Hi all,
On 2018-12-06 08:54, Ian Collier - UKRI STFC wrote:
>
> Again, tarballs are very easy to unpack onto cvmfs.
True enough, but the best people to tell us whether CVMFS (or anything
else) is always right, sometimes right, or never right is the VO who
uses it. And the best thing we can do in my view is to capture and
summarise a sufficient number of VO "operational baselines". We can then
use these to illustrate to prospective customers:
a) the application areas we are involved with;
b) the coverage, applicability and interoperability of our various
technologies to specific problem domains; and
c) the trials and tribulations of actually getting something running on
our stuff.
That would help prospective customers to set their expectations and
make sensible choices. In my view, such a summary should contain a basic
description of the science goals (the mission statement); a brief
summary of data capture, storage, workflows, transforms and so forth
(the task). Then a description of the technology selected (including,
for example CVMFS), and a rationale for choosing (or not choosing) it
(the plot). Then we need to show how well the solution performed in
action, the various hitches that had to be overcome and an honest
critique of the final user experience (the review).
If we had material like that, gathered soon after a VO reaches
"operational maturity", we would be well placed to simplify the
integration of new groups as they come along. And it would would, to
some extent, ameliorate subjectivity about the benefits of one choice
over another.
Cheers again,
Ste
########################################################################
To unsubscribe from the TB-SUPPORT list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=TB-SUPPORT&A=1
|