| brad |
| Posted: Apr 13 2008, 09:01 PM |
 |
|

Administrator
    
Group: Admin
Posts: 543 (10 robots)
Member No.: 1
Joined: 7-November 02

|
I'm planning to release a new version of RB in the next month or so, and I'd like some feedback. My primary goals are to fix bugs and to make it easier for others to build RB from source (recent versions of Microsoft's compiler and libraries have issues with the current source code). I'm not planning to add features unless they are very small and non-controversial. I'd appreciate some feedback about what bugs need fixing. I'll go through the RB bug list, but if you know of things that need fixing please respond to this thread with comments and links. |
| brad |
| Posted: Apr 13 2008, 09:03 PM |
 |
|

Administrator
    
Group: Admin
Posts: 543 (10 robots)
Member No.: 1
Joined: 7-November 02

|
For example, the reported problems with GetRandom is something I will look into.
I'm not planning to fix Registry issues, unless its something really critical. |
| tomash |
| Posted: Apr 14 2008, 03:10 AM |
 |
|

Grand Master
    
Group: Moderators
Posts: 810 (8 robots)
Member No.: 441
Joined: 11-July 05

|
Wow! Brilliant 
Certainly GetRandom is important. Are you going to correct some of the documentation also, as some of it is wrong? (that is on the bug list)
In terms of features, I like some of those on the bug list; for example the name under the robot - this would be very useful when testing, for me at least.
EDIT: Also (just remembered), is there any chance you could implement a better way to increment/decrement variables like var++, var-- ?
This post has been edited by tomash on Apr 14 2008, 03:49 AM |
| steJ |
| Posted: Apr 14 2008, 09:02 AM |
 |
|

Tiny Bot Obsessive
    
Group: Moderators
Posts: 951 (13 robots)
Member No.: 9
Joined: 2-May 03

|
var = var +1 any good 
One thing that would be nice would be the ability to have the inspector open whilst still running the match in order to monitor variables in real time perhaps? I guess it will just involve setting an auto pause to false or something?
-------------------- He who laughs last, thinks slowest. SteJ |
| chess123mate |
| Posted: Apr 14 2008, 02:52 PM |
 |
|
Logical Programmer
    
Group: Members
Posts: 587 (24 robots)
Member No.: 936
Joined: 3-February 07

|
Also, one major bug (that shouldn't be difficult to fix) is in the inspector -- some variables are overwritten. I don't have the details, and I didn't post it in the Bug list. The problem was, for example, the damage from mines (or # of mines) for the game and the entire match used the same variable, so one value was "0" and the other was just the game. This applied to some (but not all) of the statistics variables.
It would also be nice if you could run over 600 bots or so in a match AND ensure that Robot Battle doesn't instantly crash when you try and view the results. Something about the table being too large... Link: http://www.robotbattle.com/registry/index....=ST&f=11&t=1442
(PS I am currently running IE7)
Of course, making functions would be so useful, but that would be too big of a project to engage in by the sounds of it. Ah well... there are ways around it |
| brad |
| Posted: Apr 14 2008, 09:33 PM |
 |
|

Administrator
    
Group: Admin
Posts: 543 (10 robots)
Member No.: 1
Joined: 7-November 02

|
| QUOTE (steJ @ Apr 14 2008, 09:02 AM) | One thing that would be nice would be the ability to have the inspector open whilst still running the match in order to monitor variables in real time perhaps? I guess it will just involve setting an auto pause to false or something? |
Sorry, but I paused it on purpose because it made the inspector simpler. This would take more work than I'm going to have time for this release. |
| brad |
| Posted: Apr 14 2008, 09:36 PM |
 |
|

Administrator
    
Group: Admin
Posts: 543 (10 robots)
Member No.: 1
Joined: 7-November 02

|
| QUOTE (chess123mate @ Apr 14 2008, 02:52 PM) | | Also, one major bug (that shouldn't be difficult to fix) is in the inspector -- some variables are overwritten. I don't have the details, and I didn't post it in the Bug list. The problem was, for example, the damage from mines (or # of mines) for the game and the entire match used the same variable, so one value was "0" and the other was just the game. This applied to some (but not all) of the statistics variables. |
I don't remember an issue like that, but I'll take a look. If you have more details it would be helpful. Start a new thread or create a bug report if you find more details, thanks.
| QUOTE | It would also be nice if you could run over 600 bots or so in a match AND ensure that Robot Battle doesn't instantly crash when you try and view the results. Something about the table being too large... Link: http://www.robotbattle.com/registry/index....=ST&f=11&t=1442 (PS I am currently running IE7) |
Ok thanks. Seems likely to be an IE7 issue since I'm sure I tested this on IE5 and 6.
| QUOTE | | Of course, making functions would be so useful, but that would be too big of a project to engage in by the sounds of it. Ah well... there are ways around it |
Probably not going to happen |
| brad |
| Posted: Apr 14 2008, 09:41 PM |
 |
|

