| Theme

This tile is from Mainquilters Unite!

Comment: Three Trees (The death of Ben and Andy)
Checked out at: April 29, 2001
Checked in at: April 29, 2001
Checkout tile:


-3 !?
-3 ?! At least, I think I did a decent blending (not like many people...)
don't worry......
.....i vote at your tile an............ 3, to put the score image to 0 ;)
Re: don't worry......
I really hate when people give my tiles negative scores and don't even bother to justify it with a comment. I got a negative five on a tile and nobody even bothered to explain. I'm not sure who's doing it, but it sucks. I dunno. I think the whole scoring system is kinda weak.
Re: Re: don't worry......
Well the voting system was never meant to be a quantitative measure of an artist. It accomplishes its goal of filtering artists into certain groups. That's why I added it. The problem is that there's no technical way to "fix" it. The values of the dots are all relative to the min/max scores on the quilts (i.e. if I made it go from -2 to 10, we'd want to change the main quilts to have a minimum score of -4, etc.). The real problem is with people who vote low but don't comment. Not a lot I can do about that, b/c I don't want to throw out votes. I think more votes is the only way to combat the problem, but I definitely see how people get frustrated when they see low scores. I could always just hide the avg and then just filter people without telling them why :)

Re: Re: Re: don't worry......
How about requiring a comment on all negative votes, or, to be fair, all votes? Or would that cause fewer votes which makes the scoring even more granular?
Re: Re: Re: Re: don't worry......
yeah, that's exactly my worry. There will be two outcomes of that:
1. we'll get comments that are meaningless
2. we'll get less votes, especially negative votes. There would be a "penalty" associated with casting a negative vote, namely, having to do extra work. While this is nice from a "everyone hug" standpoint, it destroys the point of the voting.

Additionally, it would identify the people who voted low. If we're going to identify votes, let's identify them all, not just some.

Really, from a pure logic standpoing, there is no reason for people to get upset when someone gives them a low score. To do so would imply an amazing level of ego (i.e "it's absolutely unbelievable that someone didn't like my art"). That's not to say that I wouldn't behave in the exact same way, I just mean from an objective, logical standpoint :)

But I truly think that reducing the number of votes will only make the problem worse.

Transparency vs. Opacity
Total transparency would probably discourage arbitrary voters, who give a -5 for reasons that may not be totally objective. If it's that inobviously bad, a comment truly would be helpful. I think most people get upset when they get a -5 for something they feel they deserved maybe a -2 or -1 for, or even a 1 or 2. They look at their tile and are not sure what makes it the worst it could possibly be. They then need to go through a pretty slow trial and error feedback loop 'okay is this one right, then? How about this one?' to see if they can improve.

I agree that it would also discourage voting in general, though.

This is going to sound like a really wrong, undemocratic, idea, but how about giving more weight to "qualified" voters? I.e.: someone who is already working on the main quilt may have more weight in their votes than someone lost in the score repair/newbie quilt, or a brand new account that's never done anything. With that in place, people could be, with their votes, answering the question: 'do I want to be doing tiles next to this person's tiles?' So a main quilter would have more weight on a score repair person, and an invite-only would have more weight on a main quilter. It would distribute moderation to the people who are already proven at being able to produce for their area. It would also mitigate abuses using new accounts.
Re: Transparency vs. Opacity
Yeah, this isn't a bad idea. I've thought about it before. I guess the trickiest part would be coming up with an algorithm that worked well. How do you weight it? By the number of tiles by that artist? By that artist's score? (then we get a fun circular dependancy where everyone's score depends on everyone else's, ugh :) We would definitely need to cache the score in the db, rather than calculating it on the fly (as we do now). But I agree that it could help things, and wouldn't significant reduce votes, it would even encourage people to get involved, so their votes would be worth more.

But don't hold your breath on this, I don't know when I'll have a chance to implement it. But now that the source is out there, someone could always send me a patch ;)

some tips.
You did blend well. Don't be discouraged by bad votes like that. (easier said than done though)
One tip/critique is that on the bottom of your trees there are dark pixels right along the border. Even if the person below you had followed through with your trees' colours, there would have been a visible border, where the colour changed all of a sudden. You should really try and follow your colours through, remembering that they need to continue on to the next tile.
And keep those tiles coming!
Re: some tips.
Yeah, you're right. I didn't pay attention to the lower edge at all. I'll try to don't make the same mistake. Thank you for the tip!!
Re: Re: some tips.
The #1 thing you should worry about when you are doing collaborative art (these quilts) is how your work is going to mesh in with the other artists. What I do is make a layer on top of everything else that SHOWS me the 15 pixel wide strips I'll be passing on to incomplete tiles done in the future... that way I can make SURE that they are getting something at at least I could come up with an interesting tile from.