Peter Ansell wrote:
----- "Marco Roos" <[EMAIL PROTECTED]> wrote:
From: "Marco Roos" <[EMAIL PROTECTED]>
To: "Matthias Samwald" <[EMAIL PROTECTED]>
Cc: "public-semweb-lifesci" <public-semweb-lifesci@w3.org>, [EMAIL PROTECTED],
[EMAIL PROTECTED]
Sent: Wednesday, 20 August, 2008 8:11:36 PM GMT +10:00 Brisbane
Subject: Re: Towards a cyberinfrastructure for the biological sciences:
progress, visions and challenges
Hi Matthias,
Interesting remarks. You'll be interested to know that 'webifying'
workflows (in various ways) is one of the prime interests of people at
myGrid and myExperiment. My personal opinion, a bit from a biologist's
perspective, is that we should be careful not to loose some aspects of
workflows in our enthusiasm about Web 2.0 and wiki approaches. One of
my main reasons for liking workflows is that they represent a
publishable and reproducible 'protocol' similar to a wet laboratory
protocol. This should not be volatile in any way, which Web2.0 or wiki
approaches easily can be.
Wiki's explicitly allow for a permanent link to a particular version of
something. Hopefully an implementation of a wiki-like workflow editor online,
will have similar characteristics so that you can still use a particular
version to reproduce a past result if you need to, provided the web services
still exist and haven't changed their interface ;-) It would also be nice to be
able to get corrected versions via the wiki mechanism though and that would
suit the Web 2.0 way, as opposed to publications to which corrections are hard
to make.
If some journals are requiring raw data (e.g., microarray data) to be
submitted to a public data repository, I wonder if workflows that are
used to analyze the data should also be submitted to a public workflow
repository.
Cheers,
-Kei