On 11/16/2010 12:35 AM, Joost 't Hart wrote: Consolidation. See below.
One new question: How do we prefer to manually interfere with the machinery, should we need this. There are two machines that can/need to be stopped. The engine and the autoplay function. Actually, there is a third notion: Does autoplay drop annotations while strolling through the game? My first temptation was to hit the stop button in the engine window. This stops the engine, but not the autoplay and annotation, and when doing batch annotation, the engine is restarted when entering the next game. Maybe desirable, maybe not... Autoplay can only (?) be stopped using the small "rectangular thing" above the e8 square in the main window, (it took me ages to spot it, the keyboard shortcuts (ESC, CTRL/Z) that the help text talks about do not work). But this keeps the engine burning fuel. My impression is that after hitting both these two, we might be back to normal. Not sure though. Opening the annotation config window once more while annotation is running (especially hitting cancel) has some awkward side effects as well: It keeps both autoplay and the engine running, but the annotations are no longer produced. Hitting OK indeed restarts analysis :-) It is fairly easy to make a single-shot stop action. * Is this desirable? * Are there any pitfalls in different scid contexts that should avoid this? In other words: The implementation is a bit messy (I cannot even guess what it tries to achieve and the tests above certainly do not draw the complete picture) and I need a few guidelines. Cheers, Joost. > Hi scid user, > > This is about the behaviour that we would like to see with annotation: > > I am happily looking at the following: > > "For<blablabla> moves (only)" > Only moves of the selected color (or both) are annotated, except the > final move of game/variation, which is always annotated. > This is not different from the behaviour we had, except for the final > move thing. I think it is better to always annotate the final move. > Furthermore, as we speak final moves in variations are not annotated at all. > > "Annotate<which> moves" > "Annotate all moves" will - as it says - annotate all moves, so it even > may come up with an alternate engine move that has exactly the same > score as the move in the game. > Yet I have enhanced the code such that it will never insert a variation > starting with the same move as in the game, as this makes no sense. As > soon as the game deviates from the/this engine line, there will be > another chance to insert it (with probably more accurate statistics). > For the very same reason, the starting move of a variation is not > annotated (unless it is also the final move of that variation). How > could the variation that the engine comes up with be different from the > variation given for the main move in the game? Basically, not. > > Still on my wish list is the implementation of another filter: Add the > notion that it might make no sense to add a variation starting with a > move that is already a variation for the current game move. > But this needs some clever thinking that I have not been able to come up > with yet. Especially the fact that the user actually may want to see > this variation if he has selected not to annotate his own variations. > I will keep it in the back of my head. This will stay on the list as far as I'm concerned. It is getting time to get this over with :-) > > "Annotate only blunder moves" will only show a variation to moves that > make the score go down more than the value entered in the dialogue. Plus > the final move in the game/variation. > > "Annotate only not-best move" will only show a variation to moves that > make the score go down. Furthermore this /mode/ includes the notion of > the blunder checking of the previous option. The difference shows with > the annotation options (see below). > If you like this idea, we probably should redesign the configuration > dialogue a bit such that it gets clear that the blunder level applies to > both modes. I dropped this mode, as people do not like the verbose blunder remark anyway, so we can do with a more modest single blunder filter that one can put to zero or 3 or whatever value. > > Annotate variations is not changed. If not enabled, the automaton simply > skips variation that are already in the game. If enabled, also the > variations are handled, in the same fashion as the main moves, with the > exceptions mentioned above. > > If you do not enable short annotations, the engine will present > variations with > - it's name > - a score tag (with UCI engines, I replaced the bogus 327.xy value with > Mx (or M-x) as is done in the engine window. Good idea?) > - the known remark "engine reports blunder" for applicable moves (score > down> x) when we are in one of the two annotate-only /mode/s. This remark (from the old blunder mode) will no longer be shown. > > If you do enable short annotations, the engine will only present NAG > stuff, unless you also enable add score to annotations: In that case the > score tag is also added (but not the engine name and the more verbose crap). Added a new option "Score all moves" (I am sure that someone else will take care for the translatable string part) that makes sure that all move receive a <depth>:<score> tags (also moves that do not have a variation/nag attached). > > I have a final question: What is the idea of enabling both batch > annotation and book annotation? > > Currently Scid adds another annotator string of the form "opBlunder > <move#> (<colour>)" with every move in the game from the moment we have > left book. This only shows since my recent cvs update regarding the > extra field. In the old days this probably did not work at all. > This is nonsense of course. Except for the fact that opBlunder is > apparently an untranslated stringID and that the colour mentioned is > wrong (should be the opposite). > > What /is/ the wanted behaviour? > Somehow it occurs to me that the engine should not annotate at all in > this mode. And that we only want a quick opening mistake check. But > maybe I am wrong. > Making it such that only the first non-book offender is added to the > annotation is straightforward. But is it the idea to indeed forget about > further engine involvement? Is straightforward as well. Let me know. Decided to remove the annotator string completely and let the engine run up to the move number specified. > > Hope for some interaction. Shoot! It will allow me to quickly finish > this job. > > Cheers, > Joost. > ------------------------------------------------------------------------------ Beautiful is writing same markup. Internet Explorer 9 supports standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. Spend less time writing and rewriting code and more time creating great experiences on the web. Be a part of the beta today http://p.sf.net/sfu/msIE9-sfdev2dev _______________________________________________ Scid-users mailing list Scid-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/scid-users
