[Bioperl-l] Re: [Bioperl-l] New code

Steve Chervitz steve@trutane.net
Mon, 31 Dec 2001 01:32:36 -0800


Aaron wrote:

>On Fri, 28 Dec 2001, Steve Chervitz wrote:
>
>> Bio::Tools
>>   Bio::Tools::StateMachine
>>   Bio::Tools::StateMachine::AbstractStateMachine
>>   Bio::Tools::StateMachine::IOStatemachine
>
>For some reason, these didn't come through when I cvs update'ed ... so,

Did you remember to "cvs update -d"? These are in a new directory.

>forgive me for having not looked at them yet, but how are these different
>or better than POE, and for what are you going to use them for?  I know
>we've discussed (at length) minimizing external dependencies, but this
>just sounds like a very big ball of wax that doesn't need reinventing
>(and of course, I'm a fan of POE).

Look at the docs in AbstractStateMachine for orientation. I wrote the Bio::Tools::StateMachine::
modules for use with my Blast parser re-implementation, but they are generic
and by no means tied to the Blast parser.

They use a different strategy than POE that I think is more intuitive to
the biologically-minded coder. The main reason I wrote it was to have a
simple state machine framework that was internal to Bioperl and could
be integrated with Bioperl methodologies (i.e., Root and IO).

I'm very interested in what you think about it (positive or negative). I
put a simple example in examples/state-machine.pl. I have another example
that I'm considering checking in that is more real-world. Perhaps I can
send it to you if you need more of a taste.

Steve

>Again, forgive me if I'm way off base because I've only read the titles
of
>your modules and not seen their mechanism or proposed usages ...
>
>- -Aaron