[Pipet Devel] inter-locus communication

[J.W. Bizzaro]

Sorry for the delay in replying to this.

Rahul Jain wrote:
> 
> What exactly is the role of Paos in Loci? As I understand it, Paos is
> simply a way of moving objects back and forth. I think that Justin's
> original suggestion was good, with some details that need to be worked
> out.

Justin's original suggestion, as I understood it, was to embed workflow
information into the XML.  I agreed that it is a novel idea, but I have two
problems with it:

  (1) It would greatly increase the amount of parsing and writing involved
  (2) It would greatly diminish the role of the object server...
            (thus I asked, "why use PAOS?")

We do need PAOS for object serving, but what the XML cannot do on its own, is
handle active links between multiple loci.  What would XML do in these cases?

  (1) Workflow information has to be reported back along the path to several
loci
  (2) Several loci need to update a single XML

The best solution for this is to have a server manage XML usage.  But you see,
this is where we need PAOS.  And if PAOS can handle the workflow information as
an XML, wouldn't it be more efficient to just keep this information as data
structure objects?

> 
> With an XML parser, it should be trivial to convert a LociML file to an
> object in any language (Python, Perl, C++, etc.).

Yes.  I do see the use of XML for archiving and transferring objects, even
workflow objects.

But I think the advantage to having an XML that is biodata-only, is that it
might be used outside of Loci.  Maybe it will be more accepted than BSML or
BioML.  But if it contains workflow structures that are inseparable from the
biological, it may never be used.

Perhaps we can make BICML so that it does not *need* workflow data to be
complete, but that it can *handle* it.

I think if we BICML strongly labels biodata with ID#'s, workflow data can be
appended to the XML, kept in another XML format, or just kept in PAOS as objects
but be easier to track.

So we have 4 options for the workflow data:

  (1) Put it in BICML, mixed with the biodata
  (2) Put it in BICML, separate from the biodata
  (3) Put it in a separate XML
  (4) Leave it as pure objects in PAOS

In all cases, I would like PAOS to handle the workflow data.

Carlos, I'm curious if an XML parser can be integrated with PAOS.  I think it
would make all of this simpler, even though a parser could be separate.

[cut to save space]
> 
> This is only for the communication between the wfs server and the
> workspace. The communication between the wfs server and the loci can be
> done in a different way. That can and maybe should involve Paos
> specifically. We can worry about that later.

Thank you for the prototype.  It brings us closer to a format definition for our
XML.  But I think communication should be handled via PAOS rather than inventing
a new system that requires each client to access Internet sockets.

Can we come up with a system for options 2 and 3 above?

  (2) Put it in BICML, separate from the biodata
  (3) Put it in a separate XML

> 
> Also, I think that the definition for transfer in the glossary should also
> include objects. Whatever, it's 2AM and I feel like nitpicking. There, I
> feel much better. :)P

I guess we just wanted a short term to describe parse/write.  Of course someone
can "transfer an object".


Jeff
-- 
J.W. Bizzaro                  Phone: 617-552-3905
Boston College                mailto:bizzaro at bc.edu
Department of Chemistry       http://www.uml.edu/Dept/Chem/Bizzaro/
--