AF wrote:
Or the variable "bool metalmap;" you talk about which mysteriously dissapeared from the public releases in V2 and hasnt been seen since?
If you ever took your time to actually test it youd see that maps with no metal would be given as metalmaps and maps like wideopencombat would give shitty results. I improve the system by adding a working averagemetal value and fast enough execution to do enough spots to run something like wideopen and you say that.
This is like the huge wasteful discussion we had before because you couldnt write a simple function to pick the closes spots, blaming my class when everyone else seemed to have magically got it working. I wont delve into that though, if you dont like it, dont use it and code your own.
Oh thats the problem you insisted I didnt have such a class and that it wasnt working correctly.
Infact you're seeming to forget the raw output I took directly from your class and outputted to an image showing exactly where each and every spot was that your class found.
Infact i even remember you admitting and giving a solution!
I even remember running the same map through cains original algorithma nd getting the CORRECT results having only changed the algorithm and not the class that you blamed was at fault, thus proving it was the algorithm at fault.
I DO HAVE A CLASS THAT PICKS SPOTS AND I HAD IT THEN I HAVE IT NOW AND I'VE HAD IT FOR A LONG TIME
I keep telling you that but you dont listen, instead you ignore me.
What doy uo think is in the MetalHandler.h/cpp files of NTai then? A bedtime story? A classto find how many people are needed to change a lightbulb? It couldnt possibly be a class that takes a copy of your algorithm and uses it to get spots, then handles the palcement and extraction ranges wiht those spots, can deal with multiple extraction depths for mohos and normal mexes, even handles filtering out spots that are known to have enemy mexes on them.
The cycles of krogothe!!!!
## Bugs and errors
1# someone finds a problem
2# Ignore the problem
3# attack the problem
4# admit in private and give solution
5# pretend the problem never existed
6# goto n#1
attack any mention of it
Krogoth, you're a very strange person, because ontop of that there seems to be this whole KAI muddle of dual logic, telling people that KAI is a godlike AI that's due to be finished within the next 2 weeks and is iminent, while also telling them you havent implemented a lot of stuff and it'll be far away.
You'll say you've implemented feature x, then you'll talk about a problem with feature y, then you'll talk about how fixing feature y will enable you to do feature x, then you'll say you have xyz features that've been tested, then you'll say they'll be implemented by the end of the week. Make up your damned mind!
dude, chill! and i said public testing would start in 2 weeks, and never ever said it would be godlike, read the posts properly and without emotions fogging your mind. Last i checked you claimed NTAI was uncrashable, guess what it crashed last night, several times!
lastly i ask you to please stop hijacking this thread with rants when all i desire for it is bug reports and suggestions for improvements (i appreciate your first suggestion for a standard, but i agree to disagree).
And keep in mind its my AI, i can implement what i want when i want, and if i post incorrect information, it is my right to do so under the simple concepts of freedom of speech and human rights. Deal with it.
Krogothe, freedom of speech doesnt mean you can say whatever you want, nobody can say anything because there are consequences, and we live in a world with more than one person.
You dont exactly help by ignoring problems and blaming them on other things despite evidence to the cotnrary then providing a solution when proven wrong then denying it ever happened, by doing that you're provoking a reaction.
Also, since you have produced no proof, no crash logs, no screenshots, no crash data, no description of any NTai crashes, you must be lieing. NTai has exception handling everywhere exept the InitAI() routine so the only possible place it could crash is in the moment before the first frame, and those routines have been tested very very thoroughly, using mostly code that hasnt been edited in a long while that I've verified to be stable, I've ran ti through hundreds of debugging sessions.
So far nobody has provided any information on crashes that're of use, my guess is that you werent at your computer and it went into standby, or you deliberatly meddled with NTai's data folders, or installed it incorrectly.
A possible problem that I or anyone who tested XE8 did not encounter but could be remotely possible is that having AI exception handling turned on means NTai doesnt get the chance to catch possible errors and they're caught instead by Spring.
I could reel off a list of changes and errors and the code to fix them but I dont think they'd make a difference, so I'm just gonna keep working with it and you can incorporate when XE8.1 is released and I upload the source, because quite frankly, you think my opinion is worthless, and you've said nothing to prove otherwise.
AF wrote:
Krogoth, you're a very strange person, because ontop of that there seems to be this whole KAI muddle of dual logic, telling people that KAI is a godlike AI that's due to be finished within the next 2 weeks and is iminent, while also telling them you havent implemented a lot of stuff and it'll be far away.!
lol, the same applies to you :)
you keep telling ppl that the next version of your ai will be the ultimate killer uber ai (of course uncrashable as well) for at least one year now....
but of course epic will solve all the issues of ntai xe8 :)
Epic is far away, and I never mentioend it would be the uber killer AI for any of ym releases exept for Epic which isnt anywhere near release.
I've instead stressed the improvements, and said it will be the best relative to the other AI's, and for a long while it was true, public releases of my AI consecutively beat the opposition public releases, and it stayed that way untill KAI's pre-alpha release when i muddled up the buildtree and lost my lead in the process.
Well I was aiming more for stability consistency and overall improvement, AAI still doesnt have config files for XTAv6 without users manually renaming them? That was something I tried to prevent. Whatever NTai's abilities are, it consistently performs in the same conditions as it did the previous time if not slightly better, somethign AAI doesnt do, and neither does OTAI, altho you've lessend the degree to which ti happens eahc version.
