Getting locks and sharing: was RE: Getting concurrency

ke.han <>
Wed Jun 15 05:50:14 CEST 2005


I agree with Ulf and Joe on the issue that erlang solves concurrency in 
a cleaner fashion for real world business problems.  I have built very 
tedious Java applications and systems frameworks to remove threading 
issues from the app programmer.  Its a very difficult job.  I'm very 
hapy with my switch to erlang.  The Bray blog post dismisses erlang too 
easily but you will see this type of dismissal continue in other ways if 
erlang doesn't "grow" in ways to meet the multi-core and multi-cpu market,
Multi-core and multi-cpu system need to be supported with a single 
erlang node in order for the rest of the programming world to not 
quickly dismiss erlang.  I've read the research paper on multi-cpu 
erlang...looks promising. Is there any chance of this becoming a real 
product?
thanks, ke han


Ulf Wiger wrote:
> Den 2005-06-14 20:13:14 skrev Thomas Lindgren <>:
> 
>> And don't forget other aspects such as efficiency,
>> scalability, reusability, maintainability, ... E.g.,
>> "Does this highly optimized design I inherited
>> actually work? What if I also have to implement these
>> new requirements?" etc.
> 
> 
> I think one approach to seriously consider is to
> forget about making code "highly optimized", when
> we're talking about grid computers and multi-core
> architectures. Focus on writing conservative code
> and choosing safe concurrency patterns. When it
> comes to concurrency, optimization is truly the
> root of all evil. Use all those CPUs to buy
> yourselves the luxury to write safe code for a change,
> rather than cutting corners.  ;-)
> 
> /Uffe
> 
> 




More information about the erlang-questions mailing list