January 2nd, 2012


Си головного мозга

Копипаста из http://lambda-the-ultimate.org/deca :

I think it's a bit funny that almost no one does "systems programming" anymore other than people who are doing embedded systems and those are plenty; probably number three after web and enterprise programming. So most people who are programming in C and assembly, who actually do safety critical and bare-to-the-metal stuff have zero benefit of any modern "systems programming language"

So maybe meeting an EE professor at lunch or another one who teaches embedded systems and who is interested in safe and efficient NFC transactions or something alike may turn out to be inspiring. It's probably not really important and we don't urgently need any new programming language anyway but it would be interesting from an artistic point of view.

I would lay much of the blame on automatic garbage collection, which while extremely useful, created a hard dividing line that sucked language innovation away from systems languages.

One of the challenges here is that the target audience has a severe case of Stockholm Syndrome - they're so used to the pathological ecosystem they've learned to live with that they can't imagine why anything else would be needed.

A great fraction of systems programmers uses languages from the 1970s and no one cares.

There is an increasing number of university projects that deals with hacking cars, trains and planes. Scary, no? Unlike PCs where software always put pressure on customers buying new generations of hardware there isn't any such pressure in the embedded sector. Since it employs armies of testers anyway it may be of little surprise that it is relatively open minded towards formal methods.

No, systems software research is far from done, but the 10th billion language that targets x86 is probably not so urgently needed as many programmers might think.