Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

> Who exactly are these new C-standards for?

For me and for many colleagues in my lab? C is quite big in scientific computing and signal processing. Fortran would be slightly better, and it is widely used, but not directly around me. The C99 standard, which added complex numbers and variable length arrays, was truly a godsend in the field. I cannot imagine working without it.

If you write a numerical algorithm that needs to be run 15 years from now, then C and Fortran are possibly the sanest choices. If you do something in other, fancier, languages, you can be sure that your code will stop working in a few years.

The new C standards are really minor changes to the language, and they happen in the span of a decade. It is quite easy to be up to date. And in the rare case that your old code stops compiling, the previous (less than a handful) versions of the language are all readily available as compiler options in all compilers. You can be reasonably sure that a C program written today will still compile and run in 20 years. You can be 100% sure that a python+numpy program won't. If you care about this (for example, if you are writing a new linear algebra algorithm to factor matrices), then choosing C is a rational, natural choice.



> You can be 100% sure that a python+numpy program won't.

It's possible to use a phyton+numpy program in 20 years, but you also have to save the entire environment and make sure that it works air-gapped (otherwise external dependencies would fail). One possiblity would be to store it as a qemu virtual machine. It's very possible today to boot up stuff as VMs that is 20 years and older (e.g. 20 year old Linux distros or Windows XP iso from early 2000s).




Consider applying for YC's Summer 2026 batch! Applications are open till May 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: