To view parent comment, click here.
To read all comments associated with this story, please click here.
Wirth only says that it memory protection "should" have been superceded by safe languages and correctly implemented compilers.
While creating a safe language is certainly possible (if you're prepared to pay the cost in flexibility and/or performance), ensuring a correct implementation that is safe not only from accidental bugs but also from malicious exploits is a huge problem.
Meanwhile, hardware memory protection may not be as fain-grained as compiler-based schemes, but it costs very little in performance and one can be fairly certain that it's correct.
Well you see the latest 3rd episode (I think) of channel 9's coverage of the Singularity project http://research.microsoft.com/os/singularity/ . IIRC they had about 10% increase in perform. To put that in perspective that would your safer compiled language would run faster than your C code. Java could be beating it regularly (but not always of course, because under some situations Java just starts to suck shit).
> hardware memory protection [cut] costs very little in performance
That's not what the guys from the Singularity project said: they had the same level of performance for system call with their unoptimised implementation as Linux or Windows.
The reason why they were able to do so is that they are not using the hardware memory protection.





Member since:
2005-09-03
there is this even more famous computer vip who seems to think memory protection not really necessary see 3.7 of
http://www.cs.inf.ethz.ch/~wirth/Articles/GoodIdeas_origFig.pdf