SEDA has many lessons that can and should be adopted, as we discover the best
possible method to do so. There are things that we do better, and do not want
to discuss further:
1) Configuration--we rock, they only provide a wrapper arround Properties
2) Code standards--It is obvious that Matt Welsh has a background in C/C++
as there are non-standard method names like enqueue_lossy() or class
names like httpConstants. There seems to be no rhyme or reason as to
why some classes are standard and others are not.
3) IOC--He does it in some ways, but there are so many pointers to different
parts of the SEDA framework, that it is hard to figure out what is going
on.
4) SOC--The ThreadPool is decidedly SEDA based. It would not be something
you could rip out and use in a completely new context.
Never the less, there are some things that cannot be ignored:
1) Most tasks in a system are small and incremental, and using full threads
for each task is overkill.
2) A callback/event queue system is a poweful concept--and he made inroads
on how to simplify the process.
3) We can take his ideas to a new level of usability, and it would be
beneficial for us to take advantage of those ides.
--
"They that give up essential liberty to obtain a little temporary safety
deserve neither liberty nor safety."
- Benjamin Franklin
--
To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
