More Thoughts

[John Hopkins]

Hi Giovanni,

I also learned about Cell from Hacker News and just wanted to stop by and say how much I admire this project.

There is a lot of really great ideas here, and I think you are on to something.  I have been researching DSLs for my own purposes, and Cell is the first I have seen that is really exciting.  You have opened my mind to the possibilities of the field, and I am somewhat of a forward thinker in general!

You might consider providing education services or writing a book or something on how you have built it - your experience would be valuable to people working in similar areas.  By doing so, you would draw attention to Cell. 

I also 2nd the javascript angle - my experience is that in the JS community people are more forward looking, and in the Java community they are not as much.  

Also, if I get to the point where I can understand what you have built I would be happy to help you, but I am not a Java guy :D

Thank you for your contributions!

[cell-lang]

Hi John,

sorry about this late reply, but I've just seen your message. As you may have noticed, this is not exactly a high-volume mailing list, and I check it only once in a while.

Thanks for the encouragement. It's always good to hear there's someone who thinks all the effort I've put into this project is not a complete waste of time.

I'm currently working on the Cell to C++ compiler/transpiler, which is already generating very fast code, easily outperforming object-oriented Java or C# code for data/state intensive code in most of the tests I've done so far. After I release it (hopefully in a few months) I plan to do exactly what you suggested: to write not a book but a series of posts to do a deep dive into the implementation, and explain how it compares in terms of performance to OOP.

I hope that might generate some interest around the project just as you said, but I also hope to receive feedback that will help me improve the implementation. And after that I dream of getting a chance to present the project at some programming language conference.

As for the JavaScript code generator, I don't think I'll be able to release one anytime soon. Building and, even more so, maintaining multiple code generators is a huge hassle, and it gets pretty boring after a while. Even backporting the improvements I made in the C++ code generator to the Java and C# ones would be very time-consuming. Since my time and energies are limited, from now on I plan to focus on exploring new design ideas instead. Once the C++ code generator is finished, I'll start working on a new declarative sub-language instead, to complement the functional subset of the language. If what I've in mind pans out, I'm sure it'll be by far the most interesting and innovative part of the language, and it will make for much more satisfying work that building another code generator for yet another target language.

It would be great to have some help, if the project interests you and you find the time to contribute to it in the future. And by the way, I'm not a Java guy either, and I definitely don't want people to think that this is just a JVM language. Going forward the C/C++ code generator will be the "reference" one.

Also remember that there are many ways to contribute. If you decide to try out the language in the future, even if it's just to play with it, any feedback would be very appreciated. I'm particularly interested in hearing about "pain points", specific use cases where the language/paradigm runs into difficulties, because that's usually where new ideas originate.

Regards,