It’s been two and a half years since I started work on Plasma. With the end of 2017 I noticed that many other language projects were publishing retrospectives of their last year. Well with one developer progress is a lot slower, so rather than summarise the last year, I’ll write about the whole history!
I can’t talk or think about progress without thinking about my expectations. It’s fair to say that I did expect to get further with this project, but I don’t think that I underestimated the amount of effort required or my limited time. I knew those well going in, nevertheless I did expect to get further in this time. The roadmap is organised by milestones, each milestone is made up of multiple goals. I hoped, but now that I have retrospect, to achieve about one milestone every six months, at least for the first few 2–3 milestones.
Progress & Psychology
Initially progress seemed rapid, I was excited and had some close friends excited about the project. However in 2.5 years I am over half-way through the second milestone, less than I expected.
I published the roadmap to provide a sense of what my goals are to anyone reading along or wishing to get involved. It also helps me maintain focus on what I want to accomplish now and later (and I’ve recently been experimenting with a kanban board and github issues). However it has a side effect of making it clear what I have (and haven’t yet) achieved.
While that side affect can sometimes be unfortunate, I think I’d feel like I had less progress if I did not see the roadmap at all, I would have no lovely green ticks to see at all! While I don’t want to dwell on this — since I don’t owe anyone Plasma, there’s no "deadline" — a feeling of progress (and peer support) is just something I seem to need sometimes. Therefore I try to accept that I set myself lofty goals and when I evaluate my progress I try to do so against more moderate goals.
If you think this all sounds quiet difficult to live with then I’d like to point out that I wouldn’t have started Plasma if I was already satisfied with my career progress and contributions in the last few years.
Where is Plasma at today?
a token-based bytecode interpreter, it’ll be replaced by better technology later;
enough builtin types and functions for testing;
a compiler that generates .pz bytecode format files;
basic functions, expressions, statements, if-then-else, pattern matching;
a type system with ADTs, higher-order values/calls and generics;
the basics of the resource system;
many features in a planning state!
Without calculating it, (although the git logs are available) it seems like I’m closing off one goal every two or so months. That’s actually not bad for something I do in my spare time.
Also quite a few of the recent goals: higher-order code, generics, resources; allow Plasma to become more expressive. The language itself is taking shape. Now more than ever I can feel the implementation getting closer to my vision. Each time I complete a goal I feel excited about the expressibility that the next goal will bring. It won’t be long (relatively) and Plasma will be ready for some testing.
Next I’ll be adding some more test cases for higher order calls, support for closures, lambdas etc. Then it’s time to implement the module system and make design decisions about the Prelude/stdlib. At that point milestone #2 will be almost complete and its time to encourage some testing.