[Bioperl-l] Re: Bio::Root::IOManager: used?

Hilmar Lapp hlapp@gmx.net
Sun, 24 Sep 2000 14:20:01 +0200

Ewan Birney wrote:
> I realise that a little more consistency is a good thing here, so if we
> can collectively come to a decision about
> (a) do we need Bio::Root::IOManager
> (b) how should we use it
> that would be great. What are other people's views?

My experience is that we are duplicating methods fh()/_filehandle()
(doing the same thing), _pushback(), _readline(), close() and DESTROY (to
close files upon garbage collection) all over the place, as well as code
in _initialize() to handle and set -file and -fh parameters.

It's not that any of these methods is complicated or substantial code,
but I don't think it's smart either.

So, I wouldn't expect fancy things, but just those little ones you in
fact need all the time.

> >
> > Finally, there is a similar question concerning network access and
> > URL-based queries: do people think that each module which needs it shall
> > implement it itself, or rather delegate to a yet to be created
> > Bio::Root::NetManager? This of course affects the proxy/firewall problem,
> > too (i.e., the place of where to solve it).
> >
> See previous post by Chris. Perl again has a rich set of "standard" URL
> getting stuff. We should not need to rewrite things in my view..

In my view the point is not to rewrite something, but rather delegation
of the guts. You already quote "standard", and the fact that the
Bio::DB::Gen* modules are not firewall compatible indicates that it may
not be as straightforward as it looks like.

So, I think your point of a little more consistency and a common decision
(like what do we recommend to use) are valid here as well.


Hilmar Lapp                                email: hlapp@gmx.net
NFI Vienna, IFD/Bioinformatics             phone: +43 1 86634 631
A-1235 Vienna                                fax: +43 1 86634 727