> Pascal Georges wrote:
>
> Hi!
>
> This means another option in an overcrowded dialog box
>> (that even does not fit in low resolution screens).
>>
>
> Life is a complex thing. Nobody ever stated it is simple.
>
> Is it a really useful thing ?
>>
>
> Up to discussion, thats why I worte this mail anyway ;)
>
> I'd find it useful, but I'm always with Mark Twain: if
> you're ever on the side of the majority you should stop and
> think about it. ;)
>
I prefer Saint Exup=E9ry (hoping my translation is ok) : "Perfection is not
achieved when there is nothing more to add, but when there is nothing more
to remove."
>
> My point is, that you may have added your own annotations to
> a game you actually played. What I have in mind is a
> correspondence game that is played "without engine support".
> (Even if I spent some time to the evaluation I might be
> wrong. Acutally, being a slightly weaker player than
> Kasparovs this happens quite often that I'm wrong.) Here,
> things might get _really complex_ if the engine throws in
> it's main lines at every other point, plus probably
> crosschecking with another engine... However, the engine
> eval values in the graphical display might be very usefull
> to identify a blunder one made. To graph it, well one needs
> the values first. One could then probably check in more
> easily on the real points. And to get an estimate value I
> can run the engines at a very short time limit and then
> explore deeper at the crucial points.
>
Please have a look at the code of the Score graph window : it is tricky and
it deterred me for example to add a second graph for output of an extra
engine.
>
>
> But feel free to add it if you want (maybe using notebooks
>> widget in Tk 8.5).
>>
>
> I do not think a notebook is suitable there. It would make
> the interface difficult. (Don't get me wrong, I like
> notebooks, I come from OS/2, where the WPS uses them
> extensively.) Besides AFAIK you added this dialogue, I'll
> hardly code within it anything without discussing it first
> ;)
>
I used notebooks on Pocket PC and this makes the UI look better organized.
>
> I think one could use drop down lists instead of the pretty
> space intensive radio buttons, like for the one used for
> the opeing book selection. It could look like this:
>
> Add annoations:
> For moves by both sides [V]
> For white moves only
> For black moves only
>
> Annotate all moves [V]
> When game move is not best move
> When game move is a blunder
> Only score
>
> What do you think? That would save quite some space without
> making it difficult to use and woudl also shorten the
> dialogue for CE.
That looks ok for me. By the way I find some dialog boxes in Scid a bit ugl=
y
... Meanwhile it could be interesting to make things look a bit better (thi=
s
also may require the use of Tk 8.5).
>
>
> - To have a "full" score graph, adding the engines evaluation
>> even if the best line is played (by the engines thinking)
>> would be helpful. One probably might want to add
>> engine-lines only in case of "not best line played".
>>
>> I thought those options already exist !?
>>
>
> You can set "annotate all moves", right, but you'll then get
> the engines line as well, not just the value.
>
Yes, ok. But score with line should go together, I think. Not be separated.
>
>
> That's right, but in some cases the user may want to annotate the game u=
p
>> to the last move, and let it analyze the last position. So this would me=
an
>> yet another option ...
>>
>
> I admit, that I thought about a checkbox for "stop after
> last move" as you say both behaviours may be wished for.
This option is useful indeed, simply the layout should be carefully thought
about, to avoid an overcrowded dialog box.
>
>
> - Annotate several games: besides specifying a number, it
>> might be a very good thing to annotate the games in the
>> filter. (E.g. if you have ongoing games in a base and you
>> want to auto annotate those finished only.)
>>
>> Another option ...
>>
>
> Probably using a radio?
>
> Annotate seveal games
> [*] Filter [ ] from current to [spinbox]
>
Yes, of course, but the dialog box is 2 lines bigger ... Notebooks (one tab
with "extra options" or "batch annotation") ?
>
>
> - NAG values are currently added on the basis of the current
>> positions evaluation and not according to the difference
>> to the position before the current move. This results in
>> a sequence of moves all getting e.g. "=3D" or "+/-" or
>> whatever. I feel, only the move that led to equality or
>> superiority should get the NAG value attached to it. (Is
>> the current behaviour actually a bug or a feature?)
>>
>> A feature.
>>
>
> Is it really common to NAG _every_ move once the position
> changed at some point and the advantage just stays? I never
> saw this.
>
When looking at a line you get +/-, and you don't have to look at previous
lines to see what was the latest current NAG. Three letters are not a big
waste of space.
[....]
>
> Except on Pocket PC, Multi PV mode has little impact on
>> performance (after the first couple of seconds). But if
>> you have any figures...
>>
>
>
> Hm. If the engine processes evaluation along one line it
> uses full performance on that single line, 100% CPU time
> invested into this single line. If it has to calculate three
> lines, it has 1/3, netting a 15s evaluation to 5s.
It does not work like that. Look at code in engines : the only waste of tim=
e
for the engine is that it has to sort the lines 2 to N depending on the
scores. Hardly noticeable. There are far more waste of CPU cycle in the Tcl
code in Scid used to translate piece characters in native language. But eve=
n
there (I made some benchmarks) it is not a real problem (and unnoticeable
after depth > 6 because lines don't change so much after that, so few
translations are needed).
> This does
> not necessarily mean that you "see" it on a 3GHz CPU and it
> might be the reason why you see it on your CE.
>
> Still it should not overwrite my setting ;)
>
> I feel there's a slight "bug" in UCI init code, ie.
>> multi-PV is always reset to the value in engine config and
>> not the value currently set by the user in the engine
>> analysis window. Ie. if I set Shredder to use 3 lines in
>> the engine config, fire up the engine, reduce this to 1
>> stop and restart it it is reset to 3 lines in auto
>> analysis.
>>
>> Of course, this is the correct behavior. The engine setup
>> is there to be used.
>>
>
> Sure. I tell it "calculate 3 lines by default". Perfectly
> sensible, the engine starts up with 3 lines. Then I reset it
> to 1 line in the analysis window. Surely cause I had some
> reason to do so. This should IMHO be respected.
It is. But if you stop / start the engine it is normal to get default user
settings back.
>
>
> If you want a particular value for MultiPV, engine config
>> is the place that has the highest priority.
>>
>
> I think, the highest priority should go to the user
> overwriting the default in the analysis window itself.
I don't think so.
Pascal
-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Scid-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scid-users
