New analysis collapse/expand variation functionality • page 1/1 • Lichess Feedback • lichess.org

The new functionality is a bit vague. I don't understand what it should do. It's also very inconsistent.

Example: 1. e4 e5 (1... d5) (1... c5) *
If you collapse any of the e4, e5, d5 or c5 moves, e5 remains visible and the other two are collapsed. But if you click on e5 all three become visible.
Collapsing on e5 will show the parent node (e4) as collapsed=true, then e5 is collapsed=true, but d5 and c5 have collapsed=false. If you expand it by pressing the + sign, e4 appears collapsed=false, but e5 collapsed=true. If you collapse on d5, now e4, e5 and d5 appeared as collapsed=true.

Then there is the automatic collapse, which I don't know what means, hiding variations regardless of what I actually want. Even when they are graphically collapsed, there are no nodes that have the collapsed property true unless you manually collapse them.

If you change the collapsed nodes structure, then you change the study chapter and come back, the collapse settings are reset.

In summary:

study move list automatically collapses some nodes regardless of what I want and no way to turn this behavior off
the collapsed nodes structure is not preserved on chapter change or page refresh, which make the functionality useless
not clear what is actually collapsed and expanded as the button has no label and it's not clear what's under the + sign
readability is decreased, not increased
the underlying data structure is buggy

The new functionality is a bit vague. I don't understand what it should do. It's also very inconsistent. Example: 1. e4 e5 (1... d5) (1... c5) * If you collapse any of the e4, e5, d5 or c5 moves, e5 remains visible and the other two are collapsed. But if you click on e5 all three become visible. Collapsing on e5 will show the parent node (e4) as collapsed=true, then e5 is collapsed=true, but d5 and c5 have collapsed=false. If you expand it by pressing the + sign, e4 appears collapsed=false, but e5 collapsed=true. If you collapse on d5, now e4, e5 and d5 appeared as collapsed=true. Then there is the automatic collapse, which I don't know what means, hiding variations regardless of what I actually want. Even when they are graphically collapsed, there are no nodes that have the collapsed property true unless you manually collapse them. If you change the collapsed nodes structure, then you change the study chapter and come back, the collapse settings are reset. In summary: - study move list automatically collapses some nodes regardless of what I want and no way to turn this behavior off - the collapsed nodes structure is not preserved on chapter change or page refresh, which make the functionality useless - not clear what is actually collapsed and expanded as the button has no label and it's not clear what's under the + sign - readability is decreased, not increased - the underlying data structure is buggy

dboing

edited

#2

I am happy to discover that lichess is finally upgrading the level of citizenship of mortal chess players to access some of the SF engine priviledges (or feature now) that allowed their longer that one width worth of new branch suggestion to be truncated.

I do think that while in construction mode, such as in-game correspondance, or standalone analysis (not study, I find it clunky there to work on bushes, in constructoin mode, and even in revision mode, but in lecture mode, they come with their sibling ordering so I listen and have no need to revise sibling ordering, ... although I could ramble some about that too, for efficient use of bi-direction communcition with studies). sorry for the nesting.. but like RAV format, it might be how to compress a bigger concept into a stream.
what a CPU prison, the verbal language is.

any way.. if there wsa ALSO (because I like the invisible for construction clarity mode), a way to have the truncated but visible variant of collapsed in some way, that would be an improvement. Should I make many standalone threads.. I thought I would surf on this one. objections?

Finally to either Lichess or my favority Lichess Tools developper, i.e. its creator. Why not have alternating background transparnet shade for the SAN syllable fonts as well. Like in typical spreadsheet UI that is minimally aware of perception mechanisms... so our volatile eye gaze pointer could find its kittens at a glance without loss of train of thought information.

ok. done with my prose. too compact. too rambly. Want some expansion? ( hehehe).

Edit: this contrary to op post, is from user land point of view.. so maybe it should be another thread?

I am happy to discover that lichess is finally upgrading the level of citizenship of mortal chess players to access some of the SF engine priviledges (or feature now) that allowed their longer that one width worth of new branch suggestion to be truncated. I do think that while in construction mode, such as in-game correspondance, or standalone analysis (not study, I find it clunky there to work on bushes, in constructoin mode, and even in revision mode, but in lecture mode, they come with their sibling ordering so I listen and have no need to revise sibling ordering, ... although I could ramble some about that too, for efficient use of bi-direction communcition with studies). sorry for the nesting.. but like RAV format, it might be how to compress a bigger concept into a stream. what a CPU prison, the verbal language is. any way.. if there wsa ALSO (because I like the invisible for construction clarity mode), a way to have the truncated but visible variant of collapsed in some way, that would be an improvement. Should I make many standalone threads.. I thought I would surf on this one. objections? Finally to either Lichess or my favority Lichess Tools developper, i.e. its creator. Why not have alternating background transparnet shade for the SAN syllable fonts as well. Like in typical spreadsheet UI that is minimally aware of perception mechanisms... so our volatile eye gaze pointer could find its kittens at a glance without loss of train of thought information. ok. done with my prose. too compact. too rambly. Want some expansion? ( hehehe). Edit: this contrary to op post, is from user land point of view.. so maybe it should be another thread?

dboing

#3

Correction on my feedback. I thought it was a per branch mechanism. Now working with siblings, n=3+ at branching point and it collapses all branches. I would then wonder as op, what is the intention here. I am not really complaining unless it is not meant to be discussed and improve, harmoniously through dialog and feedback. Like well initiated here (self congratulations are in order for me and op for now, I think).

