Re: [Myexperiment-discuss] Re: [MYGRID] yourExperiment

[Paul Fisher]

Hi Marco,

Hmmm....intersting......my preference would be to show the whole 
workflow. But, as you say, this may put people off re-using it. I think 
this is where I need my post-it note idea - a little yellow lable around 
a collection of services that describe what it does, like you would put 
a post-it note onto a folder of documents or a page in a book.

Again, as you rightly say, it is the organisation of the workflows that 
is most important when it comes to browsing.

cheers,
Paul.

Marco Roos wrote:
> Hello Paul,
>
> > Your workflows are BIG.
> This is very interesting, because it seems that how workflows are 
> depicted on myExperiment (i.e. expanded) can have a big impact on 
> their reuse  =-O
> For me, the workflows should be depicted in collapsed form. Paul: what 
> is your preference? myExperment designers: can that also be made an 
> option?
>
> My workflows are not very big from a functional point of view, but 
> this only becomes clear in the collapsed form of the workflow. Like 
> Paul I use nested workflows a lot. I prefer to use small components 
> organised in subworkflows, which I think is good practice. However, 
> this does lead to larger workflows that look more complicated, 
> especially in their expanded form.
>
> The other thing is that if you upload more smaller workflows that do 
> less (i.e. more reusable), it becomes more important how you are able 
> to organise them. I hope the packs help in this respect and the 
> possibility to address subworkflows through the container workflow. 
> Now, my list of workflows is very messy.
>
> Regards,
> Marco.
>
>
> Paul Fisher wrote:
> > Hi Marco,
> >
> > Those points are the same as I have thought. This may be of interest 
> > to you
> >
> >   1. Your workflows are BIG. That has certainly put me off attempting
> >      to re-use them for the time being. To understand a BIG workflow
> >      requires a lot of time and a lot of understanding of the processes
> >      involved.
> >   2. Re-uing a workflow is a very hard task to do. The smaller the
> >      workflows the better. Even if it means hiding certain sections
> >      (i.e. nested workflows). That is why I try to modularise mine as
> >      much as possible.
> >   3. Being able to rate your own workflows seems counter-intuitive. As
> >      you can see, I have rated some of mine as 5/5!!! (for testing
> >      purposes obviously :S)
> >
> > Re: Ad 3, I had a lot of interaction with the people involved in
> > re-using my workflow. That interaction involved me sitting directly next
> > to the person who wanted to use it, and building it! The "re-user" had
> > little to do, other than make their workflow output match the input of
> > my workflow. I had to modify my workflow to get the connections to work
> > properly. As you can see, it is not a simple task. And, I can't afford
> > to spend time with everyone who wants to re-use my workflows.
> >
> > Re Additional points
> >
> >    Point 1. The ratings can be improved by (as you say) adding more
> >    metadata. The more metadata you enter, the better the automatic
> >    rating. This would give an incentive for people to properly annotate
> >    their workflows. myExperiment could then feature a monthly "top
> >    rated" workflow, or "workflow of the month". These only contain
> >    workflows that are properly annotated. They could be advertised on
> >    the front page with a note saying - "Want your workflow here - then
> >    annotate!!" or something.
> >    Point 4 . More visible human annotation would be fantastic.
> >    Something similar to the myExperiment plugin. You see the workflow
> >    and the description at the same time. Very useful!!!! That way I (or
> >    new users) don't have to learn to look in un-obvious tabs to find
> >    the metadata.
> >
> > Ad 3 (again).
> >
> >    I would also like to know who was using my workflow. At the moment
> >    it is left up to the people using it to either ask for help and
> >    therefore tell me, tell me out of courtesy, or not tell me at all.
> >    Can a digital signature be added to a workflow (questions arise
> >    about modifcation though) that would allow an update to be sent
> >    automatically to the owner/builder to tell them it has been
> >    re-submitted to myExpeiment as another modified workflow?
> >
> > regards,
> > Paul.
> >
> >
> > Marco Roos wrote:
> > > Hello all,
> > >
> > > Thank you all for starting this thread. It allows me to share some
> > > thoughts with you and make some suggestions for myExperiment and 
> > > Taverna.
> > >
> > > Some observations or assumptions:
> > >
> > >    1. Most downloads might still be by me (from different locations)
> > >    2. Workflows on top are more likely to be tried, and hence stay 
> > > on top.
