No replies
Joined: 04/19/2009

Sorry for naming the last chat wrong - that one was Jan 14, this was Jan 15. Again, I have just deleted quits and joins:

[6:13pm] TimovdL: Let me kick off. I am very pro the scoring function changes, the game obviously needs those. But
[6:15pm] TimovdL: together with the new wiggle most current scripts will be broken as they all depend on the following algoritm:
[6:16pm] TimovdL: Do something that gets the score low, let wiggle repair the damage, evaluate and do something more
[6:16pm] TimovdL: And the new wiggle is too slow to do a good repair job in a decent amount of time.
[6:17pm] spvincent: Seth: are you there?
[6:18pm] SethCooperIRC: yeah
[6:18pm] BletchleyParkirc: Sounds like the solution would be to speed up wiggle ?
[6:18pm] auntdeen2: the general consensus among the players is that we would all like to be excited about the change, the accuracy… but would like the debut put off until the script issue is solved, and the largest bugs fixed (and we are still finding them) - and wiggle made faster so that all can still play
[6:18pm] BletchleyParkirc: and possibly the new scoring function's speed
[6:18pm] TimovdL: Or get a very rude faster version of wiggle, like it was about 6 months ago
[6:18pm] spvincent: Oh good. I was wondering if you had any further news for us based on yesterdays exchange of views.
[6:19pm] katfish joined the chat room.
[6:19pm] katfish was promoted to operator by ChanServ.
[6:19pm] SethCooperIRC: do you mean it takes too long to stop running, or takes too long to get to a decent score
[6:19pm] SethCooperIRC: ?
[6:19pm] SethCooperIRC: by stop running, I mean stop getting points
[6:20pm] BletchleyParkirc: Apparently both.
[6:20pm] TimovdL: It takes too long both ways, even the new version takes about half an hour to do one minicycle of my DRW on ED on a decent computer
[6:21pm] PixieVixie: As a noob...ish.. slow is no good..I mean unless you set up scripts for while your asleep you can pretty much say bye bye to a good score and as we know instant gratification is a new gamers biggest desire
[6:21pm] frood66: sorry - to but in ' the general concensus amongst the players'? where did this come from seth?
[6:21pm] TimovdL: And then the results are such that the wiggle didnt repair enough to get gains
[6:22pm] BletchleyParkirc: @frood, auntdeen said that.

[6:23pm] frood66: oops - she did
[6:23pm] auntdeen2: sorry frood - just my analysis of all convo over the past few days
[6:24pm] auntdeen2: feel free to disagree
[6:24pm] TimovdL: The biggest concern of the few that have been looking at this is that we might loose players

[6:24pm] frood66: and now I've read properly
[6:24pm] SethCooperIRC: well, we'll definitely continue to look at ways to improve the performance
[6:24pm] frood66: aunty has it
[6:25pm] NickyCGS joined the chat room.
[6:25pm] frood66: there is little doubt that we will lose players
[6:26pm] frood66: not at all aunty - my bad - went to bikesheds
[6:26pm] BletchleyParkirc: When is this new version to be introduced as the standard ?
[6:26pm] MikeCassidytoo: Tuesday
[6:26pm] frood66: tuesdayapparently
[6:26pm] TimovdL: But might it be possible to get a different kind of wiggle doing full rounds before this release is standard?

[6:26pm] auntdeen2: Seth - we really need to know why the rush - if there is a reason we can buy into, then we can buy into it
[6:27pm] BletchleyParkirc: Is there a new puzzle that absolutely requires this new version which needs to be released to the community ?
[6:27pm] frood66: I still have no idea why there is such a rush to roll this out
[6:27pm] drjr: brow suggested reimplementing the wiggle sccuracy slider
[6:27pm] silverberg left the chat room. (Read error: Connection reset by peer)
[6:27pm] SethCooperIRC: there will be an update posted very shortly with some fixes
[6:27pm] BletchleyParkirc: in other words, is there a science problem that requires the new version ?
[6:27pm] SethCooperIRC: yes, basically all the science problems we are interested in will be improved by this new version
[6:28pm] BletchleyParkirc: ok
[6:28pm] frood66: this is a massive change for foldit - why is it not being treated with some respect?
[6:28pm] SethCooperIRC: anything we do before posting it will essentially have to be redone after posting it
[6:28pm] BletchleyParkirc: ok
[6:28pm] BletchleyParkirc: I presume the new scoring function is far more complex than the old one ?