Trying to guess the intent: might it be about having the mainline more salient? Could it be that this is the first step from engine features to human context, and that the engine post-game collapsible was never about high branching cases? Always the played games and the error from ceiling noodle that SF found, being the only new branch?

continuing my in-game experience....

Correction on my feedback. I thought it was a per branch mechanism. Now working with siblings, n=3+ at branching point and it collapses all branches. I would then wonder as op, what is the intention here. I am not really complaining unless it is not meant to be discussed and improve, harmoniously through dialog and feedback. Like well initiated here (self congratulations are in order for me and op for now, I think). Trying to guess the intent: might it be about having the mainline more salient? Could it be that this is the first step from engine features to human context, and that the engine post-game collapsible was never about high branching cases? Always the played games and the error from ceiling noodle that SF found, being the only new branch? continuing my in-game experience....

dboing

#4

ok. now with one left most node forced in variation as it is wip subtree. I get 2 branching siblings as opponent reply leftmost nodes candidates (also forced variations, sue me i don't like parentheses anywhere, I am just forced to use them here, but in chessboard-Lichess bubble I can make it fully human vision friendly). And I see that it might be better called "Collapse sub-tree from here". I would make the "s" in variation standout better, as in not needing to be seen. Less subtle signal is better in general. I And no harm in doing so.

Yes. Mainline becomes a stupid notion when exploring continuations from the only last known actual move done. (kind of obscure concept, perhaps in chess culture, explaining, why it persists in chess tool development? Well, mostly in study analysis ancestral feature set, hypothesis: PGN implementation is hard-wired into the human variations business there, and it is not in other analysis features that I mentioned). Not off-topic. Not for me. I like to step back and suggest it might be more efficient to discussion eventually. But I might be really weird thinking like that, still, I am my own user case data point. Do what you will with it.

Traditions are not all self-validating by just existing, for ages and ages. Sometimes we have additional information or have ways to look back at previous, invisible information that was always there.

ok. now with one left most node forced in variation as it is wip subtree. I get 2 branching siblings as opponent reply leftmost nodes candidates (also forced variations, sue me i don't like parentheses anywhere, I am just forced to use them here, but in chessboard-Lichess bubble I can make it fully human vision friendly). And I see that it might be better called "Collapse sub-tree from here". I would make the "s" in variation standout better, as in not needing to be seen. Less subtle signal is better in general. I And no harm in doing so. Yes. Mainline becomes a stupid notion when exploring continuations from the only last known actual move done. (kind of obscure concept, perhaps in chess culture, explaining, why it persists in chess tool development? Well, mostly in study analysis ancestral feature set, hypothesis: PGN implementation is hard-wired into the human variations business there, and it is not in other analysis features that I mentioned). Not off-topic. Not for me. I like to step back and suggest it might be more efficient to discussion eventually. But I might be really weird thinking like that, still, I am my own user case data point. Do what you will with it. Traditions are not all self-validating by just existing, for ages and ages. Sometimes we have additional information or have ways to look back at previous, invisible information that was always there.

dboing

edited

#5

2 level collapse, reporting. It seems that postcollapsing a subtree from the lefmost node (any sibling i assume), the left click decollapsing happens from touching the root node where the branches emerge (the branching node?) that which is relative mainline to the collapsed siblings.

I could perceived that only from above topology. (using forced variation for its indentation side-effect, so did not need a bigger subtree or bush). I can see how this might help in revising a bushy thing. Although I am not sure that it could not be also displaying the lefmost node of the collapsed (as a fruther keyboard or mouse even upon collapsing or decollasping features.).

or yes. maybe put those under the forced variations, as I would us that more often and before any collapsing.. In my mostly construction heavy user case. really should make my own thread.. op please let me know.

Exciting new avenur though, Lichess. I approve, if that helps anyone at all.

2 level collapse, reporting. It seems that postcollapsing a subtree from the lefmost node (any sibling i assume), the left click decollapsing happens from touching the root node where the branches emerge (the branching node?) that which is relative mainline to the collapsed siblings. I could perceived that only from above topology. (using forced variation for its indentation side-effect, so did not need a bigger subtree or bush). I can see how this might help in revising a bushy thing. Although I am not sure that it could not be also displaying the lefmost node of the collapsed (as a fruther keyboard or mouse even upon collapsing or decollasping features.). or yes. maybe put those under the forced variations, as I would us that more often and before any collapsing.. In my mostly construction heavy user case. really should make my own thread.. op please let me know. Exciting new avenur though, Lichess. I approve, if that helps anyone at all.

dboing

#6

So op is concerned mainly about the ancestral study analysis feature, and I am mainly concerned about the more recent and less historically buggy** following descendant analysis variants: 1) in-game, 2) post-game, and 3) stand alone as in Tools menu; perhaps even post-puzzle, which might be a potential fourth. Some of those might actually be the same, but since primarily use 1), I would not dare assuming that.

** (to me, on the variation context menu actual behavior)

In conclusion. It might be adding to the existing buggyness in study for the variation manipulation behavior, and I would have no clue. Should I make a new thread. or is this difference informative to keep under same hood.

So op is concerned mainly about the ancestral study analysis feature, and I am mainly concerned about the more recent and less historically buggy** following descendant analysis variants: 1) in-game, 2) post-game, and 3) stand alone as in Tools menu; perhaps even post-puzzle, which might be a potential fourth. Some of those might actually be the same, but since primarily use 1), I would not dare assuming that. ** (to me, on the variation context menu actual behavior) In conclusion. It might be adding to the existing buggyness in study for the variation manipulation behavior, and I would have no clue. Should I make a new thread. or is this difference informative to keep under same hood.