This site is in archival mode. A replacement is being developed. In the meantime, please use the PBW2 Forums for community discussions. The replacement software for this site will use a unified account system with PBW2, and any newly created threads will carry over.
Okay, progress is being made. "Project Parsebuster" - that's what I called dumping in all the remaining hulls and techs and components (those for the 6 special civilizations). In under 2 hours I got the mod to parse and load.
Testing, so far the only mistakes have been trivial graphics issues. Well, that and I had to take out a few formuas for Engines per Move; the game only allows simple numbers there too. Using a formula results in a ship with zero "Ground Movement". Kinda funny.
I'm really enjoying my test games. Although I've had to drop several planned features, the game has turned out more fun than I expected - for me at least.
Still have to write the AIs for the special civilizations. Back to top
Posted: Thu Aug 07, 2014 3:17 pm Post subject: Progress is slow
I got so involved playing my test game, that I almost couldn't bear to pull away to write AI's, until a few days ago. I actually had most of the work for the first one done early in July, but I kept procrastinating.
Anyhow, I finished it and started on a test game. There was one hitch in the research, and once I got it sorted out the thing seems to be doing very well. I had hoped for a little more variety in the hull selections than I'm seeing, but for a while it looked like there was none, and unsticking the research got it mixing things up a little. It's wonderful to see it working. So far it's designed ships using 6 custom hulls, and 2 standard.
Another big plus: I fooled around with defense base design, and got it doing more like what I wanted.
So, one custom AI done, and 5 to go. Now that I have one to use as a model, or template, or whatever, they should be much easier. Still, it's tedious and time-consuming doing some parts. "On paper", I could knock one out in 3 or 4 days, but it's much easier said than done. I gotta find a way to stay after it - thing's so close to being done! Back to top
At the rate of one custom AI per month, it'll be too long. That's the pace I'm working at, and I need to either pick it up, or issue a preview version with some AIs still not done. I'm confident I'll have the 2nd AI finished by the end of the month, but somehow I need to pick up the pace.
Oh, and I did find a way to "break" things. My fancy-schmancy research scheme works really well, but I recently discovered a way to build research facilities and get a lower total afterward. The trick is to use computers to boost some, but not all of your research. Building a new facility will slightly decrease the output of ALL of them, and with some facilities output being multiplied, it makes the reduction on them get multiplied. If the multiplied reductions add up to more than the amount the new facility adds, the grand total can actually drop.
So the trick to getting "correct" results is not to mix things much; either use computers to multiply all your research, or don't use them very much at all.
Really, I didn't know anyone was anticipating release. I think this mod is going to blow a few minds. Nearly every "ability" listed (and functional) can be used one way or another, and that's just a bare start of a description.
I'm reluctant to "advertize" too much on reduction of turn processing times. The pace is slower; there are fewer colonies and ships at any given date, and this translates to whoppin' big reductions, as the game proceeds. My scripts are bigger, and aren't written for speed; they actually take a lot longer to process; but that's just a fraction of the overall time.
I've made a policy similar to EA Sports' "If it's in the game, it's in the game". Mine goes "If it's in the game, it's in the AI". The only things humans can do that AIs cannot are cloaking, and playing intelligently. The AI will use everything except its cloaking equipment, but it isn't usually very bright about how it uses things. I did try even, to get it cloaking, but that's believed impossible, and I cannot report any success. It's a shame, too, because it'll sometimes produce sweet cloaking designs. Back to top
It's better to break up a large change into a series of small releases, otherwise a bug turning up in earlier work could invalidate later changes.
Beta testing a large release in one go usually leads to tears, ask any users of brand new Government IT projects.
One speed up technique that might work is to define large faclilies/ship components. These would be 5x or 10x the size of their originals.
A 25 slot research world could be made of 2 10x research elements and 1 5x one. (The AI would still have to use the 1x sizes ?)