[6:28pm] BletchleyParkirc: (ths taking more time to calculate ?)
[6:28pm] SethCooperIRC: yes, it has a number of improvements over what we are using currently
[6:28pm] frood66: that's what the blog says
[6:28pm] TimovdL: But loosing players will not be reversable
[6:29pm] BletchleyParkirc: Has the wiggle function itself been changed or is wiggle's slow performance due to the new scoring function ?
[6:29pm] frood66: not sure foldit central are interested in losing players
[6:29pm] frood66: not sure they could care less
[6:30pm] TimovdL: The shake function has not degredated a lot, it is only wiggle where the effect is very big
[6:30pm] pauldunn: Is it going directly to main or is it going to devprev first?
[6:30pm] BletchleyParkirc: @frood, lets stay positive, we can solve problems by being creative.
[6:30pm] katfish: Hey everyone
[6:30pm] katfish: Sorry I'm a little late to the party
[6:30pm] drjr: Maybe revert to a steepest-descent optimization?
[6:30pm] auntdeen2: hi katfish
[6:30pm] katfish: And frood, I care!
[6:30pm] PixieVixie: hey
[6:30pm] silverbergirc: Ok, I'm beginning to get tired of this "losing players" talk.
[6:30pm] frood66: sstay positive while players say goodbye?
[6:30pm] BletchleyParkirc: yes
[6:31pm] katfish: So instead of talking about the negative, let's get constructive (for my sake! )
[6:31pm] frood66: wowee - is this now an elitist 'game'
[6:31pm] bobandpenny: 1/3 to 1/2 wiggles in a DRW cycle are 'make it worse by trying something', having those be precise or slow helps no one; maybe as Timo suggested a 'crude' or sloppy wiggle for those times and a better slow one when recovering score

[6:31pm] silverbergirc: In my opinion, this isn't about the wiggle, but the points.
[6:31pm] BletchleyParkirc: Has the wiggle function itself been changed or is wiggle's slow performance due to the new scoring function ?
[6:32pm] katfish: Hold up. I have a question for everyone. Forgive me if you talked about this before I got here, but in an ideal world, what would you like to see happen to make things better?
[6:32pm] auntdeen2: ummm…. it's not just negativity to talk about losing players - it's reality, some will no longer be able to play as the client stands now - and I have had players contact me saying that if the client is this slow, they are gone
[6:32pm] silverbergirc: brb, going to have a ciggie and calm down.
[6:33pm] TimovdL: Get a wiggle that is as slow/fast as shake with a full round
[6:33pm] BletchleyParkirc: so a performance increase without losing new functionality
[6:33pm] BletchleyParkirc: significant performance increase
[6:33pm] TimovdL: Yes
[6:33pm] frood66: I will repeat for those that did not see my opening comment....I think the changes are good for science. I do not understand the panic to introduce - just for a change can we get it right b4 rolling out?
[6:34pm] auntdeen2: katfish - in an ideal world - we would have been told about this when the devs first started working on it - and asked to help out with assurance that it wouldn't go online until all the big bugs gone - user possible for all computers...
[6:34pm] auntdeen2: at this stage, the only ideal is for all of us - devs & players to figure out the implications of the scripts broken
[6:35pm] katfish: For that I apologize and have to take responsibility
[6:35pm] drjr: I figure I'll just hand-fold, repair the cutpoints, idealize and use local wiggle to score high
[6:35pm] auntdeen2: and to fix wiggle so that all can still play
[6:35pm] katfish: I've been out of the country but the communication falls entirely to me
[6:35pm] auntdeen2: and that won't happen in 6 days
[6:35pm] frood66: I see nothing proactive with players so far
[6:35pm] PixieVixie: When progressing with something all thoughts should be explored regardless of feeling imho
[6:35pm] spvincent: Might it be possible to provide a mechanism to toggle on/off the most computationally expensive parts of the scoring function?
[6:36pm] katfish: So as I mentioned before, I'm a little late to the party but I have a suggestion
[6:36pm] katfish: I am going to remote in to our next team meeting (which is tomorrow)
[6:36pm] pauldunn: I'd like to be able to switch to centroid mode to speed up puzzles in the beginning stages and switch back for later stages.
[6:36pm] TimovdL: Or of the wiggle function, so that is targets the biggest protein problems first and rude
[6:36pm] frood66: no - u r not late - we have till tuesday
[6:36pm] bobandpenny: @spvincent; good suggestion
[6:37pm] katfish: and I'll discuss with everyone and see what we can do to lay out a solid schedule and gameplan
[6:37pm] katfish: So you are aware of what's going on
[6:37pm] SethCooperIRC: I don't think we can really turn off certain parts of the score function
[6:37pm] frood66: well that would be nice

