Why is no AI is so important?
Moderator: OpenTTD Developers
-
- Engineer
- Posts: 7
- Joined: 25 May 2009 06:10
Why is no AI is so important?
Why this self created AIs are so important? I am not a programmer but a grate fan of TTD the new attempts for rebuilding TTD makes me happy in a sense that now it has more features that I never experienced before. The major problem with this open TTD version "stable (0.7.0)" there is no default AIs and I didn't have the slightest idea about what to do. Now can anybody please tell me what can I do about this. I can't just except playing with nor dead nor alive AIs. I am not a programmer, so can any body have any kind of solution for this?
Re: Why is no AI is so important?
Go to donwloadable content, download an AI.
Play.
None of them are really stable at the moment though
Play.
None of them are really stable at the moment though
Re: Why is no AI is so important?
Or load an older version of OpenTTD for example 0.6.3 the last stable release with old dummy AI!
Re: Why is no AI is so important?
I think all AI's crash quite often.
Also, 0.7.0 have bug that can cause AI to crash when it is not AI's fault (if the AI got suspended inside try/catch block), so try 0.7.1-RC1 or newer version (nightlies/svn build ...), AIs will be more stable.
Also, 0.7.0 have bug that can cause AI to crash when it is not AI's fault (if the AI got suspended inside try/catch block), so try 0.7.1-RC1 or newer version (nightlies/svn build ...), AIs will be more stable.
If you need something, do it yourself or it will be never done.
My patches: Extra large maps (1048576 high, 1048576 wide) (FS#1059), Vehicle + Town + Industry console commands (FS#1060), few minor patches (FS#2820, FS#1521, FS#2837, FS#2843), AI debugging facility
Other: Very large ships NewGRF, Bilbo's multiplayer patch pack v5 (for OpenTTD 0.7.3)
My patches: Extra large maps (1048576 high, 1048576 wide) (FS#1059), Vehicle + Town + Industry console commands (FS#1060), few minor patches (FS#2820, FS#1521, FS#2837, FS#2843), AI debugging facility
Other: Very large ships NewGRF, Bilbo's multiplayer patch pack v5 (for OpenTTD 0.7.3)
Re: Why is no AI is so important?
They do? Can't remember seeing a decent crash report of yours recently. Writing one would be a lot more constructive action to improve the quality of the AIs than nagging and whining.Bilbo wrote:I think all AI's crash quite often.
Re: Why is no AI is so important?
And remember: OpenTTD isn't really stable! Its Version 0.7.0
I think 1.0.0 would be the really stable Version...
... but you can call 0.7.0 stable, too
Perhaps in 01.0.0 the AI's aren't stable too... We'll see!
I think 1.0.0 would be the really stable Version...
... but you can call 0.7.0 stable, too
Perhaps in 01.0.0 the AI's aren't stable too... We'll see!
Re: Why is no AI is so important?
Stable releases are normally meant to be stable. Not releasing 0.7.0 as 7.0 is something completely different which has mainly to do with the objectives/goals for OpenTTD.
Nevertheless, having a 'stable' release doesn't mean everything works flawlessly or that everything you run in it works flawlessly. Think about Windows. It Windows Vista is a 'stable' release, but that doesn't mean that everything program you run under Windows Vista is stable.
Same with AIs under OpenTTD; the 'old' ones didn't visibly crash, but did they do much good besides entertaining you with stupidity? Shouldn't we aim for a better AI? But what is a better AI? Everybody has a different opinion about it, so we give everyone the freedom to build and/or chose an AI.
Nevertheless, having a 'stable' release doesn't mean everything works flawlessly or that everything you run in it works flawlessly. Think about Windows. It Windows Vista is a 'stable' release, but that doesn't mean that everything program you run under Windows Vista is stable.
Same with AIs under OpenTTD; the 'old' ones didn't visibly crash, but did they do much good besides entertaining you with stupidity? Shouldn't we aim for a better AI? But what is a better AI? Everybody has a different opinion about it, so we give everyone the freedom to build and/or chose an AI.
Re: Why is no AI is so important?
