Re: Sprint task: Refine the karma system (community input requested)
 

They might have value if karma decay were introduced. Maybe.

If we had a karma system that allowed karma from specific sources to decay, and we made t.m.o posts[1] one of those "decaying sources" due the posts' here-and-gone nature, then it might be worthwhile to recognize certain posts as being more "timeless"[2] in nature. These would be the posts that attract so many thanks that their value to the community must be greater than would be reflected by the normal waxing and waning of post and thanks karma. The "gold post" karma bonus would boost the peak from that contribution a bit higher and therefore, longer.

Obviously, this is all theoretical, and I don't completely buy into it myself, but it's the best argument for "gold posts" I've been able to concoct.

[1] Blog posts would likely be good candidates for this process also. Mailing list posts too, if only there were some mechanism by which they could be rated.
[2] As in "maintaining usefulness over a long period of time", not "written by that Mozilla developer guy".

Re: Sprint task: Refine the karma system (community input requested)
 

These are now implemented and waiting for X-Fade to roll them out to maemo.org. Then recalculating everybody's karma will take a couple of days to roll through.

Expect lower but hopefully a lot more balanced karma scores :cool:

Re: Sprint task: Refine the karma system (community input requested)
 

Thanks to everyone involved for making this happen.

On a related yet bit-of-a-side note, would it be possible to finally get some explanations regarding what the very obscure categories are? I have noticed that a lot of people in the Community are only involved in a single kind of activity (for example, in my case I'm mostly involved in TMO and IRC. Very few bugzilla, blog, wiki or brainstorm points).

Maybe we could try to lure those kind of profiles into more widespread activities? I don't know how, but if karma is going to be useful, this could definitely be one of the routes it might take: Now that karma is more balanced, an inbalance in a user's profile will be a lot easier to spot.

Another request I would have is in the high-score , could we a/ make the categories more explicit, and b/ add a thanks/posts one?

Re: Sprint task: Refine the karma system (community input requested)
 

I've never been so happy to lose so much karma. How refreshing.

Re: Sprint task: Refine the karma system (community input requested)
 

Agreed, that one should be improved on. Now the names come from the names of the karma modules that are used.

Quick explanations:

• Comments: comments to Extras Testing packages, Maemo Downloads apps or on-site announcements
• Favourites: giving thumbs up to a news item or brainstorm solution (thumbs down gives karma too)
• Blogs: blog entries based on their rating in the Social News
• Products: apps in Maemo Downloads based on their star rating
• Discussion: posts to maemo mailing lists that are imported to maemo.org community section
• Bugzilla_reported: new bugs reported to Maemo Bugzilla
• Bugzilla_comments: bug comments on Maemo Bugzilla
• Brainstorm: ideas and solutions being voted on or implemented
• Packagetesting: participating in the Extras QA process with thumbs up/down
• Itt_thanks: thanks received here on Talk
• Itt_posts: posts here on Talk
• Mediawiki_edits: editing the Maemo Wiki
• Groups: membership in Garage projects

Re: Sprint task: Refine the karma system (community input requested)
 

Darn, this version is even more embarrassing than the last. :( The only consolation is that I'm no longer on page 3.

Re: Sprint task: Refine the karma system (community input requested)
 

Holy cow! Why am I ranked so high? Is it a bug?

Re: Sprint task: Refine the karma system (community input requested)
 

Is this why my karma has suddenly jumped by almost 100 overnight ? I'd been wondering :-)

I think the fractional numbers are a bit overdone however:
Favourites 1.62019
Comments 9.79796

Maybe two decimals would be enough ? :-)

Re: Sprint task: Refine the karma system (community input requested)
 

I'm happy to see you lose it too!

See how much of a buddy I can be. :D

Re: Sprint task: Refine the karma system (community input requested)
 

I'm pretty sure that nobody lost as much as I did: 1646 of hard earned karma points! I demand a recount! :D

Re: Sprint task: Refine the karma system (community input requested)
 

Hey, it's so bad that thp pulled back ahead of me! And thoughtfix, who is off on other things, is back ahead of qole!

Re: Sprint task: Refine the karma system (community input requested)
 

Apparently the new scoring scheme really really likes bloggers, and since your bug jars count as blogs... :D

Re: Sprint task: Refine the karma system (community input requested)
 

Sheesh, and just this morning I was explaining to someone that I don't have a blog.