> > >    3. My workflow might not yet be good enough for reuse; I would have
> > >       expected more questions then. I would rate my workflow 3.5 out
> > >       of 5: 'reasonable'. I assume not many workflows would get 4/5 or
> > >       more from me if I would be truly objective. Reusability would be
> > >       my prime criterium.
> > >    4. It is not trivial to reuse workflows and often persuading to
> > >       think that building your own is better
> > >          1. It is not trivial to understand other people's work (as it
> > >             is not easy to annotate properly to make others 
> > > understand).
> > >          2. You are more in control over your own workflow and you
> > >             understand it much better. (I wonder: would there be an
> > >             inverse correlation between how much you understand your
> > >             own impressive workflow and how understandable it is for
> > >             others?)
> > >          3. A workflow may do too much, too complicated, to be reused
> > >             as a component.
> > >    5. For me the social aspects of myExperiment are an essential
> > >       bonus, but most important functions: registry and publication.
> > >    6. You can't see e-mail contact inspired by myExperiment.
> > >    7. My thesis: myExperiment has not fully matured as long as
> > >       first-time enthusiasts like me are still on top. ;-)
> > >
> > > Ad 3: Paul: your workflow has been successfully reused: how much were
> > > you (or mygrid people) actively involved in making that happen?
> > >
> > > Some requests to help address these issues:
> > >
> > >    1. Expert rating and expert review. You see this also on shopping
> > >       sites and often are more trusted than user ratings (which could
> > >       be biased, especially when there aren't many yet). For
> > >       scientists this is like peer review.
> > >    2. Automatic rating. A number of things can be rated without human
> > >       intervention, such as the amount of descriptions added to the
> > >       workflow. This could already be visible in Taverna while you are
> > >       designing.
> > >    3. Usage notification for workflow creator (cf 'return receipt'), I
> > >       suggest to add these options to the sharing tab (in addition to
> > >       science commons).
> > >          1. Allow anonymous usage
> > >          2. Request contact information, but allow anonymous usage
> > >          3. Do not allow anonymous usage, contact information is a
> > >             prerequisite
> > >          4. Reuse only upon accepted non-anonymous request.
> > >    4. More visible human annotation in Taverna. I want to see
> > >       immediately if my descriptions are still valid while I'm editing
> > >       a workflow.
> > >
> > > Ad 3. This seems a fair request to me. I do not mind anybody using the
> > > workflow that I have put a lot of time and effort in, but I would
> > > often like to know who. In some cases I may want the reuse of my work
> > > only if we officially call it a collaboration, hence lead to
> > > co-authorship (this would be option 4). In other cases it will allow
> > > me to offer help. NB I suggest that the message a user (the one who
> > > downloads) would get will be very friendly in the sense of 'the
> > > workflow creator requests your contact details to provide help and for
> > > your feedback'.
> > >
> > > In general I think we could think about how myExperiment could
> > > actively help stimulate collaboration and reuse (without becoming
> > > annoying of course). It will be interesting what BioCatalogue will do,
> > > as the threshold for reusing the smallest components might be lower
> > > than for larger workflows.
> > >
> > > Hope this was useful.
> > >
> > > Cheers,
> > > Marco.
> > >
> > > Paolo Missier wrote:
> > > > this is certainly impressive and well done to Marco, but it makes me
> > > > wonder how those numbers square with the number of reviews (0) and
> > > > comments (1, by the author) and ratings (2).
> > > > If these are real users, doesn't that make for a _lot_ of users who 
> > > > are
> > > > not interested in closing the collaboration loop, as it were.
> > > >
> > > > just wondering --Paolo
> > > >
> > > >
> > > > Paul Fisher wrote:
> > > > > Marco,
> > > > >
> > > > >    BioAID_DiseaseDiscovery <http://www.myexperiment.org/workflows/72>
> > > > >    (v1) [http://www.myexperiment.org/workflows/72]
> > > > >
> > > > >    *Viewed:* 3045 times | *Downloaded:* 1060 times
> > > > >
> > > > > WOW!!!!
> > > > >
> > > > > Well done.
> > > > >
> > > > > Paul.
> > >
> > > --
> > >
> > > Marco Roos
> > > Faculty of Science
> > > University of Amsterdam
> > > Kruislaan 403, room F1.02
> > > 1098 SJ Amsterdam
> > > tel. +31 (0) 20 525 7522
> > > http://home.medewerker.uva.nl/m.roos1 (includes links to social 
> > > networks)
> > > Note the change of e-mail address to address@hidden