Well, the way I test things is usually to set up 14 random competitors (toss the snakes in the bin and shake it ... not exactly a typical setup :), then press fast forward and walk away .... usually about 2/3 of them end up as crashed after half an hour once I return to the window.GeekToo wrote: They do? Can't remember seeing a decent crash report of yours recently. Writing one would be a lot more constructive action to improve the quality of the AIs than nagging and whining.
And why I have not reported it?
Because there were too many crashes (sometimes I restarted one of the crashed AIs and see it crashing again after short while. Much faster than I could possibly report). Which mean too much time spent with bug reports (no copy & paste, so take screenshot, compare with other crashes, compare with already reported crashes in the forum...), while I am trying to do something else (writing and testing my OpenTTD patches).
Few times I looked at the relevant AI thread, but I usually saw very similar crash already reported ....
Once the crashiness of AI's reduce to "quite rare", I'll start caring more about bug reports for AIs. I didn't mean to whine, I just noted that in general, AI's are not very stable currently.
If you need something, do it yourself or it will be never done.
My patches: Extra large maps (1048576 high, 1048576 wide) (FS#1059), Vehicle + Town + Industry console commands (FS#1060), few minor patches (FS#2820, FS#1521, FS#2837, FS#2843), AI debugging facility
Other: Very large ships NewGRF, Bilbo's multiplayer patch pack v5 (for OpenTTD 0.7.3)
My patches: Extra large maps (1048576 high, 1048576 wide) (FS#1059), Vehicle + Town + Industry console commands (FS#1060), few minor patches (FS#2820, FS#1521, FS#2837, FS#2843), AI debugging facility
Other: Very large ships NewGRF, Bilbo's multiplayer patch pack v5 (for OpenTTD 0.7.3)
-
- Tycoon
- Posts: 1828
- Joined: 10 Jul 2006 00:43
- Location: Spain
Re: Why is no AI is so important?
But the new AIs doesn't use new trains with new cargos.
Sorry if my english is too poor, I want learn it, but it isn't too easy.
- [list][*]Why use PNG screenshots in 8 bpp games.
[*]Caravan site New Industry. · Spain set. · Some spanish trains for locomotion[*]Favourites:GRVTS · ECS · FIRS
Re: Why is no AI is so important?
Is that something the 'old' AIs did? The old AI didn't even know about refitting for trains.maquinista wrote:But the new AIs doesn't use new trains with new cargos.
Re: Why is no AI is so important?
My my, quite a heated argument here. There seems to be quite a high expectation from some OpenTTD players regarding the state of AI. From the perspective of an AI writer, allow me to say a few words.
One common misconception I notice is that writing an AI is viewed as a simple thing to do. But one has to take into account that because things are considered 'easy' or 'common sense' for humans does not mean the same holds for AIs.
For example, when you start a game of OpenTTD it is quite 'easy' for humans to judge the terrain and conceive a grand scheme for your transportation lines. Later in the game humans only need one glance to notice traffic jams and act accordingly. These things make 'common sense' right? Well, let me introduce you to the world of AIs: consider you can only examine 1 tile of the map at the time, only the tile your mouse points to, the rest is pitch black. As for detecting traffic jams, the same applies, you can only inspect 1 vehicle at the time and only your own vehicles. It is kind of hard to check if your vehicle is standing still because of a breakdown of a vehicle in front of him or if it is waiting for an empty truck station. The list goes on.
Besides, most / all AI writers do not have nearly enough time available to work on their AIs as they like. To take myself as an example, my study takes up all my time at the moment so if I can work a weekend each month it is a lot. So current development goes very very slowly but it is growing. Despite some crashes and unsolved bugs I would say that the current state of the AIs far surpasses the old AI and there is no reason to complain. At this state more and more AIs are focusing on train support and I do not think it will take long before an AI is introduced which features all possible transportation types. I think there are about 6-7 active AIs under development at the moment so there should be one to your liking.
And like GeekToo suggested, if you notice a problem please do notify the corresponding author. This will both help him to fix bugs and enhance the AI in question and in doing so provide a better experience for the end user, you.
One common misconception I notice is that writing an AI is viewed as a simple thing to do. But one has to take into account that because things are considered 'easy' or 'common sense' for humans does not mean the same holds for AIs.
