RE: [Myexperiment-discuss] myExperiment "look and feel" a Flickr mockup(workflow)

[Antoon Goderis]

Hi June,

> In the gold standard work, these workflows are specifically annotated 
> "as services" - with a description,inputs and outputs and expanded 
> pointers to the included service descriptions; in line with 
> the way the 
> services themselves are described.  ... and wow, we have 
> loads of stuff.
> 
> If we then add all the other data that is associated with the 
> experimental/organisational/provenance aspects of the 
> workflow (rather 
> than the technical sum of it's parts) it gets very very massive as 
> Duncan and Antoon have said.
> 
> We need to think about presenting different views of the workflow to 
> support the different reasons people view them - and not fall 
> into the 
> trap of a massively complex one size fits all view with an 
> intimidating 
> amount of confiiguration. 
> 
> Longer term, we may also need to allow the user to configure 
> a view of 
> their workflow suitable for publication of paper purposes. - I could 
> imagine institutional looks&feels and a whole slew of logos being 
> necessary in some cases :-)

David has a few ideas on generating customised Wiki pages based on the
user's profile

Wrt the longer term comment, here's a pointer to relevant work. A while back
I attended a presentation on work done within the CLEF project
(http://www.clinical-escience.org/) on text generation. The CLEF prototype
can generate customised reports of medical histories based on the different
parties and their viewpoint. For example, a physician interested in
providing a particular treatment, a patient wanting a personal copy of their
history or the needs of the insurance company. 

Given a similar encoding of in silico experiments (workflow, constituting
services, data, provenance, related publications), we might be able to
generate customised textual descriptions of these, too, depending on the
user context. For instance, the bioinformatician's service centric view (how
often does this workflow work), the biologist's data centric view (provide a
textual summary of the genotype to phenotype transformation lineage), the
lab head's publication centric view (generate me the methods section for my
paper). 

This to complement any graphical view already present. One interesting
aspect of the work was the interest in generating hypertext, though that
part was very young.

Antoon