[MOBY-l] Comments from recent Taverna/services demo

[Twigger Simon]

Hi there,

Just thought I'd share some of the comments I heard from our lab  
folks following a taverna/webservice demo I just did. I expect these  
are not new to most people but they do represent barriers to  
potential users of this technology that we need to actively address  
so I thought a quick reminder/discussion wouldnt hurt.

Provenance:
How do we know what data a service is running on, what version of an  
algorithm, what parameters its using, when the database was last  
updated, etc?

This seemed to be the biggest concern - not knowing what was going on  
under the hood and having no way to find it out. Workflows and  
services provide great ways to automate things but theres a distrust  
of the 'black box' that this creates. From a MOBY perspective, if we  
could somehow enforce the provision information block so that service  
providers had to fill in something sensible for the service  
(algorithm version, database version, date last updated, etc.) this  
could be a very good selling point. "Use MOBY services where possible  
because at least then you can get a list of what went on in your  
workflow, what databases were searched, etc." Some sort of report at  
the end of a taverna workflow that compiled all this provision data  
into a human readable document would be great.

In reality, this is the same as getting data off a website - how do  
you know what's going on? Good websites tell you what algorithm and  
database they use, others dont. I suspect people are more familiar  
and hence more trusting of websites and havent yet got comfortable  
with webservices/workflows so the level of (blind?) trust isnt there.

Input's and Outputs - How do I know what data I should pass in?
Again, MOBY is nice in that it defines the objects and should  
ameliorate a lot of these problems but joining the dots is harder  
than it should be, particularly outside of MOBY-land.


Make it even easier to use
"Is taverna something even I could use?", "I got lost with all these  
service options"

Drag and drop, connect the dots services seems pretty straight  
forward but it may still be too much for your average (or even above  
average) lab scientist. How have other people found this technology  
has been accepted? A web front end to all these services and  
workflows seems to be one of the best ways we could get this  
technology in use by the largest number of people.


I'd love to hear any comments/thoughts about these issues from others.

Cheers,

     Simon.

--

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