For example, when you start a game of OpenTTD it is quite 'easy' for humans to judge the terrain and conceive a grand scheme for your transportation lines. Later in the game humans only need one glance to notice traffic jams and act accordingly. These things make 'common sense' right? Well, let me introduce you to the world of AIs: consider you can only examine 1 tile of the map at the time, only the tile your mouse points to, the rest is pitch black. As for detecting traffic jams, the same applies, you can only inspect 1 vehicle at the time and only your own vehicles. It is kind of hard to check if your vehicle is standing still because of a breakdown of a vehicle in front of him or if it is waiting for an empty truck station. The list goes on.
Besides, most / all AI writers do not have nearly enough time available to work on their AIs as they like. To take myself as an example, my study takes up all my time at the moment so if I can work a weekend each month it is a lot. So current development goes very very slowly but it is growing. Despite some crashes and unsolved bugs I would say that the current state of the AIs far surpasses the old AI and there is no reason to complain. At this state more and more AIs are focusing on train support and I do not think it will take long before an AI is introduced which features all possible transportation types. I think there are about 6-7 active AIs under development at the moment so there should be one to your liking.
And like GeekToo suggested, if you notice a problem please do notify the corresponding author. This will both help him to fix bugs and enhance the AI in question and in doing so provide a better experience for the end user, you.
Re: Why is no AI is so important?
Bilbo: Hope you accept my apologies for my last post, that was unneccesarily rude. I like to withdraw the nagging and whining part of that, especially since you are one of the forum members that is contributing with patches, and suggestions for the AI API. Though I still support the report issue.Bilbo wrote:I didn't mean to whine, I just noted that in general, AI's are not very stable currently.GeekToo wrote: They do? Can't remember seeing a decent crash report of yours recently. Writing one would be a lot more constructive action to improve the quality of the AIs than nagging and whining.
I feel the need to explain how I came to this post and why it was not meant personal, you were just making the wrong post at the wrong time.
Like Morloth said, we contribute to this game in our limited spare time, and when I made the post above I had justed scanned the reactions on a few projects of mine:
- First, I posted a graphic for what I thought could be a nice branch http://www.tt-forums.net/viewtopic.php? ... 60#p788160, and did find it was completely ignored, no reaction whatsoever (which I dislike even more than someone saying they absolutely hate it.
-Then, I looked at an update of my extra zoom patch, that I spent quite some time on, because it was a pretty difficult merge: http://www.tt-forums.net/viewtopic.php?p=790396#p790396. Apart from a nice guy presenting a Windows binary, and a thank you from Maquinista (who is too good for this world), ignored again.
-My previous effort in the extra zoom patch was a modification for the company colours, which took a lot of hours to develop, and it generated just a couple of reactions of people, completely missing the point what I was trying to achieve
-At this point, I was already getting a little annoyed, so let's have a look at Convoy: a guy posted a complaint some time ago, that Convoy stopped building things after a while. I actively asked about feedback, because I want to improve things when I can, but as you would have guessed by now: never heard from him again http://www.tt-forums.net/viewtopic.php?p=772471#p772471.
-And then finally, multiple weeks ago, I posted a version of Convoy with train support http://www.tt-forums.net/viewtopic.php?p=787094#p787094. I have spent an incredible amount of hours to improve and test it, and finally I thought it was good enough to put it on the forums. I know it still has a lot of problems still, and thought people would love it, or hate it, and flush me with bug reports. But the one thing that I did not expect, was that it was not being reacted upon at all, apart from one justified, solvable problem report.
Around this time, I was a the point of thinking:"why am I putting all this effort in, when nobody seems to care". Personally, when someone makes a bug report, I don't take that personally, but when it is a justified report, it is a chance to improve things, and to make Convoy better. I create things, mainly because I like to do so, but having a reaction upon it, negative or positive is always a nice reward.
Now, at that point, I read your post: all AIs crash quite often, and that led me to make my last post.
I hope I've explained a little bit my mood when I was posting that, usually I'm pretty harmless, and sometimes even helpful. I guess now it's your turn to call me a whiner
Who is online
Users browsing this forum: No registered users and 16 guests