2014.11.26 Dev Chat
( Link back to Dev Chat archive )
On Lab UI, S/N, Puzzle Next, & More
jnicol: Welcome to dev chat! [3:00 PM]
Eli Fisker: Thx :) [3:00 PM]
machinelves: thank you John! [3:00 PM]
LFP6: COuld we perhaps get an update on what's being worked on in dev land? [3:00 PM]
LFP6: Application/website wise [3:00 PM]
jnicol: I promoted the Binary coding lab to be active [3:01 PM]
LFP6: Ooh [3:01 PM]
jnicol: Restrictions on 131 bases for labs has been relaxed [3:01 PM]
Eli Fisker: Oh, what have happened? [3:01 PM]
codygeary: oh? how many do we get now? [3:01 PM]
jnicol: Rhiju and Ann have a process ready for the Jan labs to use ePCR, which removes the padded sequences from interferring [3:02 PM]
jnicol: Hi Cody, we were requiring labs to 130/131 bases so padding was not necessary [3:03 PM]
Eli Fisker: Sweet [3:03 PM]
codygeary: 'padding', is that the 3' tail sequence? [3:03 PM]
Eli Fisker: Padding could be attached at both ends [3:03 PM]
Eli Fisker: And we don't see it ingame [3:04 PM]
Brourd: I would suggest that the unpaired 5' residues be trimmed from the secondary structures in Eli's project then. [3:04 PM]
jnicol: we add random bases to pad each sequence to 150 bases which includes the promoter [3:04 PM]
codygeary: so how much larger can we design and still have room for the barcode? [3:04 PM]
Eli Fisker: I would only be happy if I could get rid of the 5' tail sequence [3:05 PM]
Eli Fisker: As it makes the data less comparable [3:05 PM]
jnicol: The maximum length has not changed [3:05 PM]
jnicol: there are 25 submissions already for your labs Eli [3:06 PM]
Eli Fisker: Yes, so that will be unfair to the players submitting [3:06 PM]
Brourd: I'm sure those players will not mind. [3:06 PM]
codygeary: I want to run a new round of KL labs, but I'd like to allow players more than 3 slots per design, is that possible? [3:06 PM]
Eli Fisker: So I change for next lab [3:07 PM]
hoglahoo: Eli whatever is fair to science [3:07 PM]
hoglahoo: we are but lab rats [3:07 PM]
jnicol: Not sure if I can do it for just 1 lab, but I'll check to see [3:07 PM]
Brourd: Bad idea, Eli. But oh well. [3:07 PM]
machinelves: by the way as a general announcement, i noticed many players saying that a large quantity of sublabs made it difficult to get through them all, so maybe we could have a limit on total sublabs per lab? [3:08 PM]
hoglahoo: Also, it appears that Daniel B has kidnapped LinkBot. An investigation is underway [3:08 PM]
LFP6: Oh, John, have you gotten to talk to whoever at Das yet? [3:08 PM]
Eli Fisker: Lol, Hogla [3:08 PM]
codygeary: otherwise I can just post multiple of the same 2D. [3:08 PM]
machinelves: @jnicol speaking of linkbot needing vacation sometimes, the Help button embedded in the chat window is still on my wishlist, link to wiki help [3:08 PM]
Brourd: I would say that if a player waits until the last day to do anything, it will seem overwhelmingl, machinelves. [3:09 PM]
jnicol: LFP6, I did, we are bringing it up again this Friday [3:08 PM]
jnicol: ok elves :) [3:09 PM]
codygeary: But I'd rather have one giant sublab to allow players to pick their own KL sequences (many of them), and a couple smaller sublabs for the controls. [3:09 PM]
LFP6: Oh, do you have weekly meetings? [3:09 PM]
machinelves: it takes players hours upon hours to do all those labs, regardless of the length of time available, i believe you will get higher quality data if there is a reasonable limit [3:09 PM]
machinelves: brourd you yourself pointed out the poor quality of your submissions... [3:09 PM]
Brourd: No. Player sequences fail due to other reasons. [3:10 PM]
Eli Fisker: Cody, you can set the amount of designs you wish for seperately for each sub lab [3:10 PM]
codygeary: @eli. but they limit 3 sumbissions per player on each sublab [3:10 PM]
