[Pipet Devel] Open DX as a loci skeleton?

[Gary Van Domselaar]

Harry Mangalam wrote:

> 
> Hi All,

Hi Harry!
> 
> I'm working on a large-scale gene expression database and analysis project
> (Open Source!) and one of teh tools that we're seriously evaluating is
> IBM's Open Data Explorer (DX), now also Open Source (a requirement for the
> project).  It's at www.opendx.org.  I just recently got it to compile and
> link on my RH6 box (50+MB of code and docs)

I mentioned IBM's Data Explorer back in the fall.  It is a very 
sophisticated and powerful data visualization environment.

> 
> It's a client/server visual programming environment that encapsulates some
> of the ideas that have been flowing on this channel (altho it's not
> Python-based).  However, it does have a modularity and architecture that
> looks looks a lot like Loci, and since it's Open Source, I wonder if it
> couldn't be mined to exploit the functionality that Loci needs.

This is very true!  If you go visit the Detailed Description page:

http://www.research.ibm.com/dx/dxDetailedDescription.html

you'll find that OpenDX's architecture and Loci's architecture are 
virtually identical.  A good example of convergent evolution ;-)

Here are just some of the striking similarities:

"The client process is the graphics user interface and it is always 
operate on a workstation. The server process does all of the computation; 
it may reside on the same of different workstation"

"The server is controlled via the data flow executive, which determines 
what tasks need to be executed base upon user requests and schedules 
their execution."

"The executive accepts a well-defined protocol (a script language), which 
the user interface generates based upon input it receives. The executive 
can be operated independently of the user interface via the 
scripting/programming language."

> 
> Has anyone here looked at it with that in mind and considered appropriating
> that which could be appropriate to appropriate?  It was once a commercial
> (expensive!) product and it runs on most all *ixs and NT (with an Xserver)
> and it has been relatively well-debugged.
> 
> The loci / module aspect of it is very intriguing.

Considering that Loci's and OpenDX's architectures are so closely aligned,
and that we have yet to hammer out the details of the 'executive', I think
we would be wise to take a peek at their code. 


Regards,

g.
--
                              Gary Van Domselaar
                           gary at bioinformatics.org
                      http://www.bioinformatics.org/~gary
            ----------------------------------------------------              
                       bioinformatics.org: The Open Lab
                        http://www.bioinformatics.org/