[Pipet Users] Re: Piper development

[J.W. Bizzaro]

cognite at zianet.com wrote:
> 
> o       What is the state of Piper development?  Looking at the cvs from
>         www.bioinformatics.org/piper
> it looks like few changes have been posted since a server was moved some
> months ago (3,5,6,9,10,11 months ago most changes) and the three projects
> began to be integrated (gMessaging, Loci, Overflow).  Is that not working,
> or something?

The state of the CVS module doesn't reflect the state of Piper very well.  As
you said, Piper is a collaboration between 3 projects, and each tends to be
developed somewhat independently.  Overflow is not in the CVS module at all,
and it is perhaps the most actively developed.  gMessaging has undergone some
major changes recently, but thay haven't been committed to CVS.  Loci is the
only one whose activity is reflected by CVS, but that's because one major
contributor moved elsewhere and the other one (Yours Truly) is busier
evangelizing than coding :-)

> o       Python is shown as the language of the Piper Build-time subsystem
> (along with CORBA communications), as opposed to C++ for its Run-time
> subsystem in the diagram of slide 10 of the Feb. 2001 slideset of
> J.W.Bizzaro posted on the site.
>         www.bioinformatics.org/piper/dpci,emtatopmbpse2000/slide_10.html
> 
>  But in the cvs, I see only one python file (a super-simple __init__.py).

You have to look a bit deeper.  All of the UIL and DL are Python -- lots of
it.

>         It looks like the role of the python may be for GUI-to-process
> ("User Interfaces Layer") as well as something called "Definitions Layer".
> Why would python not be used for the Brokering layer of the Run-time
> subsystem?

The UIL and DL were originally "Loci" and written entirely in Python.

The BL was originally "gMessaging" and written entirely in C (now C++).

The PL was and still is "Overflow" and written entirely in C++.

The different languages are artifacts of the collaboration, but since they are
"connected" via CORBA, and the whole system is modular, having more than one
language isn't an issue -- except library dependencies, but we managed to
reduce it to 2 languages Python and C++.

> -- The other informationflow? diagram in the
> www.bioinformatics.org/piper/design subpage is far more complex; the bottom
> 2 rows don't seem to correspond directly to the slide. Perhaps you could
> explain it a bit more?

The former was written more recently by Jarl van Katwijk, and it's mainly
about his part: brokering.  The slides were written by myself a year ago, and
it gives a rough idea as to how Piper is modularized.

> Thanks!  Best wishes.  The idea of overlaying inputs from various URLs onto
> a user screen set up with XML, with commands using variable parameter values
> underneath, built with process diagrams having enterable value slots, looks
> like certainly a coming thing.

Yes, we've seen hints of it around, and Piper is an attempt to make a Free
Software implementation with some other very unique ideas.

Cheers.
Jeff
-- 
J.W. Bizzaro                                jeff at bioinformatics.org
Director, Bioinformatics.org        http://bioinformatics.org/~jeff
"As we enjoy great advantages from the inventions of others, we
should be glad of an opportunity to serve others by any invention
of ours; and this we should do freely and generously."
                   -- Benjamin Franklin
--