Hi Sarah,

There were a couple of caveats with regards to the completeness of the usage statistics. In particular, it is possible for users to opt-out of sending usage statistics, so we may not get a full picture of everything which is used. There seems to be two broad themes why people might want to opt-out:

1. They want to keep privacy. However, it’s worth noting the usage statistics we collect are completely anonymous and contain no personal data. It’s essentially just an increment to the particular feature that was used. We should then ask, why give the option to opt-out?

2. They are concerned reporting usage will cause a slowdown in their workflow (as it sends a usage report each time an algorithm is

executed). At SNS, they turn off usage statistics by default for this reason. A potential solution would be to only report usage statistics upon exiting Mantid, so usage statistics are only sent once per Mantid session.

If we can address these two issues, I think the statistics from the usage service would work really well for our algorithm deprecation policy. The statistics from the usage service would also perhaps be a really useful way to quantify what is used/isn’t used at our MPB meetings.

Kind Regards,

Rob

To your list of present communication routes, Sarah, I would add the Science Advisory Forum.

Other fora that might be considered are the IDG Meeting (hits a cross-section of inst scientists), the Neutron Users Meeting (hits a subset of UK users), and an additional popup window when you first run a new version (hits all users) saying ‘The following algorithms are deprecated in this release and are due to be removed in vx.x.x. If you have any concerns please contact…’?

Steve

We having been working on drafting a policy with SNS whilst we have been out there. It will be based on usage data but there are some caveats I believe. Rob was working in the group looking at the usage data.

Rob – what was the outcome of that session?