[Cryptography] It's all K&R's fault

[tpb-crypto at laposte.net]

> Message du 22/04/14 05:52
> De : "Lodewijk andré de la porte" 
>
> 2014-04-22 2:50 GMT+02:00 ianG :
> 
> > > I'm sorry Ian, I have to raise an eyebrow at what is essentially saying
> > "kids today are no damned good." When I was a kid, back in those days, kids
> > were no damned good. C and its ilk were for lazy people who didn't want to
> > learn how to program the *machine* dammit.
> >
> >
> > Couldn't think of a snappy reply to this in less than a few days, so I
> > guess I'm now old and cranky...
> 
> 
> Lately all the talk of distrusting your compiler has me thinking this is
> the right way to do it. It's not, of course. Kernels of the future will be
> written in C# (or devirtualized Java) and hopefully Haskell (of maybe F#,
> if we're not fully lucky).
> 

Me and a few friends tried to create a JavaOS a few years ago and it was awful. Too many abstraction layers.

Brace yourself for decades of OSes made in C ahead.

Maybe someone will dare making something with C++. But we are no longer in the early 90's that a kid in college could jumpstart an OS and have it adopted by the masses. The amount of devices and platforms and services and protocols to support is mind boggling.

People once were content with a terminal, now you got retina displays, sound cards with stereo surround 5.1 7.1, dozens of optical formats, hundreds of usb devices, digital cameras, ... it is an undertaking requiring lots of money and a daring entrepreneur.