The game seems to process each element on a ship/planet separately each turn. The above example would appear as a 3 element planet as opposed to a 25 element one - shorter lists mean faster processing. Back to top
It's better to break up a large change into a series of small releases, otherwise a bug turning up in earlier work could invalidate later changes.
Beta testing a large release in one go usually leads to tears, ask any users of brand new Government IT projects.
Thank you. I expect that's good advice, but it's too late now. I did break things up, but I've been the only one testing. Nobody showed much interest before...
Perhaps I am mistaken, but I'm confident I have found most bugs. My unit pick-up code for the AI remains untested, and that's a concern. I introduced hybrid carriers, and I need to know whether or not they'll manage to load up fighters & drones. I also coded to have combat ships pick up troops if they have cargo bays, and then get assigned "core" in the landing TFs, and that is untested as well.
The first of three potential new unit types is bugged, and probably won't ever work right. I don't plan to proceed with the other two in the initial release, and I may not ever bother with them. They'll definitely have to wait until I'm confident I can persuade the AIs to pick them up, and in the case of one type, deploy them. They're just extra frills, but I wanted to get most of the preliminary work done, so it won't be quite so hard if I decide to go ahead and add them.
Quote:
One speed up technique that might work is to define large faclilies/ship components. These would be 5x or 10x the size of their originals.
A 25 slot research world could be made of 2 10x research elements and 1 5x one. (The AI would still have to use the 1x sizes ?)
The game seems to process each element on a ship/planet separately each turn. The above example would appear as a 3 element planet as opposed to a 25 element one - shorter lists mean faster processing.
That's a clever idea, and should work. I might consider it for later on. I insist the AI has to use everything, and this suggestion would be a fair project on their end. I haven't familiarized myself too very well with how the AI adds facilities. I only have a few to bother with, and my code there isn't elegant; I just find a good place and insert the code to add the facility I want when the conditions are met. I didn't even write in any potential to scrap my new facilities. My rule is only that the AI use everything - not that it must use everything intelligently.
Thanks again for the encouragement. If things go really well, and I do pick up the pace, I might be able to finish two of the custom AIs within about a month. That'd give me a total of three out of the six, and I just might release a preview version so volunteers can start in. Shoot - if anyone could round up a trio for multiplayer, all three of the remaining custom civilizations could be "covered".
The main drawback there is that the mod has a learning curve, and it's better not to start with the custom "races", but instead just stick with stockers. I've gone bankrupt thrice myself, including my first two games in this economy. I recovered, but I don't like the idea of people getting a bad taste in their mouths. Once a player learns what's what, it is still possible to colony race and tech race considerably (that's how my third bankruptcy happened - I got cocky). First-timers have no chance, if they try to do more much more than survive.
Another drawback is that without experience, a player doesn't really understand the mix of advantages and disadvantages, which the custom races have.
---
Now I suppose I might also explain how come I've been so secretive. There are lots of surprises in store, and "spoilers" can ruin things. In order for a player to enjoy the full "artistic effect" of the mod, the surprises need to remain unknown until they are encountered in game. Ideally, I desire players to underestimate the mod, to think they know what it's all about when they've only begun to discover. Back to top
Strip out the custom races in an early pre-release, just have the basic stock ones present.
The underlying changes can get a workout whilst gaining familiarity with the mod.
The extra complications/subtleties of the newbies can wait for later.
I only do single player, but I'm happy to give it a try on that basis. Back to top
From a strategic perspective, 10x scale facilities are useless. Since you cannot gain any output until the whole thing is completed, you lose out on a lot of production points. If you build a 10x miner, then your planet will not produce any resources for 10 turns as it gets built. On the 11th turn, you will have netted a total of 10 units of production from that final turn. If you build the 1x facilities, on the other hand, you will get 1x unit on the second turn from the first completed miner. On the tird turn, you will have generated 2 + 1. On the fourth turn, 3 + 2 + 1. By the 11th turn, when all 10 miners are finished, you will have produced 10 + 9 + 8 + 7 + 6 + 5 + 4 + 3 + 2 + 1 units of production.
