mattydoincode

came here to write this comment, +1!

Author here! I'm not arguing that software engineers should necessarily be paid in proportion to the end economic result of their employment. Although that's an interesting point for another article perhaps! Rather simply that companies tend to undervalue domain specific knowledge and forget how costly and time consuming it is to hire and retrain new team members. This should be taken into account when considering compensation structures to retain your top talent when considering the open market is always available to them.

Author here! My intention with "domain knowledge" point is mostly about all the non-coding stuff. Things like how your company functions, who to talk to about different issues, deep product knowledge and understanding the customers. I actually think transferrable skills grow as well, usually more on the technical side but also on the soft skills side in a generic way. I think the non-transferrable stuff is somewhat unavoidable because of the nature of products, domains, and organizations being meaningfully different.

I think that from the optic of someone who is not a physicist (like myself!) it's natural to think "hey... we can't actually know that it's dark matter vs some other unexplained way that gravity works". And in fact there are many physicists working on trying to figure that out, but there is actually a ton of evidence that what we're seeing is in fact particles, and most physicists consider it to be the most likely current theory. I don't think anyone in the physics community treats anything about dark matter as a given either (or hawking radation for that matter). In fact part of being a true scientist is keeping an open mind.

I actually did a school project that was very similar for a compilers class! It was also Hindley-Minler in Ocaml and we tried to basically stuff any random features of a programming language we could into one language, and just compile it to java (because you can do anything with java).

Here is the paper we did: http://www1.cs.columbia.edu/~sedwards/classes/2013/w4115-fal... and the full information is available here: http://www1.cs.columbia.edu/~sedwards/classes/2013/w4115-fal... our project was called "pubCrawl".