[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Intel: timely article

On 10 Jul 2006, at 16:04, Barry Cook wrote:

Another article in the same issue - "Can We Make Operating Systems Reliable and Secure?" - (Tanenbaum et al) pp44-51 also makes interesting reading (another familiar message). There is mention of another language "Sing#" which "fully supports channels in the language, including formal typing and protocol specifications". The syntax even includes '?' and '!' !!! Also implemented is data owned by one process and ownership passed via channels. Again no references to the proponents we know.

Just a note to say that we may well have to tolerate the idea that 'they' (whoever it may be) thought of all of it first, when CPA does filter through as the way to program concurrency (and reactive behaviour).

It would not be the first time, though Hoare's classic CSP book, is hard to ignore, not to mention the transputer and occam. Parallel evolution will then follow.

An article in IEEE Computer would be powerful, though suspect an alien invasion would be resisted. Editors will be loath to upset the OOP crowd. Am afraid any such paper probably will not be RAE- countable - IEEE Computer is classed as a 'magazine', not a journal. The same status is attributed the now legendary CACM.

IMHO, the only way to win is to build real software with tangible (ie financial) advantage. How can anyone read 3,000+ journals anyway! Truth is no-one does. I don't think writing anything but salable software will effect change.

Perhaps aiming at the right people and trying harder to solve _their_ problems might work better. In my humble experience, getting a modest contract has actually been easier than getting a grant, or a paper in an RAE-listed journal, for that matter. Pays better too.