I'm getting an instance crash when opening the puzzle as well.
Thanks for the report. We're looking into it.
It seems to be inconsistent, so retrying a few times may help. Some other things that have been reported as a workaround are to try loading the puzzle from within another puzzle (versus from the main screen), or tweaking your view options in another puzzle prior to loading 1855. Also, some people have had success with deleting the corresponding puzzle files (those starting with "0002009845") from the Foldit running directory to force a re-download of the puzzle from the server. -- All of that might be voodoo, though.
We'll try to keep you updated as we figure out more what the issues might be.
thx 4 the info. starting after opening another puzzle worked for me at least one time (1st try).
We found a memory corruption bug which we believe to be the cause of the issue. We should hopefully be rolling out a patch for that sometime soon. (After we dot 'i's and cross 't's and push it through the release pipeline.)
I accidentally downvoted Xartos' post, and can't fix it.
I'm also getting an instant crash on 1855 using 20200618-bbc43b61c3-win_x86-devprev.
In 1986 I published (my only refereed paper) the mathematical basis for James G. Nourse's Configuration Symmetry Group.
https://pubs.acs.org/doi/abs/10.1021/ci00050a009
That was part of the sugars database for automated reaction design in DENDRAL, back in the very early days of chemical AI.
Sorry to be missing this puzzle! Am in discussions to get an actual graphics card for my antique client.
Turns out Microsoft quietly threw my onboard chipset under the bus about 5 years ago. When Foldit's server side's graphics went to OpenGL 3.x it bricked my client as presently configured. Here's the details.
https://forums.intel.com/s/question/0D50P0000490WWlSAM/update-opengl?language=en_US
"That is correct the Intel(R) Q45/Q43 Express Chipset supports OpenGL 2.0 and you will need a system that supports up to 3.0, as Al Hill mentioned above if you have a PCie slot you can buy a graphics card for you system, other than that you will need to buy a new computer to work with this type of software."
me again: That might be the way ahead for Bletchley's legacy client too. Sorry for off-topic ramblings, hope to be back soon :-)
Thanks John!
This was one of the theories on our side, so I think it is a possible culprit.
However, we are trying to figure out what we changed that has suddenly caused your computer to be unsupported, as we don't think we've done anything that should cause this. If possible, we would like to continue supporting older hardware, as other people who try Foldit may run into the same issue.
I have two questions regarding this puzzle.
1. Some of the fragments have green endpoints sticking off of them. Is there anything special about these?
2. Often, when I try to change reactant fragments, it will change, but then immediately reverts back to the previous design. Why is this happening?
The two green atoms sticking off of some of the reactants are chlorine atoms. A number of drugs and drug-like molecules have chlorine (or fluorine) atoms on them, as they have structural and interaction properties that are different from the other atoms typically in molecules.
I'm not sure what might be happening with the reverting. I will say that selecting the reactants/reagents in the panel only sets them for the proposed (ghost) structure. Until you click the accept button (the two test tubes into a beaker icon), your designed ligand won't be realized in the protein. Exiting out of the panel by the red X close button will discard the proposed structure and return you to the ligand you had when opened the panel.
Thanks. I am hitting the button at the bottom of the reaction design panel to run the reaction.
I selected my reaction design in the menu and pressed the button at the bottom that looks like the remix button.
Once pressing it, it applies the reaction, then snaps back to the previous design before I can do anything. This can be shown in the undo menu, where there is a yellow V shape. The base is the applied reaction, and the rise back up is from it automatically undoing the change.
I cannot get the image uploading to work to provide further info unfortunately.
Just downloaded this puzzle. Can't see the reaction tool anywhere!
Found it! Well hidden inside the structure!!
The reaction tool is a train wreck.
I have no idea how to choose/unchoose what I want, and no one seems to be able to tell me.
What is the significance of the row on top? And why three columns? None of this was answered in the tutorial.
There is nothing to tell me which structure(s) I have chosen. Why aren't the current choice(s) highlighted? There is also no system for keeping track of permutations chosen. Do I really need to get a pencil and paper to play this game?
I appreciate seeing and trying new things, but this level of frustration I feel from being left cluelessly in the dark is exactly why new players mass exit this game as fast as they come in.
The lack of detail (text) to tell us ANY sort of info about the choices, makes this all the more difficult to use the tool (or at least use it properly/to the fullest extent).
