It’s always been my intention to develop the Loci programming language as a tool that I would regularly deploy myself and this year that’s going to start happening!
The language and compiler have reached an early stage of maturity, with a substantial mix of implemented language features and a small (but quickly growing) set of useful standard library modules. Hence this seems the right time to move some of my projects, currently written in C++, over to Loci, both to take advantage of the strengths of the language, to guide future design decisions of the language and as a real-world test of the compiler. Chain Reversi has already been ported to Loci and has been particularly useful when optimising the compiler; I consider quick build times extremely important and I’m pleased to say both that the design of Loci means it can be expected to compile much faster than similar C++ code and that this is borne out in practice.
The first project to move over is my ‘second’ (but slightly neglected) project, OpenP2P, which is basically a set of libraries that make it easy to develop applications that use peer-to-peer networking. One of the aspects that held OpenP2P back was the lack of good event notification/handling in C++ (which was basically implemented as a layer over boost::asio) but fortunately Loci has std.event, which in my opinion follows a very clean approach to event signalling. So I’ll be moving OpenP2P across and integrating it neatly with the standard library modules.
There’s another reason for making this change. Loci has presented me with a couple of problems:
- It’s too much fun to work on!
- It’s too successful!
Hence it’s starting to dominate my projects time, which has led to it becoming relatively well structured but on the other hand my other projects have been neglected. By moving these projects over to Loci I get to work on both at the same time! I also get the thrilling feeling of the last 5 years of designing Loci and building the compiler coming to fruition in a very practical way.