[6:38pm] spvincent: oh well: it was just a thought
[6:38pm] katfish: Thanks for helping fill me in, everyone
[6:38pm] auntdeen2: Seth - despite the script and slowness problem, are you still planning on a Tuesday rollout?
[6:39pm] katfish: I'll have more details in my informational post, auntdeen
[6:39pm] katfish: Gimme a day to put it together

[6:39pm] TimovdL: In my measurements the degredating of shake performance is minimal, but wiggle is totally different
[6:39pm] frood66: honestly guys - most of this is nonsense.....all we need is a little cooperation
[6:39pm] auntdeen2: it might be good to leave the separate clients as is until at least the script issue is resolved
[6:39pm] SethCooperIRC: so far
[6:39pm] drjr: Local wiggle works, though
[6:39pm] katfish: I feel that, frood
[6:40pm] disturbedxxx666: it seems like this chat room has cought a katfish hehe
[6:40pm] katfish: hah!
[6:40pm] SethCooperIRC: well, in some sense, i think that we need to change the way that wiggle is scripted regardless
[6:40pm] katfish: Lively discussion is good bait
[6:40pm] disturbedxxx666: haha
[6:40pm] frood66: katfish: not sure u do,,PR here has been a disaster and all the work that that the devs have put in will be for nothing at this rate
[6:41pm] BletchleyParkirc: Why is that Seth ?
[6:41pm] katfish: For that I take full responsibility, frood
[6:41pm] katfish: and offer my apologies
[6:41pm] SethCooperIRC: the current meaning of an iteration as running to completion seems to be problematic for making changes to wiggle's behavior
[6:41pm] frood66: u can take all u like - makes no difference
[6:41pm] disturbedxxx666: cooper's here too?!
[6:42pm] auntdeen2: frood - katfish is in another country on different project for game center
[6:42pm] disturbedxxx666: did i miss something?
[6:42pm] SethCooperIRC: it caused problems with the wiggle precision before now, and now causes problems with the newchapter changes
[6:42pm] auntdeen2: she hasn't been involved with foldit in a while
[6:42pm] BletchleyParkirc: Seth, is wiggle so slow because of the scoring function ?
[6:42pm] spvincent: What kind of changes did you have in mind Seth?