Which I, too, found the lack of selection indicator to be frustrating when re-opening the tool after a couple hours. :(
I could repeat Boots' other points, but I'll simply end on the note of:
This is coming from someone who not only played, but "beat" the Tutorial level for Reaction Design! Sadly, that tutorial taught us nothing other than how to access the tool in Original Interface. Mind you, I do understand that it's always able to be improved and made better, but it almost seems better to have waited before rolling it out. The text in the Puzzle Description was sufficient to get the ball rolling, all that was needed was a quick bit of text that explained how to access the tool and it would've been on par with the tutorial lol
(also... while we're nit-picking... renaming the Intro/Tutorials to "Campaign" wasn't a good idea, since in every game I've ever played the Campaign(s) are the full-blown single-player side of the game; unrelated to the tutorial or introduction of game mechanics.)
<3
-Form
Since we're not allowed to delete or edit our posts....
I found what the top row and the columns represent, here: https://fold.it/portal/node/2009869 . For anyone who wants to know, that page explains the new tool... somewhat.
All the rest of my questions remain.
And now I have more questions. Can I choose more than one item in the same column? Why is the ligand the same color as the protein, making it difficult to find?
(Note: The Boots' last question about colors is because he is colorblind; there are a handful of active folders that are)
In DevPrev, they appear to have already addressed the Menu not showing selected options :) I don't know if that will be visible to you, but hopefully it is (they are teal, which is inbetween green and blue). The menus are also (at least in Dark Background mode) also now equipped with a dark background. In actuality, the fact that Mainline has a bright background MAY be hiding the selection boxes (even to a Trichromate like myself lol)
No green or blue or anything inbetween om my legund. it is orange or orange-brown depending on how one sees color varients, jusr like hydrophobics, i.e. no color distinctiveness between hydrophobes and legund. And yes I do have color vision but near blind in other regards.
I might have misrepresented what I was trying to explain... (it was 2:30am and I was very tired)
What I meant was that in the Reaction Design tool window where you pick the options to build your Ligand, that the thumbnails in DevPrev (the Test version of Foldit) have a Teal (Green-Blue) outline around their boxes to show that it is selected.
In the Main version it appears to lack any outline and that makes it very hard to determine what configuration you are using.
Maybe I am missing something here, but this almost looks like one of the cases where an automated algorithm would do a better job? After all, it appears that there are only 6 options in the first column, 15 in the second, and 2 in the third, for a total of 180 combinations.
Is the computational complexity so great that a brute-force approach won't outperform humans, or is this a situation where pattern recognition actually matters (because fine-tuning the fit via manual adjustments is also needed)?
Is that we've been given this as a real-world Training Exercise.
The Tutorial level is to give you a quick and dirty "What this is, and how to use it" intro to Reaction Design. Granted, it's.... not very good at doing that right now, but that's besides the point lol
The point being that once you've been shown how the Tool functions, you need something to practice on and understand what and how you play the Ligand type of puzzles.
Again, there is a lot that we're not being given, information-wise, that would help us... but it is early and these sort of things continually receive polishing.
Beyond that, I would almost argue that having so few choices actually makes it ideal for us human meat-bags to be the ones to play it. If we had 30-choices in each column... can you IMAGINE how daunting that would come across? We'd need a month in order to figure out the best combo!
I would say then, that the Algo's tackle the giant list of candidates and refine it into a smaller one which "seem to work well". Then we get our fingers on them and work our magic to come up with creative ways to shoehorn that square peg into a round hole. :)
That's just my 2-cents though.
Having LUA commands like at:
https://fold.it/portal/node/2009887
to pick the reactants and make the ligand would help.
Since sharing solutions with Group doesn't seem to work for 1855
(Crashes 6-7 at https://fold.it/portal/node/2009873#comment-42080),
I'm guessing that sharing with Scientists won't work either.
With all the different ligand designs possible in Puzzle 1855,
this is a shame. I'd imagine there will be a great diversity
of solutions from each player. Because of this, I hope that
the Foldit Team will mine the many solutions automatically
shared with the server as we play Puzzle 1855 online.
Perhaps posting a sequel to Puzzle 1855 after fixing all the
bugs with 1855 would help. It would also be nice if you would
let us load solutions from Puzzle 1855 into this sequel puzzle.
Thanks!
JeffThe wiki now has reaction design puzzle with a general description, and puzzle 1855 reaction design with more of a deep dive.
The big discovery for me was that the initial preview image does not match the initial selections when you open the reaction design panel.
When you open the panel, the top of each of the three columns is selected, so "1,1,1". But the preview shows you "4,4,2". If you actually click on those options, you'll notice the preview doesn't change.
I didn't figure all this out until about a week after the puzzle closed.
Cant open the new puzzle the client always crashes when i try to open it. Even after redownloading the puzzle still the same.