Patch?

[MarkShot]

That is some list. I suspect there is other stuff that Dave has tweaked that never got an official TT attached to it.

We, the DDT/BETA/DEV group, have been playing the same scenarios over and over again so much that no one uses their names anymore. They all have short abbreviations like FCAV or ERG. I am beginning to feel like I am really in the military now that I am learning to speak in abbreviations.

Anyway, I have played ERG (Elasson Rear Guard) so many times that I am thinking about writing a book about my experience.

[Arjuna]

Mark,
 
Yeh but I doubt the world is ready for another POW saga. [;)]

[Kabbers]

POW - LOL

now this has been one a wonderful thread so far, I have to tell you I have really enjoyed reading and sharing a little bit of your journeys to the patch ... you couldn't just, put it off for another week or two and keep the thread running could you? [:)]

Of course I am actually very excited to see what you've done in action (pulping me probably)... it's just been a good journey thus far, thank you eversomuch for keeping in touch

[Arjuna]

Well I for one am not going to delay the patch any longer. The last build has passed muster. I'll be uploading the files to Matrix today. They will then prepare the installer, we'll test to make sure it works and then it will be made available to you. So not long now.[:)]

[Pergite!]

ORIGINAL: Arjuna

Well I for one am not going to delay the patch any longer. The last build has passed muster. I'll be uploading the files to Matrix today. They will then prepare the installer, we'll test to make sure it works and then it will be made available to you. So not long now.[:)]

"Finally! It feels like we have been waiting for ever to get on with the figthing and do yet another daring air drop into the loving Cretean countryside..." [;)]

[Pergite!]

"Oh my! I got so excited by the talk about that
the patch is on its way that I somehow managed
to misplace my pants."

[TheHellPatrol]

ORIGINAL: Arjuna

Well I for one am not going to delay the patch any longer. The last build has passed muster. I'll be uploading the files to Matrix today. They will then prepare the installer, we'll test to make sure it works and then it will be made available to you. So not long now.[:)]

Damn!!! I was going to play golf this weekend for a change of pace[8|]...Oh well[:'(][:D][;)].

[BigDuke66]

I bet they won't get in ready before next week [:(]

[MarkShot]

It's Friday here. Once the actual patch has been packaged into something which can be installed by customers it still needs to be tested, at least, in a cursory fashion. So ...

On the other hand, at least you are not dealing with EA, Ubi, Activision, ... Those companies can be sitting on a complete patch for four weeks in order to make it ready for distribution. Matrix is much more nimble than that.

[GoodGuy]

ORIGINAL: MarkShot

Those companies can be sitting on a complete patch for four weeks in order to make it ready for distribution.

I'd say many of their games may be way more complex:
graphically (2D/3D -> DirectX), sound-wise (DirectSound, EAX, 6.1 channel stuff and whatever else) and may contain some AI which may display a class of its own, partially at least. And all that stuff has to be compatible with common hardware.

These games' install sizes exceed 3 GB easily (yeh i know quality counts, not quantity [8D] ) nowadays, and sizes of patch files may range from let's say 80 MB to 500 or 700 MB .... so pushing those biggies through the QA might take its time.

So, hardly comparable, imho.

[MarkShot]

Big graphic and sound files doesn't make something complex to test for a patch. Granted pushing video and sound technology does add to testing. But consider that COTA is one of the few multi-threaded engines on the market today and has been since RDOA.

The COTA code base and AI is quite sophisticated and have been the subject of development for many years. Although in some ways our testing is less professional than these other companies, those doing the COTA testing understand the game mechanics and subject material I would bet far better than those working at the larger publishing houses.

[GoodGuy]

ORIGINAL: MarkShot

Big graphic and sound files doesn't make something complex to test for a patch.

Dunno, I wasn't referring to file sizes only...
Let's say you've got issues in a 3D environment (gfx - scenarios, texture filtering, FX and whatever else), in a 3D-sound environment, and maybe even on the AI side, then you got a sh*tload of things to evaluate, and at the end of the day things have to be compatible with the most common machines to get a patch through QA. The overall effort might be somewhat bigger, that's what I was trying to say.

COTA's AI / core engine shines, and it's a real gem. But the stuff around the engine (gfx, sfx) is pretty much sub-standard.
Don't get me wrong, COTA displays a revolution in a niche sector which is ruled by hex games. Plus it's the only game in this sector that works for me, since I can't stand turn-based games..... (maybe except for the ancient Panzer General and Battle Isle back then).
The effort testing/evaluating gfx/sound bugs in such a relatively simple 2D environment might be somewhat less time consuming, though? No?

The COTA code base and AI is quite sophisticated and have been the subject of development for many years. Although in some ways our testing is less professional than these other companies, those doing the COTA testing understand the game mechanics and subject material I would bet far better than those working at the larger publishing houses.

Well, I wasn't bashing COTA's AI. I was referring to the possibility that games from these publishers may have things in their feature lists which may display artful AIs as well, and hunting down all those things mentioned above might add to a bigger QA-queue.

There might be 2 reasons for COTA's AI-programming taking yrs to develop/tweak: 'Coz it's a micro-company doing it (working on this single product), and because Dave is an eager perfectionist who tunes each and everything 'til it rocks, still adding things (by user request or own initiative) where a big publisher would have dropped the product already, to head for a different game (okok maybe after a few sequels to squeeze the customers -> EA [;)] ).