Well, I'm sure this will touch off the next round of karma formula refinements. There's no way my name can be left there above those who do real work.

Re: Sprint task: Refine the karma system (community input requested)
 

Oh quit with the humility. Your bug reporting is extremely valuable, and if it isn't automated, is something that would make me nuts doing.

But back to...

I do love the irony. :D

Re: Sprint task: Refine the karma system (community input requested)
 

Ouch, man, that stings. Bugzilla sure took a beating. Not in love with the fact that the top 10 Bugzilla contributors in bug reports and comments earned only 1188 karma between them. While the top 10 bloggers have earned 7249 karma between them in a single category, and the top 10 Talk contributors in Thanks! and posts have earned 5601 between them.

Oh well.

Re: Sprint task: Refine the karma system (community input requested)
 

Well, that's a bug among others it seems, we didn't proposed any alteration to bugzilla karma, nor for favs that is also very low.

Also there was a bug with the blogs karma, we reduced it (now max.10p) and guess what my blog karma is suddenly 3 times more :( oh my.

Perhaps the blogs metrics need to be revised :D

Re: Sprint task: Refine the karma system (community input requested)
 

Here's another one that ought to be revised, and why not completely removed: Groups!

The first one in that list has 162 group karma points, and those are all of his karma points.

Also, if you look at jakunnar's profile, you'll notice that he has 0.7 karma points from favourites, however they're not counted in his total...

Not that I'm asking karma points be revised to take tenths and hundredths in the final number, but 0.7 is closer to one than it is to 0, so why not 163 karma points?

If it is so easy to get karma points through groups, I'm going to start making a list and request to join every single group on garage... If not, what's the interest? It doesn't count contributions, just membership, which is worthless; so why not remove it outright?

Re: Sprint task: Refine the karma system (community input requested)
 

I totaly agree with you. The bugzilla contributions seem to count much less and other forms of contribution are counted to much.

If I have a look into my actual karma count I see:
82 point for t.m.o thanks
89 points for mediawiki edits
64 point for bugs reported in bugzilla

If I ask myself how I invest my spare time into mamo.org I would answer I am using 70% of time to report and handle bugs and I also invest much more time into t.m.o than into edit pages in the wiki.

Beside of this I think a wiki edit is more worth than a t.m.o.

All in all bug reporting / triaging should bring more points. I don't need much karma points but there should be a signal to everyone who is willing to invest time into th maemo community: "Reporting bugs is important and a good way to get karma points." At the moment the signal is: "Post as much as you can to t.m.o."

Just my 2 cent...err karma points :)

Re: Sprint task: Refine the karma system (community input requested)
 

Please, this is a bug as I said no need to make a storm around that, I can't fix it otherwise would be fixed for sure:).

Re: Sprint task: Refine the karma system (community input requested)
 

Ooops, you are right and you already wrote it. I'll write to the blackboard for hundred times: "I will read all posts following the post I'll answer to before I write an answer!" :)

Ciao jukey

Re: Sprint task: Refine the karma system (community input requested)
 

Don't forget to post the video when done. :p

Re: Sprint task: Refine the karma system (community input requested)
 

we should just forget karma and go for high scores on n900fly

Re: Sprint task: Refine the karma system (community input requested)
 

In general any karma item where the top karma receivers have more than 300 or 400 karma from that category sounds quite unbalanced. Well, maybe except products (coming from popular apps published by the user to Maemo Downloads).

This is why we did a little bit of tweaking to bug reporting and thumbs karma, in order to ensure they don't overwhelm the other karma categories.

Blog karma probably deserves the same.

Re: Sprint task: Refine the karma system (community input requested)
 

I wish more people would understand that things like karma are always open to process improvement. No reason to get all balled up when a certain metric looks out of whack... just participate in the corrective process. ;)

Re: Sprint task: Refine the karma system (community input requested)
 

And now bloggers are about to be nerfed.

Re: Sprint task: Refine the karma system (community input requested)
 

I agree in part with that, but bug reports are IMHO the second most important contribution from the community (#1 apps in extras), and now if I get it right, a bug report worth the same than a wiki edit(6*sqrt(# edits)) and less than a TMO thanks(8 * sqrt(#thanks)), IMHO a bit wrong.

Re: Sprint task: Refine the karma system (community input requested)
 

With bug reports it would be best if we could include bug validity into calculations. Otherwise you get karma for posting duplicates and completely beside the point bug reports :rolleyes:

Anyway, happy to tune the karma algorithms as needed. The current ones are already a lot more balanced than what we had, but especially blog karma could use improvement.

