Re: [Myexperiment-discuss] a requirement while I remember!

[Stian Soiland]

On 8 Aug 2007, at 18:45, M. Scott Marshall wrote:
> Wouldn't it be interesting to know how often a workflow has been  
> run? I
> imagine that if, by downloading a workflow from myExp (counter++), you
> actually got a slightly modified version: a version with a Taverna
> tracking code/workflow identifier inserted into the SCUFL (similar to
> the code you insert for Google Analytics or other web stat  
> service). So,
> when you run the workflow, a counter corresponding to that workflow  
> gets
> knotched up by one iota on a myExp server. And when it runs to
> *completion* you get a few more iota's. This would not require any
> special effort on the user's part (but admittedly some special  
> effort on
> the part of my* developers..).

There's some privacy issues with such reporting, but we can do it in  
a slightly similar way if we include it as options:

[  ]   Report anonymous workflow usage statistics to myXP
[  ]   Report anonymous service usage statistics to myXP

This would be a part of an myExperiment plugin (that would also be  
able to tell you if there was a new version of the workflow, let you  
submit the workflow straight to myXP, etc).

All that needs to be inserted into the downloaded workflow would be  
some URI or another identifier (obviously the never-changing LSID is  
not very good for this) to report which workflow is in question.

Similar identification issues would arise when reporting service  
usage, but that shouldn't be too hard as at  least we know how to  
call the service, and that can be used as some kind of identifier.  
(Filtered through some sha1/md5 hashing to avoid a massive URI and  
avoid exposing which internal services have been used)

--
Stian Soiland, myGrid team
School of Computer Science
The University of Manchester
http://www.cs.man.ac.uk/~ssoiland/