The fact that the developers you mentioned are big hairy and fat companies, where the left hand often doesn't know what the right hand does, doesn't necessarily mean that they aren't able to come up with some sophisticated stuff, (last but not least) even in the QA-department, imho [:)].
Also, I tend to think that such a micro-company needs dedicated testers (who are using their brains, where a somewhat deeper knowledge of mechanics might help the programmer to get things tracked down more quickly), rather than a fat dev/publisher.
On the other hand, software engineers in a fully fledged SQA department aren't just robots either, they often have to create/tune their own test evironments.

[Arjuna]

Patch Installer uploaded for testing now. Greg at Matrix has pulled out all stops. Thanks Greg.
 
Hopefully we can sign off in a few days. [:)]

[Kabbers]

Tell us something nice again about the new AI while we wait... one more time? The aggression is obviously a high issue, but what is the thing that most tickles you as programmer Arjuna, or MarkShot from your persp? or anyone who gets a tickle Other than it attacking more vigorously? you are both very eloquent already but more more more if you can if you are kind enough to ease the hours

Arjuna any sights you'd be willing to share on how you programmed the AI - nothing too deep and certainly I don't mean to pry, just any interesting aspects - for example, the core algorithms, how might they compare with how the a human cognitive system makes decisions, has context etc? I'm a pretty novice wargamer who's written AI in another space - little chess programs so far (Z80A ass way back, recently c#, their est ELO rating probably only circa 1800) and I know little enough from both attempts that there are tiiiny things that please and upset when you're developing AI, micro-details that others might not look at if you don't point out --- what are the behaviours that tickle?

I felt I stymied the thread a bit, earlier, so if I do the same again I'll jump on my login and beat my final beat only 2 posts in, so please remember I'm just jumpin' around waiting for the patch, really, praying Greg works the weekend, but truly: relationship to an AI is a strange thing, what makes it tick Arjuna?

K

[MarkShot]

Kabbers,

You asked for a perception about the AI engine. What makes the AI engine in this game special compared to other games also regarded as having good AIs. (Note: I am not going to explicitly mention any other products in this response, nor will in futher discussion either.)

Much of the strength RDOA/HTTR/COTA AI is within the game engine itself. Effectively, the game engine has enough semantic understanding of terrain, tasks, forces, military process and strategy to put in a decent showing on the attack or the defense. This is in contrast to other games which are much more dependent on the scenario designer to craft a situation which creates suspension of disbelief or immersion for the player through simulating reasonable military process --- as if the scenario designer was choreographing a Hollywood feature with an animation tool. Thus, in COTA scenario design is primarily objective designation with a focus on points/type/location/time window and the setting up forces. Beyond that the AI, understands how to attack, defend, delay, etc...

Other games are much more handicapped when it come to AI. This may take the form of the AI typically only defending well due to the one time surprise which is create by FOW events. Or may take a form such as comprehensive scripting where the action of each and almost every unit in the scenario is controlled via commands (resembling a computer programming language) thus allowing the scenario designer to make things happen in a believable manner.

I have recently said in the Beta discussions that I prefer the PG approach. A well implemented AI yields a significant return repeatedly across many scenarios. Crafty scenario design yields a return only for the specific scenario which incorporates it.

Disclaimer: I mean no disrespect to DDT scenario designers (and map makers) who are true artists and invest a great deal of research and work to produce scenarios that recreate history and explore what might have been. However, you asked what sets this AI engine apart from others which I am familiar with and this is by far one of its most unique characteristics.

[Arjuna]

Kabbers,
 
Good question.
 
I suppose the main reason I spend my hours designing AI routines for wargames is its creative. I derive great joy from seeing the AI do something that simulates what happens in reality. It's this focus on reality which is at the root of why our AI engine is different from that in other games. There the tendency is all too often to abstract - ie do enough to create the illusion or effect - and then as Mark said to choreograph an arrangement of effects using the traditional "scripting" approach.
 
The trouble with this approach is that it is case specific. It's not generic. It works once for the particular scenario. With a few smarts scripting can throw in some randomness but ultimately there is no real situational awareness and/or underlying strategy behind it that the AI can utilise to reach its own decision. After one or a few plays it offers no more challenge. In fact it becomes down right predictable and boring.
 
So for us the focus has been and will be on creating a generic AI. Effectively we have created an expert system. Like all AI it will never be perfect. So for a self confessed perfectionist like me it is a continual challenge. [:)]
 
Got to go. I'll be back.

[alaric99x]

Congratulations on your energetic support of the games you produce.  Of course, everybody would like the next patch yesterday, but you guys are still remarkably fast and effective.
 
Reading this forum has convinced me that I "need" this game.

[JudgeDredd]

...need...

a better adjective could not be used!

[TheHellPatrol]

ORIGINAL: JudgeDredd

...need...

a better adjective could not be used!

Apparently that goes for the patch as well[;)].

[Arjuna]

Hi all,
 
Testing results of the installer on Thu and Fri led to a flurry of work for me over the weekend. By end of Monday I had fixed all reported bugs and put out a new replacement exe ( ie game program ). Autotesting overnight gives a clean slate. No reported bugs in from testers. I will ask Matrix to supply a new installer for testing. Hopefully this time it will get the all clear.