On Starting a New Project

Published · 3min

I was talking to Peter last night about Laszlo. There are these times you need confirmation of why you have difficulty going off and either learning or writing something. I’ve a few projects like this, and something he said, quite off the cuff mind you, was a trigger.

He said, though I paraphrase, that to learn something you need some reason to learn in, and in the case of programming, that means a project that’s well suited to what you’re trying to learn. If you don’t have that right kind of project, you won’t learn the language or the technique.

Quite apart from Laszlo, I’ve been looking for some excuse to build something useful in ML, be it SML or O’Caml. But nothing’s came up. I’ve had a similar problem with Python: none of the projects I’m developing right now seem well-suited to it, and those that might be I’m doing in other languages because they offer some end user invisibility. That’s why I want to use Lua for Expresso (what I intend on calling MochaServ from now on), and for my eventual rewrite of Mocha (I’m installing Visual C++ and wxWidgets this weekend). Anyway, it’s a shame to waste such an nice icon.

Ah, but what to write in ML. O’Caml’s compiler generates code as good as, if not better than most C++ compilers with optimisations all on. But there’s no way to tie the likes of Python or Lua into it, nor are there any wrapper libraries for the likes of wxWidgets. So despite it being a better language than C++, and one I really want to build something useful in, I can’t, or at least not without even greater difficulty than if I was lazy and wrote the core in C++. And there’s no sense in knowing a language in the abstract, theoretical sense, if you can’t build something in it that’s practical and useful.

And then there’s Python. I know Python, and know it well, but it’s about a year since I’ve written anything more than a quick hack in it. I don’t feel I write enough in the language. Maybe I should dig around for an SDL wrapper and write a game in it, maybe convert my old RISC OS game, Reaxion, to it. But do I’ve the time, or will I just start putting it off like I tend to do?

BTW, I’ve a project that could benefit from being written with Laszlo, both in terms of implementation and learning opportunities. I’m just wondering if its dependancy on Flash might be a disadvantage to its saleability…