Citizen Science Grid

[Bok]

So, I'm in two (or more) minds on what to do with CSG now it has begun exporting stats of it's own with subproject data consisting of the original projects - SubsetSum@Home, DNA@home & Wildlife@Home.

Do I

a) add it and remove the older projects (meaning we'd all lose some MM's).
b) add it and leave the older projects also (meaning we'd count the credit twice as the total credit for CSG is the combined of the other 3)
c) don't add it
d) something else?

I *think* the best option is a) but I still don't like it...

[Maxwell]

My memory of the original intention of this was to have one centralized place to issue WUs for three separate projects that already exist. Which means my initial thought on this would be not to have a project of CSG, but issue updates to the other three projects. I don't know how this would work for people who don't have credits in the old projects, however.

Another thought would be to include CSG as a new project, but subtract out the credits for the old projects, which would only count new CSG credits toward MMs. So:

[CSG New Credits] = [CSG Export Credits] - [(Current Subset Credits) + (Current DNA Credits) + (Current Wildlife Credits)]

The more I look at what I typed, the more I like this adjusted option. It keeps all MMs intact, gives us a new one to shoot for, preserves rankings on the previous projects, etc.

[STE\/E]

My memory of the original intention of this was to have one centralized place to issue WUs for three separate projects that already exist. Which means my initial thought on this would be not to have a project of CSG, but issue updates to the other three projects. I don't know how this would work for people who don't have credits in the old projects, however.

Another thought would be to include CSG as a new project, but subtract out the credits for the old projects, which would only count new CSG credits toward MMs. So:

[CSG New Credits] = [CSG Export Credits] - [(Current Subset Credits) + (Current DNA Credits) + (Current Wildlife Credits)]

The more I look at what I typed, the more I like this adjusted option. It keeps all MMs intact, gives us a new one to shoot for, preserves rankings on the previous projects, etc.

Agree ...

[Bok]

That's pretty inefficient though. No efficient way to do it within the database when I do the calculations for combined credit.

That's already the most costly sql calculation of all and to do this would mean something like

sum where proj NOT IN ('wil','sub','dna')..

or sum all, then subtract for each of those projects by cpid.

yuk to the yuk.

[Bok]

One other thought...

I *could* include subproject data in the MM count for those projects which support it...That should be feasible, though it's not trivial the way the data is held. Could still include the main project in the count also.

That way I could deprecate the other projects, but the MM count's would not change (or might even increase..)

[Fire$torm]

What exactly is the project doing about the credit mess?

Oh, just had a thought. If the "New" project is going to have the stats merged, old into new, then how about appending the old names to the new names. Ex. New project name@csg [Original Project Name].

This way you don't have to muck around with a ton of formulas. Old MMs should be "Frozen" and allow the new MMs to move along.

[Bok]

I think I've pretty much decided to retire the older 3 projects which will have an effect on a number of the MM categories. It's affected across the board so it should be minimal. Already added the stats for csg so right now they are counting double for credits for those that had data in wildlife, subset and dna@home.

I tinkered with adding the various levels of the subproject data into the MM categories but it just wasn't efficient at all. So I thought, WTH, just create new MM categories for subprojects standalone. I've done 10k, 100k and 1M, others range from trivial to quite a bit of work.

Biggest issue right now is the amount of real estate the MM list will take up on some screens, might just have to list top 15 or so and link to a new page with full list. Probably only affects a 100 users anyway.

I know removing the older data isn't the ideal solution but this is a one off merge of those projects and I just don;t see any other practical solution. Data will be archived so I could still show it in some way, but not in amongst the current projects.

[Maxwell]

How is this going to affect people who have not linked their accounts at CSG? For example, I know Mumps has tons of credits in some of the projects (DNA, Wildlife, Subset), but hasn't linked his account, so he's got zero CSG credits.

On the subproject stats, I love it. I think it's awesome. Personally, I'd love to see it for all the other MM categories eventually, but I'm a stats whore myself. If I had to pick one other category, it would be 500k, as that tends to follow the badge structure of most projects.

As far as which MM categories to display, my initial thought would be to display only the numbers and ranks ones - 1k, 10k, 25k, 50k, 100k, 250k, 500k, 750k (which, IMHO, could be dropped), 1M, 2.5M, 5M, 10M, 25M, 50M, 100M, 250M, 500M, 1B, Top10, Top100, and. Top 1000. A link to a separate MM page with all the other fun ones would be perfectly workable by me.

And (dream, I know) while we're on the subject, I'd still love to see the WUProp hours MM stats (e.g., Most 100h apps, most 250h apps). Perhaps on the new MM category page?

I'll stop. I'm not obsessed or anything... :p

[Bryan]

As far as which MM categories to display, my initial thought would be to display only the numbers and ranks ones - 1k, 10k, 25k, 50k, 100k, 250k, 500k, 750k (which, IMHO, could be dropped), 1M, 2.5M, 5M, 10M, 25M, 50M, 100M, 250M, 500M, 1B, Top10, Top100, and. Top 1000. A link to a separate MM page with all the other fun ones would be perfectly workable by me.



+1

[Bok]

I did have some thoughts on the linking and was going to ask travis. Not sure why it could not be done automatically, surely it's just link a team wide import where it creates the founders accounts too. At least I don't see it as any different.

I could easily mine the data out of my tables for him if needed but he should already have it.

And for MM's yes, I agree with doing most of them with subprojects too and the work I have just done on the table structures enables it as well as the wuprop hours side (though I still need more work on that bit).

*EDIT* http://volunteer.cs.und.edu/csg/forum_thread.php?id=1903&postid=4645

[Mumps]

How is this going to affect people who have not linked their accounts at CSG? For example, I know Mumps has tons of credits in some of the projects (DNA, Wildlife, Subset), but hasn't linked his account, so he's got zero CSG credits.
Fine. Whine and snivel for me and I get the point... :D

DNA and Subset accounts linked. Maybe I should get back on that pair and reclaim #1 spots... :)