Administrator
    
Group: Admin
Posts: 543 (10 robots)
Member No.: 1
Joined: 7-November 02

|
| QUOTE (tomash @ Apr 14 2008, 03:10 AM) | | Certainly GetRandom is important. Are you going to correct some of the documentation also, as some of it is wrong? (that is on the bug list) |
Yes, I will fix documentation issues.
| QUOTE | | In terms of features, I like some of those on the bug list; for example the name under the robot - this would be very useful when testing, for me at least. |
Probably won't get to that, sorry.
| QUOTE | | EDIT: Also (just remembered), is there any chance you could implement a better way to increment/decrement variables like var++, var-- ? |
While making RSL better is one of the things I want to do the most... I probably won't have time for any real changes there. |
| brad |
| Posted: Apr 14 2008, 10:41 PM |
 |
|

Administrator
    
Group: Admin
Posts: 543 (10 robots)
Member No.: 1
Joined: 7-November 02

|
Another thing I'd really like to do is incorporate David's work on making RB more portable |
| chess123mate |
| Posted: Apr 15 2008, 12:44 PM |
 |
|
Logical Programmer
    
Group: Members
Posts: 587 (24 robots)
Member No.: 936
Joined: 3-February 07

|
Oh, oops... mis-interpreted the statistics variables... sorry about that. I thought that "[0]" meant for the current game and "[1]" for the entire match, but it was the other way around... |
| jamie |
| Posted: Apr 18 2008, 04:46 PM |
 |
|
Semi-Retired Robot Battler
    
Group: Registered Owners
Posts: 324 (6 robots)
Member No.: 13
Joined: 2-May 03

|
Something that's been on my wish list for a while is fractional angles (for body, gun, radar), so they don't have to be whole degrees.
When tracking at long range, the angle quantization can be a real headache.
Angle quatnization in bodyaim means driving to an exact spot is sometimes impossible without using multiple segments.
I think it would make robot behavior simpler to understand.
Perhaps you could use some kind of LRU hash table to avoid recalculating all the sines and cosines over and over. Even without fractional body/gun/radar, some kind of sin/cos cache might be a good idea since robots can always call sin() and cos() with odd angles anyway. |
| brad |
| Posted: Apr 27 2008, 08:08 PM |
 |
|

Administrator
    
Group: Admin
Posts: 543 (10 robots)
Member No.: 1
Joined: 7-November 02

|
Yes, I should have done that long ago when I converted movement to be fractional. It might also be reasonable to implement approximation algorithms for the trig functions since the standard C routines are probably a lot more accurate than needed.
Of the top of my head I'm not sure how much code the would effect in RB, however, so I can't promise it will get done now. |
| ospector |
| Posted: Sep 1 2008, 07:16 AM |
 |
|

Serious Player
  
Group: Registered Owners
Posts: 88 (19 robots)
Member No.: 127
Joined: 8-February 04

|
Hullo RBfolks.
Visiting the site after 3 years! Contributed much then, and could be useful again ... Brad, I could send you math functions for efficient approximate trigo etc. Just let me know at ospectoro at yahoo dot com. Work in C++, but could write in C, and it'll be simple conversion for any other language. And thanks for having this wonderful game still kicking.
Ospector |
| steJ |
| Posted: Sep 3 2008, 07:32 AM |
 |
|

Tiny Bot Obsessive
    
Group: Moderators
Posts: 951 (13 robots)
Member No.: 9
Joined: 2-May 03

|
Gosh, has it been that long!!? Time flies hey 
Get a Rally Cup entrant in and spice it all up
-------------------- He who laughs last, thinks slowest. SteJ |
| brad |
| Posted: Sep 8 2008, 09:45 PM |
 |
|

Administrator
    
Group: Admin
Posts: 543 (10 robots)
Member No.: 1
Joined: 7-November 02

|
Sure, go ahead and send the code to brad@robotbattle.com. Progress has unfortunately been very slow. Every time I think I'm going to have some extra time for RB something else comes up. It will happen someday... just not as soon as I'd hoped. |
|