On 'Building Server-Side Web Language Processors'

I think this article misses mentioning exacly why running programs on the web is exciting: in theory any computer that can browse the web will be able to run the program without any platform-specific processing on our end, because the web is the platform. I think running programs in the terminal would require work on platform-specific distribution on the side of the programmer or a compilation step on the side of the user of the program (I guess it depends on the terminal). I think anyone who believes that the web will keep on growing as a computing platform would agree that it is a good idea to give CS students the chance to build projects for the web.

I really like what the Hydra compiler project is aiming for: compiling into web assembly. I think that depending on wasm's broader success (or lack there of) it might turn out to represent a very actual project, because if wasm succeeds then I think that lots of other projects will want to target wasm also.
In my view, the ultimate  'for the web' class project would be something that extended what it is possible to do on the web. Given our compilers (or translators) approach I think that such a project would be maybe making a translator that would allow building any language (or any app) into wasm. I could also think of projects that aided the research effort on wasm performance and optimization techniques such as evaluating the applicability of known techniques for wasm.