From the 12th turn on, the marginal production output of each strategy is the same each turn. The opportunity cost of the 10x facilities is rather significant, though. In a competitive environment, a naive player using the 10x (or even 5x) facilities will fall behind because they will not have as many resources to expand with in the early game. If you program the AI to use them in place of the 1x facilities, then any human player that figures out the cost will gain another unfair advantage against AI empires.
If you remove the 1x facilities and only have 5x versions in the mod, then there would be no competitive disadvantage. Smarter than your average Texrak. Back to top
Give the larger facilities a reduced maintenance and/or construction cost to compensate?
They would also be harder to knock out completely, I suspect Weapon Platforms etc. would last longer with these on planet during a bombardment.
It was suggested primarily as a speed up device in the context of what appears to be doable via mod scripts.
If it were possible to get at the underlying code, a much better way would be to re-factor it to cut out a lot of the (unnecessary?) list processing. Back to top
Posted: Wed Aug 27, 2014 11:33 pm Post subject: Progress Report
I haven't been moving as quickly as I had hoped, but I did get the second custom AI finished. I put together a 7-zip preview for uploading (105 MB) but I'm going to hold off with that for a couple of weeks.
I just discovered a slew of problems with one of the custom civ's. Their techs, hulls, and components don't work right. They don't get some things they should, and they're getting things they shouldn't. It's a big deal, since they aren't supposed to be able to colonize early, but rather rely on conquest. They need all their advantages, and I need to get colony ships out of their reach. I though I'd done all that, but something's very wrong.
I wish I had more time for testing, but I don't. Anyhow, in about two weeks, I intend to share what I've got so far.
I think I have a faster way to do the AIs now: I'll work on them in parallel. It'll save the time it takes me to figure out what I did before and need to do, each time I go to work in a given area. Some of these things are confusing, and getting confused once beats getting confused four more times. I'm really pleased with how things went, and I expect this second custom AI to play a fairly good game.
I also discovered a problem with my first custom AI, and corrected it. They should do a proper job selecting hulls now. It had been satisfactory, but wasn't really what I intended. My test game is so much fun I hate to abandon it, but it'll be time to move on when I next have the chance to play some.
So... two custom AIs are done, compiled and ready. Four more remain. One has non-AI issues I'll need to either sort out ...or cancel that civilization. Back to top
Posted: Thu Sep 11, 2014 11:33 pm Post subject: I Tried
I'm sorry, but it's going to be another couple of weeks. Real world issues keep taking me away. I have to leave again. I tried to upload a 7z of my mod here, but after 3 failures, I'm out of time. I don't have time to go hunting for file hosting either, or figure out the problem.
Zero progress was made on the AIs during the last two weeks, but my most recent playtesting detected and allowed me to fix a major, major bug in the techs of one custom civilization. Simple really: 'Avian' was misspelt "Avain" in a couple of critical places, but not everywhere.
Sorry for the way things turned out. Even this incomplete version is a lot of fun, and I really want to share as soon as I have the time. I'd still like to present the complete, finished mod, with all the AIs and everything working, but that's looking to be months down the road. I just don't have the time I used to.
Anyhow, when I return again, getting this thing hosted will be a priority. Back to top
My bad, on that earlier deal. The file is too stinkin' big. Anyhow, it's available now. I'll be starting new threads to discuss the Space Empires at War Mod. Back to top
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum You cannot attach files in this forum You can download files in this forum
All logos and trademarks used on this site, all comments and stories posted for reading, all files hosted for download, and all art work hosted for viewing are property of their respective owners; all the rest copyright 2003-2010 Nolan Kelly.
Syndicate news: - Syndicate forums:
Page Generation: 0.29 Seconds
.
?
- Syndicate forums: