On Fri, 17 Sep 2010, Juha Manninen (gmail) wrote:
I honestly wonder the attitude against threads here. I hear comments of Pascal
being a dying language from the past. Maybe it is true after all.
How can one take seriously a developer community which claims threading is
inherently wrong and bad?
The community doesn't claim so.
It is the result of independent research on the subject.
You must separate 'parallel computing' and 'threads'.
Threads are one way of implementing parallel computing.
Note that the community doesn't say that parallel computing is bad; of
course not; The opinion (of some, obviously) is simply that threads are
not the best way of accomplishing parallel computing.
BTW, TThread class has a good and intuitive interface, IMO. A new language
syntax for parallelism would be nice but it wouldn't make any fundamental
difference. Start a thread and call WaitFor later. How much easier can it be?
No-one said that it was not easy to program threads.
I use them myself extensively.
But the possible side effects were pointed out, and the consequences that usage
of threads has IN THE IDE. And the opinion was raised that if there is a way
to accomplishing the same task with our without using threads, then "no
threads" is preferred for the IDE. This does not mean that a threading way
will be rejected; you'll just need to argue the need for threads a bit more.
I understand that it may be harder and less easy for your particular task.
Maybe you can submit your work so the IDE experts can offer better and more
responsive ways of doing the task.
Michael.
--
_______________________________________________
Lazarus mailing list
[email protected]
http://lists.lazarus.freepascal.org/mailman/listinfo/lazarus
