China Makes A Java Version Of Core LLVM

[tammy]

Never tell a developer that their pet project is a waste of talent. If it wasn't valuable to them, they wouldn't be doing it. Many useful things have come from pet projects.

Said well!

[JanC]

Really, the main reason why C/C++ isn't being taught anymore is because is HARD. Its not a fun language to teach the basics in. [I actually recommend people learn an easier syntax, like Pascal or even Algol 68 first].

C isn't too difficult to learn because, despite some inconsistencies in its syntax that make working with pointers more complicated than necessary, it is a fairly simple and small language. C++ on the other hand has so many inconsistencies and is so big and complicated that there are maybe one or two people in the world who really know it (and C++'s inventor has professed he is not one of them).

So I agree that teaching students how to work with pointers can better be done with a modern (Object) Pascal implementation (I don't know Algol 68, so difficult to say anything about it). For teaching basic programming, Python or another language with a simple syntax and no boilerplate overhead (so no Java or C#) would be good too.

[allison]

Never tell a developer that their pet project is a waste of talent. If it wasn't valuable to them, they wouldn't be doing it. Many useful things have come from pet projects.

Yes, I agree with you

[gamerk2]

I can see the exercise in rewriting the code in another language to be an acceptable very beginner-ish exercise. The end product will be of absolutely no value though.

The way I read this, is that, at least when I went through it, comp sci degrees tend to start off with Java. Hence java becomes the "most familiar" language for a lot of students. That doesn't make it actually a good language, or even remotely suitable for high performance computing needs.

What this really is, is a BEGINNERS C++ course with a major project being to *read and understand* the C++ code through translating it into Java.

Might be a fun learning exercise.
This is not news.

Yeah, thats what I'm seeing.

Really, the main reason why C/C++ isn't being taught anymore is because is HARD. Its not a fun language to teach the basics in. [I actually recommend people learn an easier syntax, like Pascal or even Algol 68 first].

[aironeous]

They actually made something? Oh it's younglings - ahhhh . Soon they will grow up and be stealing our technology, counterfeiting our goods, pirating or music, movies and software and stealing our trade.

[CARMEN]

It's not about not understanding LLVM in C++, it's about using it in Java projects, which is more difficult if it's C++ code.

Yes, I agree

[Stephen_A]

Oh. Why don't they just use Scala?

Yeah, Scala is ok enough

[droidhacker]

I can see the exercise in rewriting the code in another language to be an acceptable very beginner-ish exercise. The end product will be of absolutely no value though.

The way I read this, is that, at least when I went through it, comp sci degrees tend to start off with Java. Hence java becomes the "most familiar" language for a lot of students. That doesn't make it actually a good language, or even remotely suitable for high performance computing needs.

What this really is, is a BEGINNERS C++ course with a major project being to *read and understand* the C++ code through translating it into Java.

Might be a fun learning exercise.
This is not news.

[randomizer]

I understand students and learning but it just seems like a total waste of good talent.

Never tell a developer that their pet project is a waste of talent. If it wasn't valuable to them, they wouldn't be doing it. Many useful things have come from pet projects.

[pingufunkybeat]

I don't think so.

The number of students who effectively work on software of this complexity is very small. Marek & co are exceptions, really.

This way we got a bunch of students who really understand LLVM. Most other students are writing Angry Birds clones, or dreaming about the next big app that stores your aunt's recipes on the cloud, or some similar nonsense.

I don't know if these guys will end up contributing to LLVM or some similar compiler project in the future, but I'm pretty sure that the "Angry Birds clone" guys won't.