Eli Fisker: Ah, got it, Cody [3:10 PM]
machinelves: is it possible to make duplicate sublabs? [3:11 PM]
machinelves: then you could get multiples of 3 [3:11 PM]
Eli Fisker: Elf, ys [3:11 PM]
codygeary: I'd like each player to submit maybe 10-12 different KL sequences (its only 6bps, and I want as many different possible ones). [3:11 PM]
Brourd: The most important reason being there has been little data about multistate RNA design. It's building up though. [3:11 PM]
jnicol: Cody, that s a good idea to allow more submissions, duplicate the sublabs :) [3:11 PM]
codygeary: I'll make a separate sublab for a control group, but it doesn't need nearly as many sequences. [3:11 PM]
machinelves: i saw half a dozen different players wondering why so many sublabs [3:11 PM]
mr.mola-mola: Are you guys in, like a cool people group? [3:11 PM]
machinelves: i feel this affects their time per design [3:11 PM]
machinelves: mola this is dev chat, where we mention bugs and things, welcome :) [3:12 PM]
hoglahoo: Do people generally feel obligated to offer submissions for every sublab? [3:12 PM]
Brourd: I guess so, hoglahoo? [3:12 PM]
jnicol: I like to think of this as "ideas to help EteRNA become better" [3:13 PM]
hoglahoo: oh this is devchat, I just realized [3:13 PM]
codygeary: so jnicol, I can just post multiple copies of the same sublab, but this adds to the "too many sublabs" problem a bit too.. [3:13 PM]
Brourd: There can never be too many sublabs. [3:13 PM]
machinelves: jnicol love that tagline, adopted :D [3:14 PM]
hoglahoo: Brourd wants sublabs for the sublabs [3:14 PM]
jnicol: Cody, I wouldnt worry if you really want more submissions then add them [3:14 PM]
Eli Fisker: Agree with Elf :) [3:14 PM]
LFP6: So, what fun things have the devs been working on lately? Much coding? [3:14 PM]
LFP6: Or all lab fun? [3:14 PM]
codygeary: I have another question. Brouard, maybe you are the guy to ask. [3:14 PM]
LFP6: Or just getting ready or the holiday parties? :P [3:14 PM]
machinelves: maybe we can take a poll on what a 'reasonable' number of sublabs per lab is [3:14 PM]
Brourd: I doubt it, but plenty of people here who can answer. [3:15 PM]
Brourd: If not [3:15 PM]
jnicol: LFP6, mostly lab coding lately [3:15 PM]
machinelves: i just want to make sure you get the quality of data you wish for [3:15 PM]
codygeary: For the next lab I'm planning, I'd like to run a script to double-check that people pick complementary bases in the KL. I can't build that into the puzzle since it's pseudoknot. [3:15 PM]
Brourd: Actually, you can codygeary. [3:16 PM]
codygeary: really? I'd like to do just exactly that. [3:16 PM]
LFP6: What is lab coding, exactly? [3:16 PM]
Brourd: You create a second state, where there is a single helix connecting the two unpaired hairpin strands. [3:16 PM]
codygeary: but that's not a switch [3:17 PM]
codygeary: oh, wait, it's a trick, does that work? [3:17 PM]
jnicol: Brourd, for your competing lab, I chose only sequences that repeated the same 5 bases [3:17 PM]
Brourd: Indeed it does! [3:17 PM]
Nando: actually, the SCRIPT primitive has been coded as a constraint on puzzles, but I don't know whether it's functional yet (Jee would know I guess) [3:17 PM]
Nando: hi guys [3:18 PM]
jnicol: not function yet :( [3:18 PM]
Nando: ah, I see [3:18 PM]
jnicol: Hi Nando [3:18 PM]
hoglahoo: hi Nando [3:18 PM]
machinelves: hi nando! [3:18 PM]
Eli Fisker: hi Nando [3:18 PM]
Brourd: As a trick, codygeary, you can theoretically code any number of pseudoknots into a target, by making them alternative secondary structures. [3:19 PM]
LFP6: What is this lab coding though? [3:19 PM]
jnicol: Cody, I could do the same for your lab as I did for Brourds, only chose sequences that followed the complementary base rule [3:19 PM]
