As we all learned the points awarded for finishing a WU are based on a number of technical parameter like base credit, TPF, timelines, konstant factors with an exponential curve.
While I technically understand the parameter and how they influence the curve I wonder how are those determined ? In some discussions the famous word of "scientific value" came up... what is that ? How is it defined or measured ? How can it be compared between projects and cores ? What's in the base credit ?
Note: cross posted on Reddit: https://www.reddit.com/r/foldingathome/ ... fic_value/
What is the definition of "scientific value" ?
Moderators: Site Moderators, FAHC Science Team
-
ChristianVirtual
- Posts: 1576
- Joined: Tue May 28, 2013 12:14 pm
- Location: Tokyo
Re: What is the definition of "scientific value" ?
Some theory surrounding those concepts:
There is a certain amount of scientific value associated with the completion of a WU from a project (and it has to be completed before it expires). Base value represents that achievement. (Very) Roughly speaking, it represents the number of FloatingPoint Operations required to complete the WU.
The time-value of the WU is also part of the scientific value. A WU that's completed quickly is worth more than the same WU that's delayed until just before it expires. The bonus points which are added to the base value represent that scientific value.
How does a change in the time-value of a WU compare to a change in its normalized number of FLOPs? That's a question I can't answer.
In practical terms, the project owner(s) are responsible for setting those values associated with the bonus points formulae, but I don't know how/why they decide.
There is a certain amount of scientific value associated with the completion of a WU from a project (and it has to be completed before it expires). Base value represents that achievement. (Very) Roughly speaking, it represents the number of FloatingPoint Operations required to complete the WU.
The time-value of the WU is also part of the scientific value. A WU that's completed quickly is worth more than the same WU that's delayed until just before it expires. The bonus points which are added to the base value represent that scientific value.
How does a change in the time-value of a WU compare to a change in its normalized number of FLOPs? That's a question I can't answer.
In practical terms, the project owner(s) are responsible for setting those values associated with the bonus points formulae, but I don't know how/why they decide.