[kepler-dev] SEEK 101
rods at ku.edu
Fri Jun 11 07:23:05 PDT 2004
I think it would help to start to understand how the "average Joe"
scientist would approach the SEEK technology. Here is my initial take on
this, some of which we briefly discussed in out Kepler Santa Barbara
The first question a scientist may ask is "What is the /value add/ of SEEK?"
(If I take the time and effort to import my model and data into SEEK
what more will I get out of it?)
The answer may be "Enabling your modeling to access additional datasets
within the SEEK 'grid'"
(Hopefully their are additional compelling answers to this question.)
The initial steps:
1) Install SEEK
2) Possibly run an existing "sample" model that is somewhat related to
the work they do
3) Input an existing model they have
a) What is the model written in?
b) How easy is it to import?
c) How can they feel comfortable that what they have in SEEK
represents what their model is?
4) Import some or all of their existing data
a) What format is the data in (text, spreadsheet, MS-Access, etc)?
b) How do they get help importing their existing data?
c) How much describing of their data do they have to do before they
can use it in SEEK?
5) Running their "new" model and verifying the results?
a) (Insert "wow factor" here)
6) Now, how do I integrate new and additional datasets to my existing model
There is also a different "use case" which is the one where a scientist
has been exposed to SEEK and wants to start from scratch and enter a new
model and access both their own data and other data.
We need to continue to flush this out with more and more concrete steps
with increasing levels of detail on how to make someone's first
experience with SEEK a successful one. Then we will know what tools are
needed on the front end.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Kepler-dev