Brourd: well, those that follow the NN rules. The helices may be inverted and such. [3:19 PM]
Nando: won't stop people from submitting designs that don't comply, but it's maybe better than nothing at all [3:19 PM]
codygeary: exatly, that's what I want jnicol [3:19 PM]
Brourd: Yeah. And if you throw in that helix state. It gives players a foundation to design on. [3:20 PM]
jnicol: You can warn plyeres that the sequences will be filtered out if they dont follow your requiement [3:20 PM]
LFP6: Oh, post-analysis? [3:20 PM]
codygeary: yes. that would be ideal [3:20 PM]
Brourd: So they won't be required to guess if their kissing loop is correct. [3:20 PM]
Brourd: Just ask Nando. I come up with convenient hacks of the system all the time :P [3:21 PM]
jnicol: Ann now has a web page to collect sequences directly once a lab has closed, faster and more reliable [3:21 PM]
Eli Fisker: @John, awesome [3:21 PM]
Brourd: hey jnicol, how difficult would it be to post process a sequence and secondary structure, like switching a loop? heh heh heh [3:22 PM]
codygeary: I like you suggestion Brourd. I will totally try to implement that as a guide. [3:22 PM]
machinelves: anybody: what does "exp." mean in the view option: "use continuous colors for the exp. data" [3:22 PM]
Nando: experimental [3:23 PM]
Brourd: You should see the hack we have for mimicking the loop we mimic. [3:23 PM]
codygeary: experimental? [3:23 PM]
machinelves: ohhh thanks! :D [3:23 PM]
Eli Fisker: Thx, Nando [3:23 PM]
jnicol: I have not been into the flash code yet, so implementing a script constraint for the labs will not be available for the next round [3:23 PM]
hoglahoo: Brourd: show us [3:23 PM]
Eli Fisker: @John, could we have the image say experimental too: [3:23 PM]
Eli Fisker: http://prntscr.com/53kw4q [3:23 PM]
Eli Fisker: I mean in the gear settings [3:23 PM]
Brourd: I'm wary about posting any screenshots in chat, hoglahoo! [3:24 PM]
hoglahoo: Eli can do it [3:24 PM]
machinelves: thanks Eli that's the one [3:24 PM]
Brourd: That's just your way of making me try it again :P [3:24 PM]
jnicol: Eli, yes, when I start with the flash coding [3:24 PM]
Eli Fisker: John, thx [3:25 PM]
machinelves: @jnicol btw you've probably seen the misalignment bug in sublab table, but i noticed the sequences are not misaligned, if that helps know how to fix it: https://getsatisfaction.com/eternagame/topics/ui_issues_report_your_game_problems#reply_15061131 [3:25 PM]
jnicol: I've seen it and Jee said he would take care of that [3:26 PM]
Nando: @john: is there a chance you could describe to me how to build the Flash applet? I tried downloading various tools, got the Flex SDK 4.6, CrossBridge (as Alchemy replacement) and a number of other things, but never managed to build the whole thing... [3:26 PM]
machinelves: @jnicol that's great news, thank you! [3:26 PM]
jnicol: Nando, I have not built it yet either [3:26 PM]
Brourd: You can build it together! [3:27 PM]
Nando: aha, ok, that limits the options a little... :P [3:27 PM]
jnicol: Maybe get a session with Jee :) [3:27 PM]
codygeary: When I go to vote for sequences, the hilighted sequence doesn't align to the selected one. [3:27 PM]
LFP6: Any idea if/when we'll get those UI updates from NOVA? [3:27 PM]
Brourd: How many grad students did they get to help with the Eterna project this time around? [3:28 PM]
Nando: trying to catch gonna be a b.... [3:28 PM]
Nando: him* [3:28 PM]
machinelves: @cody that is the bug i mentioned in the forum link above, sounds like they will fix [3:28 PM]
machinelves: @jnicol another oldie but goodie: we had another new player ask about the next puzzle button, whether it could be customized like how we have sort options for general puzzle portal ( most solves, switch, etc... ) [3:28 PM]
