RE: [Myexperiment-discuss] non-taverna workflows

[David De Roure]

Hi Paul

 

That would be great – we want myExperiment to be able to be used that way, and we would work with you to try that out.  We are opening up the API for writing on a case-by-case basis at the moment (and providing a sandbox for developers) so that we understand all the issues.

 

Web-based workflow authoring is outside the scope of myExperiment just now but we’ve recorded a lot of interest in it, and where a workflow system provides this then we are interested to see how myExperiment functionality can be integrated. However, the web based interface to connect workflows up in order to execute (rather than author) workflows is definitely in scope.

 

Part of our interest in Galaxy is that it has a web-based workflow editor (which people have commented as being a bit like pipes) - it’s open source and, talking to James Taylor yesterday, it should be possible to make use of that part of the system separately.  So we’re planning to take  a look at that over the summer (in conjunction with the Taverna team).

 

Incidentally, our colleagues in Triana are also very interested in web interfaces for workflows.

 

Thanks

 

-- Dave

 

From: Paul Allen [mailto:address@hidden
Sent: 24 July 2008 13:28
To: David De Roure
Cc: Antoon Goderis; address@hidden
Subject: Re: [Myexperiment-discuss] non-taverna workflows

 

Dave,

That's great. The site I am working on (no good name yet) will be focused on authoring and executing Kepler workflows, somewhat like Yahoo Pipes. It certainly has need for a workflow repository, so I'm thinking that one option might be to use MyExperiment as that repository if the MyExperiment API is opened up for writing in addition to reading. That would open up MyExperiment to use purely as a back-end and allow a completely different UI to be layered on top of it?

Cheers,
-Paul

David De Roure wrote:

Hello Paul

 

Just to follow up on Antoon’s comments, the myExperiment codebase is designed to share any “contributable type”, and currently we’ve focused on Taverna workflows and on packs.  We are developing support for a variety of other scientific artefacts including, for example, experimental plans used in the context of electronic lab notebooks in chemistry, and also statistical models. Kepler is very much part of the myExperiment picture, and we have enjoyed several fruitful conversations with Bertram about this. In our workplan we are working now with Triana and will be coming to Kepler in the period starting this fall.

 

We follow the Web 2 design principle of cooperation – we’ve put a lot of work into the myExperiment API so that myExperiment functionality can be brought through in other interfaces (eg google gadgets, wikis, the Taverna plugin).  It would be great to discuss how we might do some things together.  We’re currently planning some integration work with Galaxy, for example.

 

Let’s look at some options and see what we can do.

 

Many thanks

 

-- Dave

 

From: address@hidden [mailto:address@hidden] On Behalf Of Paul Allen
Sent: 23 July 2008 20:12
To: Antoon Goderis
Cc: address@hidden
Subject: Re: [Myexperiment-discuss] non-taverna workflows

 

I am happy to hear about a roadmap that involves support for Kepler workflows in MyExperiment. There are several efforts already underway or in the planning stages for a web front-end for dealing with Kepler workflows. I think that Hydrant (https://www.hpc.jcu.edu.au/hydrant/) is the only one that is public.

I hope that those of us working on Kepler-based web sites can find ways to collaborate and leverage each other's efforts instead of directly competing.

Antoon Goderis wrote:

Hi Jan,

On Wed, 23 Jul 2008, Jan Aerts wrote:

All,

Is it possible/advised to put non-taverna workflows on MyExperiment?

You can do it by uploading them as a File. You get all the usual goodies

but they don't show up under the workflows tab. At the moment workflows,

when uploaded, have to parse as a valid Taverna workflow, because an

image is rendered and documentation extracted based on the Taverna format.

It's definitely within the remit of the project to have different types

of workflow on myExperiment - so the current situation is temporary and a

matter of resource. Explicit support for Triana workflows and Kepler

workflows (two other environments) are next on the to do list, for

instance.

Antoon

 > Michael Barton

has blogged about using rakefiles (the ruby equivalent > of makefiles) on

http://www.bioinformaticszen.com/2008/05/organised-bioinformatics-experiments/.

I have used that approach a lot and believe it is a very good

alternative for doing scientific workflows where it is impossible or

not advisable to use webservices for the steps of the process.

Michael has put one of his workflows online on github:

http://github.com/michaelbarton/cost_in_evolution

Or is MyExperiment.org too much tied in with Taverna?

jan. (saaientist.blogspot.com)

_______________________________________________

Myexperiment-discuss mailing list

address@hidden

http://lists.nongnu.org/mailman/listinfo/myexperiment-discuss

_______________________________________________

Myexperiment-discuss mailing list

address@hidden

http://lists.nongnu.org/mailman/listinfo/myexperiment-discuss