[MOBY-l] Moby in workflows

[Simon Twigger]

Hi Tom,

I havent used Taverna but have heard about it on the list and from 
Martin. We arent using Taverna at this point but I think this type of 
workflow tool is wonderful and will be a great help to people trying to 
chain services together. I gave Taverna a try for the first time this 
afternoon and had some issues on Mac OSX, I submitted a bug report on 
sourceforge. Consequently, I havent had a chance to really play with it 
so my thoughts so far are limited to rather superficial observations.

I agree with Rebecca's point that a non-gui implementation would be 
great so that we can design workflows in Taverna but then pipe things 
into them behind the scenes. I'd love to be able to create a custom 
workflow in Taverna and then wrap it up as a 'component' or other black 
box and then be able to pipe input data to it from Java (perl?) and 
have the output come back such that I could use it appropriately. This 
would be a great way to quickly create code to solve complex problems.

Thinking about giving this to average lab end users, is there any 
chance we can create workflows by dragging connections around on the 
Workflow diagram? The Model explorer no doubt gives more flexibility 
but it would be so intuitive if people could drag connections between 
objects rather than building in the explorer.

I wonder if a drag and drop approach would work for running workflows - 
drag your input file onto a workflow icon to send it on its way, 
results being returned to the desktop (or similar).

I'll try again on OS X and XP and explore some more and try some MOBY 
services and see what happens.

Simon.

On Oct 7, 2004, at 2:40 AM, Tom Oinn wrote:

> Hi all,
>
> I run the Taverna project (http://taverna.sf.net), a workflow design 
> and enactment system which supports BioMoby and I figured it was time 
> I de-lurked and started to pester you with questions, I hope you'll 
> forgive me!
>
> I'd really just like to get a feel for whether people are using 
> Taverna; if so for what and how could we improve it, and if not then 
> why and how can we persuade you. I'm aware of a couple of shortcomings 
> in our Moby support, there's an issue with empty results and we don't 
> pick up annotations at all at the moment - it's the second of those 
> that I'm curious about (we can sort the first out, it's mostly a 
> technical issue). If you were chaining multiple service together how 
> would the annotation information manifest in this flow?
>
> Cheers,
>
> Tom Oinn (EMBL-EBI)
> _______________________________________________
> moby-l mailing list
> moby-l at biomoby.org
> http://biomoby.org/mailman/listinfo/moby-l
>
>
--

Simon N. Twigger, Ph.D.
Assistant Professor, Department of Physiology
Medical College of Wisconsin
8701 Watertown Plank Road,
Milwaukee, WI, USA
tel: 414-456-8802
fax: 414-456-6595
AIM/iChat: simontatmcw