Still thinking on it.
First most proposals are about 2 people. I’m considering larger participation (2 included).
How to include AI(s) analysis specifically (as these are quite common intervenant, which makes a total of 3, not 2 many times ).
We could have help to (auto)classify different contribution around a node in the game at any time. An alert system when more contribution come in on a specific point (with a kind of bookmark for which point seems of higher interest).
Some help to eliminate redundancy (like eliminate it but still keeping some statistics like x players had same variation.).
Having a like or dislike or doubt system (quick way for feeling feedbacks) May add wrong, overplay, winning, similar, hard to see, [empty] and more? May be easy choice with popping list, autocompletion or similar.
May have help for tewari analysis.
Could be through some labels you put directly on the game tree. Colors (or code or full name) for label/text (developed in game chat window) ownership.
Each participant may hide/delete whatever and propose it to others too. Or highlight.
Resume of a full branch on request.
Remind me on the democratic collective game we had here… so we could add some poll system (with a bit more automation) so we could use the tool for team play.
Try playing a whole teaching game with variations. I suppose it’s possible, but wouldn’t that get tedious? So much more clicking just to see what the other person did.
The more-than-two-reviewers option did also cross my mind.
A consequence can be that it becomes a discussion between reviewers.
This can be a good thing, but there is also a possibility that the person(s) who asked for the review (the player or players that played the game) get lost in a discussion that is way over their head.
That is a risk that is inherent to this set up. No idea how to prevent this.
It should be possible to create Shared Demo Board the same way as Direct Challenges work - by clicking name of someone… and if they accept, it starts.
That’s basically a minimum of abilities that should be available in Shared Demo Board: https://forums.online-go.com/t/features-everyone-secretly-wants-on-ogs-but-will-never-be-implemented/32125/595
But also there should be button that will allow opponent to place stones of any color.
Logic is: anything that is possible in real life Go club when 2 people sit near Go board should also be possible on online go server.
Interesting … but then the creator would only get it back if the person who got control from them actively gives it back? That would be sub-optimal, to say the least. Thus it seems a little more complicated to me — but I am not a coder.
There shouldn’t be a concept of “control”. Just let a few of us collaboratively put stones on a board. It shouldn’t even require accounts for anyone but the creator.
Something like this would make it much easier to conduct a lesson over Zoom, for example, especially if the teacher is not an OGS expert.
I still think that some memorization of what happen on the board should happen. For example if a teacher propose a L&D problem, to be able to reset to the starting position. Or to compare then different sequences proposed by “students”
Something great imho: by clicking on someone name in the list of participants, you can get or hide his moves on the board (for you )
it sounds simple, but also real life go boards don’t have delay that can potentially go upward mutliple real life seconds as somebody works in their reality, and another in theirs, while those two realities contain conflicting events
Sort of? but it’s a much simpler handle, since as the review owner (required for the Take Control) button, you have priority over control passing, and I’m pretty sure that overrides a board editing event from the other person’s perspective and they end up on a local branch not shared to main.
Not necessarily, since that “Take Control” action is now an event. If A has control at T=0, and B inputs an event at T=1, it will send the “takeControl” and “inputEvent” flags at the same time, and if A inputs their own event at T=2, then at T=3 the server recieves A’s event, and at T=4 it recieves B’s event, what happens? Does the server prioritize B’s taking control cuz it was done first? or A’s event because it was recieved first?
Currently, I’d suspect that the takeControl event has priority due to review ownership, but now we don’t have necessarily have that heirarchy, and the control flag is sort of redundant
and then there’s the option of having both events happen, but in a specific order, and only not happening if that order doesn’t constitute a legal board state. What happens if one player goes back a move, and the other places at the same time?
Or to take it further from what’s possible on a real board, one of them switches branches?
On the current structure we can just have one of them desync and manually resync, based on control, but if we do that now, we can have concurrent edits to different parts of the tree, potentially deleting entire sections as one adds more to a section
Cuz you’re not moving stones on a board, you’re editing an SGF. There are nice simplifications to limiting that to real board actions, but you lose out on benefits that are actually kinda nice about it being an SGF such as perfect move-order memory and branches
Shared Demo should look more like how games look now than how review looks now.
There is no “give control”. Both place stones freely. If both place stone too simultaneously, let random generator to decide the order in sgf.
There is no time travel with branched timelines in real board. So by default players see newest version of the board only.
There may be a problem if player A wants to place a black stone at intersection a and player B wants to place a black stone at intersection b simultaneously. If player A is a few milliseconds faster, then a black stone is placed at a and a white stone at b - which is not what player B intended to do.
A way to circumvent the problem would be: if a player places a stone on the board, then another player cannot place a stone before a certain delay, like 1 second (like in a real-life review, you have to wait for your partner to remove his hand from the board).
I also think that simultaneous editing is desirable. It’s acceptable to have only one controller initially, but it would be fine to add more later.
Regarding simultaneous operations, I believe it’s sufficient to process them on a single thread in the order they are received by the server. If an error occurs when attempting an action, refreshing the screen would be acceptable.
This is because, in reality, when we want to perform simultaneous operations during review, we are likely having conversations, and mutual exclusion is managed through those conversations.
I think that alternating moves would be more effective for review than trying to cover simultaneous operations.
I wonder is a simpler solution just to make it much easier to pass control to someone during a review.
At the moment you have to find the players name, either in the participants list or if they’ve sent a message in the chat, then click to bring up the player card and then pick to pass control.
So maybe just some quicker interface near the game tree or somewhere that’s generally in view, press a pass control button and then pick a player? Would that be better than the current system.
I feel like anything with race conditions for the timing of who click what can’t be good. It can be headache in scoring for instance if two people are disagreeing and clicking or submitting at the same time or fighting with the autoscore etc
Maybe my imagination is not good since I don’t ever do reviews, but I can’t imagine a scenario in a review that would be a headache and not be trivially resolved by backing up a step or by the reviewer disabling the viewer’s access.
The problem in scoring seems to be that there’s no way to resolve an actual disagreement, but that wouldn’t be the case with a review since there’s a defined owner who is in charge.
A “Give Control” button which brings up a list of players available, and when one doesnt have control, changes to “Take Control” for thé review owner, and greys out for others, perhaps ?
It would be a lot more visible, findable and self-explanatory than the current interface, and easier to click for passing control back and forth.