jnicol: I think Mike was looking into the Nova code integration, but his time is limited [3:28 PM]
LFP6: Lol, next puzzle and questing. [3:29 PM]
LFP6: Please [3:29 PM]
jnicol: Elves, that is something I can do fairly quickly :) [3:29 PM]
machinelves: wow! yay!!! :D [3:29 PM]
machinelves: thanks :) [3:29 PM]
Brourd: Could lab projects be implemented into the next puzzle suggestion? [3:30 PM]
hoglahoo: yeah that would be nice [3:30 PM]
hoglahoo: re the next button [3:30 PM]
codygeary: so, again, my tall order.. @john, the scripting tester is buggy [3:30 PM]
jnicol: Cody, several bugs in the scripting interface have been reported [3:30 PM]
codygeary: :-) It's probably a less popular feature on eterna. But I like it lots. [3:31 PM]
machinelves: @jnicol & cody, especially while we have nando here, if there was a particular fix you were talking about in updating scripting, mentioning which has highest priority can help [3:31 PM]
jnicol: I'll ask for a list of issues when the bug tracker becomes available [3:31 PM]
machinelves: ooo bug tracker :D [3:31 PM]
machinelves: *waves pompons* [3:32 PM]
LFP6: When???!!! [3:32 PM]
Eli Fisker: :) [3:32 PM]
LFP6: :) [3:32 PM]
codygeary: I tried to add Nando's new parser module to my scripts, but to no avail. [3:32 PM]
Brourd: @ codygeary - in regards to the "trick" for your lab puzzle. You may want jnicol to implement an ungodly free energy bonus into it, when he puts that as a second state. [3:33 PM]
jnicol: noted [3:32 PM]
hoglahoo: 666kcal? [3:33 PM]
jnicol: ungodly large? [3:33 PM]
Nando: lol [3:33 PM]
Brourd: -666 kcal [3:33 PM]
codygeary: 'ungodly free energy' got it. That might fool the system. @hog. bwahahaha [3:33 PM]
hoglahoo: ungodly small? [3:33 PM]
mr.mola-mola: Is there a group that can help me learn how to program? [3:33 PM]
machinelves: @mola try codecademy [3:34 PM]
mr.mola-mola: thanks [3:34 PM]
machinelves: :) [3:34 PM]
Brourd: Also, I would suggest it be put into the puzzle so it ignores the usual tests for secondary structure in the check of the bonus. No clue how that will affect everything else, but it's a start. [3:35 PM]
Eli Fisker: By the way, Machinelves mentioned an idea for puzzles, that I liked a lot: [3:35 PM]
Eli Fisker: I tried to read the descriptions but I was going so fast through many that often I would click to open the puzzle before I remembered to scroll for the description. [3:35 PM]
codygeary: cool idea brourd [3:36 PM]
Eli Fisker: This is the view we get: [3:36 PM]
Eli Fisker: http://prntscr.com/5ai45d [3:36 PM]
Eli Fisker: This is the view I would like: [3:36 PM]
Eli Fisker: http://prntscr.com/5ai4dg [3:36 PM]
codygeary: I can't see puzzle previews in Chrome on my mac [3:36 PM]
machinelves: oh yeah i forgot thanks eli. yes i would like to see the description when opening puzzle, with no scrolling [3:36 PM]
Eli Fisker: I mean, no need to see the shape of the puzzle twice [3:36 PM]
Eli Fisker: on top of the page [3:36 PM]
machinelves: and keep the other info, just put under description [3:36 PM]
Brourd: so then, like this KL puzzle would have this as a second state http://prntscr.com/5aikco [3:37 PM]
machinelves: yes exactly [3:37 PM]
jnicol: I like also Eli [3:37 PM]
codygeary: cool, but what about all the other 2D? does it somehow ignore that for the 2nd state Brourd? [3:37 PM]
Eli Fisker: Thx, John [3:37 PM]
LFP6: Even better, I would like to see description available in the interface [3:37 PM]
LFP6: (Er, Flash app) [3:37 PM]
machinelves: Thanks Eli & John! that would be awesome [3:38 PM]
Brourd: That's what the ungodly free energy bonus that is not forming an internal loop does, codygeary. [3:38 PM]
machinelves: @LFP which part of UI? in puzzle itself? I totally agree [3:38 PM]