Re: Sprint task: Refine the karma system (community input requested)
 

That would be awesome, not knowing the system, don't know if this would be doable.

Re: Sprint task: Refine the karma system (community input requested)
 

Currently we get a CSV dump of bugzilla accounts and their bug creation and comment numbers. The script that generates the dump could possibly be tuned to ignore invalid bug items

Re: Sprint task: Refine the karma system (community input requested)
 

Just as long as it pops that overinflated Gadsby back down to size. :D

/me ducks and runs

Re: Sprint task: Refine the karma system (community input requested)
 

About products karma. There was a Brainstorm (I think) about apps getting karma based on the amount of downloads. Has this been implemented? Or was this rejected?

Re: Sprint task: Refine the karma system (community input requested)
 

Invalid and duplicates would be great.

Re: Sprint task: Refine the karma system (community input requested)
 

Just as long as it's only DUPLICATE reports. INVALID resolutions are too often more "Nokia is stupid" rather than actually INVALID.

Re: Sprint task: Refine the karma system (community input requested)
 

That is now implemented for Maemo 5 apps. And app karma has a decaying algorithm, so to stay on top you need to consistently get new downloads and reviews.

However, user karma for "products" doesn't take this into account because of the decaying component. It only looks at the stars the app has received.

Re: Sprint task: Refine the karma system (community input requested)
 

I want to thank that but you have enough encouragement. :p

Re: Sprint task: Refine the karma system (community input requested)
 

Oh come on.. you have to see the benefit of giving him Karma for a post joking about removing Karma.

It's just all around the perfect post to thank :D.

Re: Sprint task: Refine the karma system (community input requested)
 

Remembering some old game-designer-days... How about doing this top-down ? I mean, we can set up a scoring scheme and set on to discover who is the big winner in the new scenario, but maybe it would be easier to go the other way round. Say there is a karma pool for an area (wiki, blogs, etc), and then split that karma across participants. Their internal relations will not change (more karma is still more karma), and karma inflation will be more easy to control. This would also have the side effect of moving karma whores to areas that have less people working on them (=better karma density). It's also more indicative then a general number and arguably reflects 'activity' better. Obviously to maintain real value and prevent deflation, you need to accomodate the total karma in accordance with the GCP (gross community product), for example scaling the karma pool with the number of (active) members on talk, number of applications in extras, etc. I also like achievements-style notices (Ubuntu does this, too) like 'Extras general (10+ apps in extras)', 'Bugsquad lt. (25+ bugs reported/triaged)', 'Wiki sage (100+ wiki edits)', 'Forum master (1000+ thanks)' both for (not condescending) bragging rights and getting to know what people actually DO in a fun manner without digging through karma tables. Okay, so, that would be the completely out there off-topic gaming approach many will dismiss off-hand :) , but now, seriously, a step back to gain a little top-down insight could be useful in this case to clarify (again) what goals we want to reach with the (approximative metric) of karma.

PS. Also, it's hard to define the usefulness of karma. Yes, sure, we see the destructive effects as they tend to be more noticeable (like when the N900 DDP program turned the karma game into a deadly serious device contest). On the other hand, the small little plusses that karma motivation gives, while present, will hardly manifest directly, making a final conclusion difficult.

Re: Sprint task: Refine the karma system (community input requested)
 

Right, so it wasn't a bug after all but intentional, and done without community consensus or even debate :-(

While I don't see karma as very useful, as long as it's there having such a low value for bugzilla activity (does this look right to anyone?) sends completely the wrong message.

To everyone who filed bug reports and especially what's left of the bugsquad: please don't be discouraged, your contributions are valued regardless of what some pointless vanity number in your profile page says.

Re: Sprint task: Refine the karma system (community input requested)
 

It would be nice to see how karma is calculated. This wiki page seems to have old info.

I would like to see that the most valuable contributions would get most karma. Personally I feel this should be the order of value (some of them in shared position):

1. Products
2. Bugzilla, Package testing
3. Wiki
4. Blogs, Brainstorm
5. Comments
6. Groups, Discussion
7. Talk thanks
8. Talk posts
9. Favourites

Re: Sprint task: Refine the karma system (community input requested)
 

Updated: http://wiki.maemo.org/Karma#Weights