A lot of people are talking about the problem itself - which is interesting - but very often you run into real problems which cannot be simplified so easily (NP hard problems, many different other cases). In addition, when you need to do important calculations on a server for your product and you're using Ruby, you may very well end up requiring 10x as many servers. This is pretty massive when you consider an AWS ser…
Do you think we as a community should be pushing for a potential rewrite of MRI to get rid of the GIL? I feel as if the GIL is what is limiting true ruby performance. It also seems like Matz and ruby core team don't really give a shit about performance. It is very possible to achieve a high level of performance on an interpreted language!!
There's an argument that removing the GIL would require a lot of effort for little gain (and in some cases decreased performance). It's just not a priority for them.
Of course, every use case is different (sometimes very different), but I haven't ever hit a problem that has caused me to curse the GIL.