machinelves: often wish to see after clicking , and have to go backwards [3:39 PM]
Brourd: Or should do. I do reserve the right to say there are always weird side effects. [3:39 PM]
Nando: from a general point of view, the web could use a few collapse-able divs [3:39 PM]
LFP6: I mean, you have the link in the top-left. What if you did a rollover that brought up the @Elves description, or something? [3:39 PM]
LFP6: Yeah, in the puzzle [3:39 PM]
machinelves: yes! beautiful love it. roll over description tooltip on top left link [3:39 PM]
LFP6: Also allows better use of the 'next puzzle' option [3:39 PM]
machinelves: ohhh good point [3:39 PM]
Nando: for example in scripts, when I open one and just want to run it, I don't really need to see the code, yet I could be interested later in inspecting it [3:39 PM]
machinelves: yes that's a realy good idea, thanks! [3:40 PM]
LFP6: So you don't have to navigate out, which practically midigates points of the enxt puzzle [3:40 PM]
LFP6: *out of the puzzle window [3:40 PM]
mr.mola-mola: Whats dev mean? [3:40 PM]
jnicol: Quiet because I'm taking notes from everyone :) [3:40 PM]
machinelves: @nando yes it would be cool to have collapsible ( floating? :D ) windows to fit more info in according to what one needs [3:40 PM]
machinelves: thanks jnicol :) [3:41 PM]
hoglahoo: dev is short for developer [3:41 PM]
codygeary: So, can we consider a "colorblind" color mode that makes C and G not look exactly the same? [3:41 PM]
Nando: or devious puzzle author [3:41 PM]
hoglahoo: or devilish free energy bonus [3:41 PM]
Brourd: Jnicol scribbling in notebook *collapsible divs everywhere* [3:41 PM]
machinelves: @cody great idea [3:42 PM]
LFP6: @Brourd: No, *Jnicol scribbling in notebook* "collapsible divs everywhere" [3:42 PM]
LFP6: Or loose the quotes, not sure [3:42 PM]
LFP6: No, should be italicised! [3:42 PM]
LFP6: The part in quuotes [3:43 PM]
codygeary: horribly colorblind here. I use the "letters" mode, but when there are locks I cannot tell the G/C apart. [3:43 PM]
hoglahoo: is this screenwriting class suddenly [3:43 PM]
Brourd: I missed an "is" so it should be quotes. [3:43 PM]
LFP6: Yerp. :P [3:43 PM]
jnicol: Cody, like different spikes on the bases? [3:43 PM]
machinelves: @cody are there particular colors that work well and have good contrast? [3:43 PM]
Eli Fisker: Cody, here is a forum post on the topic: [3:43 PM]
Eli Fisker: https://getsatisfaction.com/eternagame/topics/adapt_properly_for_differening_color_perception_by_default_basic_usability_flaw [3:43 PM]
LFP6: @B: Oh, well, yeah then [3:43 PM]
codygeary: blue yellow is the best. but if the green was darker that would help. [3:44 PM]
Eli Fisker: It includes a color scale [3:44 PM]
machinelves: thanks Eli! [3:44 PM]
LFP6: @John: https://chrome.google.com/webstore/detail/spectrum/ofclemegkcmilinpcimpjkfhjfgmhieb?hl=en [3:44 PM]
Brourd: Hmmm, and an adjustable color scale for how SHAPE is displayed would be great as well. [3:44 PM]
LFP6: Eli linked that ni the forum post [3:44 PM]
Nando: @john: if you ever look into color blindness issues, https://chrome.google.com/webstore/detail/spectrum/ofclemegkcmilinpcimpjkfhjfgmhieb is a cool tool that allows you to emulate various types of color blindness [3:44 PM]
jnicol: Thx nando [3:45 PM]
Brourd: Nando lost this round. The winner is LFP6 [3:45 PM]
machinelves: Nando thanks that is really helpful to know of [3:45 PM]
LFP6: W00t!! [3:45 PM]
Brourd: Oh, looks like an upset. [3:45 PM]
Brourd: Nando wins, Nando wins. [3:45 PM]
LFP6: Hehe [3:45 PM]
Nando: lol [3:45 PM]
LFP6: Aaaand it all just goes downhill, but who cares any more? *Throws up handss, but stops with them up* [3:46 PM]
LFP6: *Sits back down* [3:46 PM]