[6:42pm] SethCooperIRC: basically what we have changed it to in newchapter

[6:43pm] SethCooperIRC: that an iteration is just a reasonably small unit of work
[6:43pm] frood66: anyways - I think this is going nowhere...so I'm gone. good luck with 'lastchapter' - really badly handled in my opinion
[6:43pm] disturbedxxx666: my newchapter works perfectly
[6:43pm] spvincent: I would really like to know whether I should make an effort starting now to change some of my scripts to the new vstandard or wait

[6:44pm] TimovdL: But do the devs realize that all scripts will no longer be useful?
[6:44pm] BletchleyParkirc: And I am typing to pinpoint if the slowness of wiggle is caused by the scoring function
[6:44pm] BletchleyParkirc: trying
[6:44pm] SethCooperIRC: I think it is a combination of two things
[6:44pm] BletchleyParkirc: because if it is, improving the scoring function performance will improve wiggle and all the work on changing scripts can be avoided.

[6:44pm] SethCooperIRC: one is that the scoring does take longer to do
[6:44pm] BletchleyParkirc: ok
[6:45pm] silverbergirc: back
[6:45pm] SethCooperIRC: i think the other is that it takes more "work" to stop getting points
[6:46pm] SethCooperIRC: in a sense, wiggle would be able to find more points than it was before
[6:46pm] SethCooperIRC: so it can run longer
[6:46pm] TimovdL: Is it possible to get the less precize button back?
[6:46pm] SethCooperIRC: we had also talked about adding a timeout to the scripting API
[6:47pm] disturbedxxx666: hasnt wiggle always been slow?
[6:47pm] TimovdL: It was always faster than shake
[6:48pm] disturbedxxx666: right...
[6:48pm] SethCooperIRC: so one could make sure that it didn't run longer than a given time
[6:49pm] BletchleyParkirc: right
[6:49pm] disturbedxxx666: oh thats interesting
[6:49pm] frood66: I just need one answer - why the rush to roll this out - it makes no sense to me right now
[6:49pm] disturbedxxx666: sorry, i just got here. trying to catch up
[6:50pm] auntdeen2: Seth - whatever you did in today's fixes seems to have fixed the Mac client issue I posted in feedback
[6:50pm] auntdeen2: I can close and reopen that client now
[6:50pm] silverbergirc: Could someone answer froods question?
[6:50pm] frood66: any chance of that?
[6:50pm] auntdeen2: I also asked yesterday

[6:50pm] auntdeen2: we do deserve an answer
[6:50pm] frood66: I notice a certain reticence
[6:50pm] SethCooperIRC: basically all the science problems we are interested in will be improved by this new version
[6:51pm] frood66: still my question is unanswered
[6:51pm] auntdeen2: Seth - there are HUGE issues with this
[6:51pm] SethCooperIRC: anything we do before posting it will essentially have to be redone after posting it
[6:51pm] auntdeen2: please be honest with us - what is the rush
[6:51pm] SimplySue: (for thoses of us who have recently joined: can you give us the basics Pros and Cons of this debate?) Please....
[6:51pm] frood66: and I'm getting less happy - surely it can't be difficult
[6:52pm] disturbedxxx666: oh wait! Isn't foltit getting an award this monday at the science awards?
[6:52pm] frood66: now I'm really not happy at all
[6:52pm] frood66: and I now many are also looking for the same answer
[6:52pm] auntdeen2: and / or is there a paper deadline as mentioned in the original post?
[6:52pm] katfish: Production schedules rely on a lot of different factors here, and science puzzles are a big part of that
[6:52pm] frood66: so now's the chance

[6:53pm] drjr: Tactically, dramatic change implemented fully and simulataneously has advantages
[6:53pm] SimplySue: wait....stop.... let the dev's type...
[6:53pm] frood66: u haver got to be kidding
[6:53pm] katfish: The thinking on rolling it out is to help get the very best results for upcoming puzzles
[6:53pm] TimovdL: We all get that and i am all for improving the game and would like to see the game experience stay nice
[6:53pm] katfish: However I completely understand that there are still concerns
[6:53pm] frood66: and the question is STILL not answered
[6:53pm] auntdeen2: we all want what it is bringing - AFTER the issues have been sorted
[6:53pm] katfish: Frood, can you help me out? Not sure how I'm not answering the question
[6:53pm] frood66: answer the question - it's not hard
[6:54pm] SimplySue: which will be in the younger members lifetime?
[6:54pm] auntdeen2: kat, seth - why this tuesday instead of next?
[6:54pm] SimplySue: sorry.

