Open Source GUI Wish List

[HumbleProgrammer]

As a professional software developer, I have been using "home-rolled" chess GUI software for a year or so; however, since Matthias clearly wants to move on from ChessGUI, I plan to release my software as an open source project. It is written in C# for .NET, which means that it will initially be only for Windoze, but I'd really appreciate someone with Mono experience willing to test-drive it. (Dear trolls: please don't turn this into a "why we hate Micro$oft" rant.") If there's enough interest, I would be looking at an Alpha delivery around Christmas, with a Beta by Easter and a finished product by summer of next year. (If that seems like a long time, please keep in mind that the software is already 100,000+ lines of code...)

Here are my minimum requirements:
(1) Windows XP or better / 32-bit and 64-bit compatible / multi-core aware
(2) WinBoard and UCI compatible
(3) Human vs. Human / Human vs. Engine / Engine vs. Engine
(4) Blunder search and unattended analysis of PGN games
(5) Full PGN import/export/annotation capabilities

And a list of things that are NOT going to be offered (yet):
(1) 3D graphics -- I simply don't have the artistic skills
(2) ChessBase compatibility -- it's proprietary and I don't want licensing "issues"
(3) Internet or ICCS play

What would be the most help is ideas, suggestions, and requests from the user community (that's you guys & gals) about what features are most important. Feel free to refer to existing GUIs (Aquarium, Fritz, Winboard, ChessGUI, etc.) for observations like "I really like having two different engines analyzing the same position at the same time." or "I want to be able to create my own piece images." or "I love/hate the ribbon UI".

Your throughts?

Humble Programmer
,,,^..^,,,

[Matthias Gemuh]

As a professional software developer, I have been using "home-rolled" chess GUI software for a year or so; however, since Matthias clearly wants to move on from ChessGUI, I plan to release my software as an open source project.
Humble Programmer
,,,^..^,,,

Hi Humble Programmer,

don't let my ChessGUI interfere with your GUI in any way.
The more GUIs we have, the better !

Cheers,
Matthias.

[Peter C]

I'm working on a Free GUI written in Java (*cough*true cross-platform support*cough* ), though I don't know when it will be ready. Time for some competition.

No ribbon, please.

Peter

[Uly]

I hate the ribbon UI - If it came with it I'd avoid the sofware like the plague

Anyway I'm looking forward for a software that reunites my two main used features: Solid analysis of Correspondence games, and 2: Consistent storage of analysis. Currently I'm switching all the time between Bookup and Shredder Classic, and while it's very effective, only having to run a single application would be nice.

What I use on Shredder GUI:

• Installation of Winboard and UCI engines, with easy access to change the engine parameters, creating personalities, and access to the engines' text files so one can edit buggy engines if needed (like Rybka 3).
• MultiPV support. With the F4 and F5 keys, I'd like to have access to this on a toolbar.
• Searchmoves support. With Shift + Drag piece a move is excluded from analysis, with CTRL + Drag Piece only this move is considered, and repeating the action adds more moves. I'd like that this feature could be done without needing a keyboard.
• Feature to force the engine to play a move at a given depth/time, or to play against itself/another engine from the current position until end of game/the user stops it. Every move would include in the notation the score/time per move of each move.
• Supercustomization: The user can move around any part of the layout and put it everywhere, change its size or make it disappear, with dockable windows (so that it allows max usage of space).
• The GUI looks Pretty , but anyway, some anti-aliased, shaded piece fonts could do the trick, even Winboard has that.
• The GUI is lightweight, it starts up immediately and only uses 14MB of RAM while running.

And, that's it , sounds simple, but presentation would be a high priority for me.

What I use on Bookup:

• A Book-like format that has the option of just adding moves by dragging the pieces.
• Powerful transposition support (with a click all moves that transpose to other positions in book appear).
• A big comment box where one can write whatever wishes, different comments can be stored per move. Transpositional moves share the same box. This is the killer feature because every user can use the box for its needs.
• Variation count with backsolving (this is a complex topic and no GUI does it right, I'd be happy with just a number that appears beside the move that increases the more positions in book).
• Ability to easily surf the tree (classic ¦< < > >¦ buttons).

