Is not turning in fah work assignments bad?

[bufke]

I'm trying to get permission to set FAH up on some university computers, however they have deep freeze installed and thus after a restart anything new on the hard drive (like fah data) is deleted. Assuming the computer is restarted at least once per day if not many times, the client would then fail to turn in assignments every day. So the question is does this matter?

[7im]

Welcome to the forum.

Yes. Not turning in work units hurts the project. Any work unit that is lost will not be reassigned to another person until the work unit expires after several weeks. That unnecessarily delays the science of the project from moving forward.

Run the fah client from a thumb drive so the data is not lost after a restart? Run the client from your personal folder on the network.... surely they provide space to save files on a network drive somewhere and don't wipe that out each day?

[codysluder]

If the data is really discarded every day, nothing would ever be completed. Most WUs assigned take more than one day to finish.

[DamianJansen]

Should there not be something that a user can do to 're-assign' their work units? i.e. log in and either download those units, or abort work on them. Or better yet, flag them as failed so that anyone can download them?

[7im]

That's a good suggestion, but the infrastructure to enable such a feature would take a lot work and rework to support it. And to not much benefit. Yes, a few work units are lost/deleted/ect. but very small numbers in the big picture, and fah already has a method to get those few missing work units back out to be folded.

If something isn't really broken, they prefer to concentrate on the science of the project, and on newer clients that can make very large improvements in performance. The ability to report a few lost work units is a bit further down on the priority list with their limited development resources, and the small improvements it makes.

P.S. Welcome to the Folding Forum, DamianJansen.

[DamianJansen]

Thanks for the welcome. I did Seti@home once upon a time, but this seems more... significant I guess.

I have no idea what sort of infrastructure they have, or what it would take, so I won't comment on that other than it's a bit of a surprise. However, being a software QA engineer myself I agree entirely with "if it ain't broke, don't 'fix' it".

[7im]

I started out on SETI I myself, and I agree that F@h is more significant. The time frame for return on the investment is much much quicker with fah than waiting to find a signal from ET, and then returning a signal, and then waiting for them or us to meet. Then again, some seem to think ETs are already here.

Keep the suggestions coming, as Pande Group will occasionally incorporate these ideas in to the clients while doing other major updates.