just that it served our purposes better.

I first learned programming in Fortran, then in BASIC. My first job programming was in C. I would say that C/C++ is what I am most comfortable using for my own coding, although teaching AP Computer Science has made me pretty comfortable with Java. If I am writing a simple program for myself I will often write it in Java just because it is easier to make a simple GUI and get it working. But one thing I tell my students is that they should never make their decisions about what language is best while they are in high school based on their perceptions of the real world. Even if their perceptions are accurate, what they learn in high school will not be what they use for most of their lives. When I was in high school (early 80s) my programming teacher told us we really needed to learn COBOL if we wanted to ever be employed in the field. So I did, and have never really used it. She wasn't wrong about the field at that time, but it changed. There was a big upsurge in COBOL jobs for a while leading up to Y2K, but now they are going away fast.

Now that I am at the end of my "open yourself up to new experiences" sermon, my two cents about language choice. I find that Java has a lot to recommend for this kind of robot programming. The way Java handles events, threads and exceptions is straightforward and easy to implement. It makes for much more robust code than typical C++ code. As for Java being filled with "fluff," most of the so-called fluff is just a standard implementation of coding standards that you will be forced to use by an employer if you ever code for big projects in any language.

[synth3tk]

I don't know (or care) if it's "the best", but we used Java. Our three (3!!!) student programmers either left, or didn't return, and none of the freshmen had no desire to learn programming at all. So I was stuck with the role. While we had default/leftover code from previous years written in other languages, myself and the team mentor decided to use Java. Randomly.

I'll tell you what, though, I knew absolutely nothing of Java as of January 9th, 2010, beside the fact that it's a programming language. That first week of build, I started learning the syntax, rules, and studied the default code. By week 4, the robot drove, and the pneumatic kicker worked. Had I been given more than 2-3 days of exclusive access to the robot, I probably would have had a nice autonomous working.


Also, quick note, never count out other languages, ever. Sure, you may love the most popular language used, but knowing X or Y may cost you a job/project.