> Now ask the users, and the number one complaint I hear most often is "why is this thing so damned slow?"
Good point. My viewpoint is rather focused on numerical linear algebra, since I worked there and the article is about it. If you skim through the paper mentioned, in the article you can e.g. compare numpy and armadillo. You will see that the speedups from using a (arguably) much more complex C++ framework instead of numpy are marginal and will not be visible for the user. The increased production/maintenance costs due to a more complex code, will be visible for the user.
Theoretically you are correct, but taking the skills of the user into account, I think the very opposite is true. In practice, hardly anyone has the skills to actually implement anything with high performance anymore (especially in numerics) and it is thus (in most cases ) more performant to simply stick to stuff like numpy.
In most cases, "working code, fast" is more important than "fast working code".
(From my very limited understanding) making commits requires you to find a specific hash, thus making new commits exponentially harder over time.
At some point it would cost significant money (besides the developers time) to make commits. It's basically an economic game like bitcoin, that those people who have economic interest in the code, get to decide what's in it.
A Problem with accelerator physics particularly is that there is a large scientific industry around the actual experiments that depend on these experiments continuing. The engineers just get told what to build and they get going, without having a deep understanding why this particular experiment is really necessary.
I worked in accelerator physics for 5 years and on all the conferences, meetings etc. I had yet to meet anyone who understands, why we are doing this (with an understanding that goes deeper than Wikipedia).
And i didn't either, I called it a tool for serverless architecture. In fact some tools for serverless architecture like AWS Fargate require you to use Docker.
The long startup time is somehow what keeps me from really liking Docker. In the end (if I understood correctly), it is supposed to be the go-to tool for serverless architecture. If my serverless function needs more then a second to startup, it's not usable for me.
Even the hello-world container, which is only a few kB in size needs roughly a second to startup.
Good point. My viewpoint is rather focused on numerical linear algebra, since I worked there and the article is about it. If you skim through the paper mentioned, in the article you can e.g. compare numpy and armadillo. You will see that the speedups from using a (arguably) much more complex C++ framework instead of numpy are marginal and will not be visible for the user. The increased production/maintenance costs due to a more complex code, will be visible for the user.