And I think that would suffice for a start, with just a program that included a good implementation of such features I'd be ready to make it my main GUI, though of course it's important how much it takes to use it (for instance, with Aquarium it was faster to switch between these two GUIs than to just use Aquarium).

Though for the fusion to work, new features would need to be implemented, like "asynchronous" trees, where you can have a different position in the Notation and in the tree, but those are just details.

Also, you don't need artistic skills to have a 3D board, you could just implement isometric board rotation, here's an example:



You can have that kind of look by just using different coordinates for where the board and pieces are drawn. But it's just a gimmick.

I'm always looking forward for a be all, end all GUI , one that makes all the other ones unnecessary, so every time I hear about a new GUI I think "maybe this is it!"

[AnthonyTheSage]

I think simplicity is the key. There are enough guis with so many features that just waste time.

I would like to see a gui devoted to analysis. I think arena has the E Vs E covered. And I think there are enough database programs out there.

I barely even use most of the huge databases that I have anyways. I mainly use them for openings and after so many moves your out of the database anyways..........So all you really need is a huge opening tree instead of a database. The key to the opening tree is annotation. Such as Kings Indian etc. etc., winning percentage, and what the moves lead to. Opening tree is one key piece that I think alot of newer guis get wrong. It should be simple and right in front of you, you shouldn't have to scroll up and down and all over the place to find what you want. I find myself going back to chessmaster 9000 alot to study openings. Take a look at CM9K opening tree, awesome annotations simple, and easy to read.

Another thing that really annoys me about most guis is the move list. Why can't they make the move list simple and easy to read. For example the old scid used to have and option to view the move list in column view with two moves per line, and then variations in between. It was pretty good, but I think it can be improved upon. Shredder classic is a great example of how not to do it. Once you start adding annotations and comments it just becomes a mess. Aquarium is slightly better than Shredder but still a bit messy. I'm sure it can be improved. Also don't forget to add the option to have algebraic notation, I'm old school and I know alot of people that don't like to use the modern chess symbols. Mainly because when I play a otb game I don't want to waste time with the symbols. So if I"m always using the same notation it makes it simpler. English algebraic short with the options for changing check symbols etc. Also option to change font and colours of mainlines and variations etc.

Easy option to insert mainline from engine........ALSO option to add second or third line in multpv mode.

Setup position is a must have. I don't understand how some programs can miss this option, a bit ridiculous. setup position should be simple, with options to play engine vs engine, human vs. engine, or analysis. Again Chessmaster does this well.

a Good gui should have alot of different skins and piece sets. Reason being, serious chess players spend many hours analyzing games, and its really hard to look at some skins for a long period of time. For example to baby blue of aquarium is getting really annoying.

Winboard support would be good. I know all the strong engines now adays use uci, but there is still some cool Winboard engines out there.

MultiPv is a must. With option to analyze with many different engines at the same time.

A simple auto analysis. This is were chess assistant really gets annoying. I have Chess Assistant 10 and it used these really annoying levels for auto analysis, 1 sec blunder check, 5 sec main check, blah bla bla. Take a hint from shredder on this one>Seconds to spend analyzing each move, thresh hold value(when to insert the line) whether to analyze both black and white, or just one colour. thats about it for auto analysis. I like to analyze openings myself as long as there is a good opening tree.

I'm a big fan of floating windows such as the old winboard, or scid, or take a look at Kvetka. Kvetka is a small program that has alot of potential. With some missing options added I think it could be a winner. I know there are also alot of people that like docked windows, so i guess an option to switch back and forth is a must.

Another flaw in some guis is they make it so hard to open and save games. Sometimes they want you to create a new database, convert, etc. etc. It should be simple to open one game at a time if you like, also to save one game at a time. Sometimes you just want to move one game around.

Well I gotta go, thats all I can think of for my wish list right now. If I think of anything else I will post it.

[AnthonyTheSage]

Just thought of one more must have before I run. The ability to save layouts is a must have. Once you get your windows the way you want, whether they are docked or floating, you have to be able to save them that way. Some programs miss this option, and you have to set everything up everytime you start the program, very annoying. This is also good because often a layout that is good for analysis, is not good for engine vs. engine etc. Cheers.