[Pipet Devel] Re: Overflow/Loci/GMS collaboration

J.W. Bizzaro bizzaro at geoserve.net
Tue Mar 14 18:58:41 EST 2000

jarl van katwijk wrote:
> Use corba to transport the XML data?

We could, but Brad has some good points about continuing the use of Loci's
command stream for core-to-GUI communication.  This API allows...

  * Fronts to be made in any language, without the use of an IDL
  * Multiple fronts to read and respond to the core SIMULTANEOUSLY
  * The command stream to be saved and then recalled at a later time
    (and have an 'instant-replay' feature!)
  * The front(s) and core to be separated via simple Internet socket

And I don't think GMS's API for core-to-GUI communication will have to be
changed.  Brad can probably plug Loci's mechanism for core-to-front
communication directly into GMS.  We'll see.

> This supports having multiple 'clients'
> attached to the core (DFP, or dataflow prosessor) even remotely across a network.

Can one client replicate the actions of another client that is operated by the
user?  For example, the user connects 2 lines in one client and a second
client does the same exact thing ON ITS OWN.

You see, this is important to Loci, because we want to have a console-based
client that can control the GUI client.  Fo instance, the user could type...

    "connect node 3728 to node 4578"

And the GUI would do it.  Better yet, with speech recognition (maybe Jean-Marc
can help with this), a user (perhaps a blind user) can speak...

    "connect node 3728 to node 4578"

And the GUI would do it!  How does this sound to you?

>  I suggest that communication between cores takes place by extending the pipeline
> functionality by a sharing ability.

Hmmm.  I like the idea in Loci of self-referencing (recursive) cores.  Could
we do that?

> All other points are already started or are fine to me. XML is the mayor item gms
> is missing.

Well, you're lucky to have Brad help then.  He's a whiz when it comes to XML.

BTW, Overflow and Loci take VERY similar approaches to the use of XML.  It
would be nice if we all standardized...maybe even just for the sake of
neatness and comprehension.

> But first everybody should  have a good look at each others code\situation. I
> might be better to do some more work seperately before any intergration attempt
> is done. I want to complete verson 0.4.0 which will take about 1-2 weeks.

Yeah, I think we need some time to look at each other's ideas, plus a
cooling-down period :-)  I want to re-package and post some of Loci's more
important ideas in that time too.

                      |           J.W. Bizzaro           |
                      |                                  |
                      | http://bioinformatics.org/~jeff/ |
                      |                                  |
                      |        BIOINFORMATICS.ORG        |
                      |           The Open Lab           |
                      |                                  |
                      |    http://bioinformatics.org/    |

More information about the Pipet-Devel mailing list