machinelves: oh haha i am a slow reader, thx lfp too :D [3:46 PM]
LFP6: Nando gave the explanation [3:46 PM]
LFP6: It counts for something! [3:46 PM]
hoglahoo: a shorter url too [3:47 PM]
Brourd: Everybody is looking at the referee, hoglahoo, who's watch conveniently stopped when the first link was posted. [3:47 PM]
machinelves: i'm happy to have so many helpful hands, so more help is only better and appreciated :D [3:47 PM]
codygeary: I'm not sure this is dev related, but Eli told me that the signal/noise ratio on the KL lab was higher than usual, are there any explanations? [3:47 PM]
LFP6: oh, by two characters, bah! [3:47 PM]
hoglahoo: Brourd: if I didn't stop my watch, all my jokes would get old [3:47 PM]
LFP6: lol B [3:47 PM]
machinelves: ( hogla, this is why i don't wear a watch. problem solved: never grow old! :D ) [3:48 PM]
Eli Fisker: John, the high SN numbers on Cody's lab was something Meechl discovered [3:48 PM]
jnicol: rhiju wanted to discuss the S/N in the news item [3:48 PM]
LFP6: Well, this just disolved into bad puns. Back on topicnow [3:48 PM]
Brourd: codygeary - theory 1: The number of unpaired residues in the kissing loops project was low. [3:48 PM]
Eli Fisker: There is a forum post on it [3:48 PM]
hoglahoo: machinelves: I don't wear one either actually. gives me more excuses when I miss deadlines [3:48 PM]
Eli Fisker: https://getsatisfaction.com/eternagame/topics/awesome-signal-to-noise-ratio-super-stable-3d-rna-designs [3:48 PM]
machinelves: hehe [3:48 PM]
codygeary: And, does lower "shape threshold" relate to higher SN ratio? I didn't really understand exactly what the threshold meant, but I thought it meant there was a good contrast in the data. [3:49 PM]
Brourd: fewer residues were modified, therefore, all the data/signal was concentrated in a few bases. [3:50 PM]
codygeary: ah. I thikn that's a good explanantion actually [3:50 PM]
jnicol: ask Rhiju a question here: http://eterna.cmu.edu/web/news/5323230/ [3:50 PM]
jnicol: or post the link to the S/N forum duiscussion and ask him to comment [3:50 PM]
Brourd: Yeah. [3:52 PM]
Brourd: This is the longest you have stuck around in a while, jnicol! [3:52 PM]
LFP6: Oh, and as a thought, should all the links on official EteRNA pages be set to the eternagame.org domain instead of the other mirrors? Professionalism and consistancy thing in my mind [3:52 PM]
machinelves: thanks for all your hard work, John! [3:52 PM]
jnicol: are you saying its time for me to leave? [3:52 PM]
LFP6: Thought for the Wiki too. [3:52 PM]
LFP6: *That was a [3:53 PM]
Brourd: Rather, I am saying my devilish charm is doing its thing. [3:53 PM]
LFP6: er, is a [3:53 PM]
jnicol: LFP6, links should be relative [3:53 PM]
LFP6: Really? [3:53 PM]
jnicol: so if you start in a domain you should contine in that domain [3:53 PM]
codygeary: I'll post conclusions for the KL lab soon. But I wanted to thank you guys, the data made a lot of sense and was better than I had hoped for. [3:53 PM]
LFP6: I'm talking about static links, not buttons [3:53 PM]
Brourd: I don't believe I made any submissions, but that is great, codygeary. [3:54 PM]
jnicol: which static links? [3:54 PM]
machinelves: @cody thanks for participating and sharing your knowledge with us, we're happy to have you here! [3:54 PM]
LFP6: Not sure, hold on [3:54 PM]
Brourd: I will soon be posting results for the riboswitch targets. [3:54 PM]
LFP6: I think I'm just kidding myself [3:54 PM]
LFP6: Never mind [3:55 PM]
jnicol: Thanks everyone for the dev chat, your contributions are appreciated! [3:56 PM]
LFP6: Maybe I'm just thinking of news articles, or the old code of conduct page [3:56 PM]
LFP6: Thank you! [3:56 PM]
Eli Fisker: Bye John [3:56 PM]
Nando: thx John [3:56 PM]
machinelves: bye john, thank you! :) [3:56 PM]
