Crosscut

Let's move on to the last in the previous prototype's list of problems: that it targets the browser as the primary platform, which causes a lot of complexity. I've already alluded to this earlier.

I made this decision, because a former prototype had achieved the capability to support actual games. I had written one (a Snake), and wanted the ability to publish this game (and future ones) in a way that would allow people to easily play it. Because if I can't actually publish stuff, nothing I'm building is of any use.

But then I went on to not use that capability (even the Snake game went unpublished for a good while), because I was busy with figuring out the language design, compiler implementation, and other technical aspects. And despite not being used, the browser-centric architecture introduced a lot of complexity that translated into an ongoing burden.