[6:54pm] frood66: no katfish - I can't - I'm just asking the question that others will not
[6:54pm] auntdeen2: why not wait until we have the script fixes and speed fixes?
[6:55pm] disturbedxxx666: is it really that big aof a deal guys?
[6:55pm] frood66: now - an answer will probably satisfy many
[6:55pm] auntdeen2: *deep breaths*
[6:55pm] frood66: ok - I'll back off
[6:55pm] disturbedxxx666: the devs can do what ever they want
[6:55pm] auntdeen2: frood, we've dominated chat with a question that will apparently not be answered, so be it
[6:55pm] katfish: There are upcoming puzzles that will benefit from the change, as I mentioned before I will put together a post explaining in a little more detail once I remote in tomorrow
[6:55pm] frood66: but we have no answer - read into that as u will
[6:56pm] SethCooperIRC: well, we now we're going to switch to the new scoring function
[6:56pm] SimplySue: If the scripts don't work, the roll out won't work. Period...End of Comment.
[6:56pm] drjr: Local wiggle works well
[6:56pm] frood66: the answer is clear now - thx for that katfish
[6:57pm] TimovdL: Same with any question about alternatives for the current wiggle functionality
[6:57pm] disturbedxxx666: scripts work fine for me though and im using newchapter
[6:57pm] SethCooperIRC: so, we'd like to get to be able to run science puzzles on this as soon as we can, rather than continuing to use a scoring function that we know to be inaccurate

[6:58pm] auntdeen2: sigh
[6:58pm] auntdeen2: soon is good
[6:58pm] frood66: these would be players with big machines?
[6:58pm] pauldunn: will the new puzzles be reasonably small at first while the bugs are worked out?
[6:58pm] auntdeen2: but working properly is better
[6:58pm] disturbedxxx666: cooper has his reaons

[6:59pm] TimovdL: One other issue: loading the ED puzzle in newchapter takes minutes, very very slow
[6:59pm] SethCooperIRC: for the most part, things that we do now with the less accurate scoring function, we'll need to redo again with the improved one, and there is only so much time before CASP
[6:59pm] auntdeen2: ...
[6:59pm] auntdeen2: we of course don't want you to have to redo anything
[7:00pm] auntdeen2: of course, we will be redoing thousands of scripts...
[7:00pm] frood66: well - u sort of have to have guys here for casp in the first place
[7:00pm] SimplySue: so, we will re-do the re-done puzzles again...
[7:00pm] auntdeen2: thanks, and I have reached my limit for the day, afk

[7:00pm] disturbedxxx666: i get it! seth, ya'll want to fix all of these issures before CASP, right?
[7:00pm] SethCooperIRC: it's not just us redoing things, but players as well
[7:00pm] frood66: and....to be fair - casp only came into it when it was mentioned by a player last night
[7:01pm] frood66: seth - I'm affraid u r being disingenuous
[7:01pm] SethCooperIRC: we want the science to be as accurate as possible so that what everyone is doing is most useful
[7:01pm] disturbedxxx666: that makes since
[7:01pm] SimplySue: afk
[7:02pm] frood66: absolutley - I can grasp that - but u have to grasp the other half - FUN

Sitemap

Developed by: UW Center for Game Science, UW Institute for Protein Design, Northeastern University, Vanderbilt University Meiler Lab, UC Davis
Supported by: DARPA, NSF, NIH, HHMI, Amazon, Microsoft, Adobe, RosettaCommons