Re: Hokay, I officially give up...
Gordon Milne <[email protected]> Fri, 22 May 2009 15:24:03 +1200
| Newsgroups | gmane.comp.programming.language-of-the-year |
|---|---|
| Message-ID | <[email protected]> |
Perhaps now is the time to look as Scala and see if its actor implementation helps. I haven't tried this myself since I find it hard to justify the time to try something new when what I have already works but ... I have this nagging feeling that all that explicit locking and synchronized keywords stuff will come back to haunt me and bite me in the behind when I can least afford it. I need to spend time on this but I cannot justify it. This sort of stuff is destined to drive us all mad. Regards, Gordon J Milne 2009/5/22 John Carter <[email protected]> > I have sort it with needles, I have sort it with care, I have pursued > it with forks and hope... > > ... I have found bugs, in my code and fixed'em. I have found bugs in > ruby, and fixed them. But there are more and I can't see the bottom of > the pool. Each time I thought, "AH! I have FOUND IT!", but I have > merely decreased the probability of occurrence. Each time I think, if > I just get this right, it's going to be a small contained no-risk > tweak to my app... it's worth ploughing on. > > Threading is a truly awful paradigm and I truly regret ever having > anything to do with it. > > Threading in an interpreter is intrusive and incredibly subtle, > putting hooks into every corner. Maybe one day they will get it > perfect... but I doubt it. > > In fact, the next generation of ruby uses pthreads.... which leaves me > with a feeling of cold dread. Why? Because last year I wasted more > time than I care to account for tracking a race in a the glibc pthread > library routines. > > Threading is very easy to do, easy to get 99.9% right, but incredibly > hard to get 100% robust. You can get it 99.9% right, or 99.99% right or > even 99.999% right... but right always and in every case seems to be > incredibly hard. Not just for me apparently, but for OS, library and > interpreter writers as well. > > And alas, unless _everybody_ gets it perfect, the end result is still > flawed... and as we get more and more cores on the CPU, the problems > are just going to get bigger. > > I'm giving up on them and ripping out the word Thread from my entire > app. > > Alas, threads still exist as cruft within in the OS, libraries and > interpreters making them all much bigger, much more complex and a lot > buggier. > > Sorry, it is time to admit Moore's Law has just crashed into the > immovable barrier of Grunt's Law (Human aren't getting smarter).... > > I'm just plain not smart enough to do threads. > > John Carter Phone : (64)(3) 358 6639 > Tait Electronics Fax : (64)(3) 359 4632 > PO Box 1645 Christchurch Email : [email protected] > New Zealand > > > > ------------------------------------ > > Yahoo! Groups Links > > > >