re: needed clarification + methodology
by A. Turin on Monday April 19th 2010, 20:36
>>I guess you refer to the MPE here. Wasn't able to get it running either.
The entire software development kit independent executables, including chickenkiller, damagetool, not just MPE.
>>Well not really part of the game, but..
I know. I tossed it in since I was able to get the game to run properly in widescreen on windows, and it can be considered a mod.
>>A d3d error shouldn't have anything to do with Phaser accuracy.
It doesn't. Between normal inaccuracy and the fact that the beam weapons 'stutter' instead of immediately firing as they should, it greatly reduces the efficacy of beam weapons as a whole. It is most painfully clear when fighting enemies using dominantly disruptor and/or torpedo-style weaponry. Most AI (including Felix) seem to try to fire, and when that fails to properly fire the first time, they stop trying to fire their beam weapons for a little while, cycling through their AI subroutines. So this also makes enemies that have mostly beam weapons way too easy. Due to circumstances beyond my control, I cannot continue to troubleshoot the issue as I'm to the point where I need more hardware.
>>To check you can try to disable "Foundation Tech" in KM 1.0 or KM2010.02.
A side note about this. A very large amount of current-gen mod content depends on F-tech. Heavily modified vessels can sometimes depend on it entirely. Back when I was thinking that it was related to inaccurate phasers, I had to locate the two specific scripts that directly controlled it and comment specific lines out. Once I can find my case notes of how I did that, I'll probably post it for reference. Either way, the removal of phaser inaccuracy had no effect on the incidence of the glitch.
>>Also please note that your "stability issues" probably mainly come from BC, not wine. So your experiences that some wine versions play the game more stable can be just luck.
For testing purposes, I established a baseline in Windows XP and Vista (as well as their equivalent wine modes) with the same hardware setup (swapping in and out multiple system drives), using defined settings that do not change (because any time I do change those settings, I have to re-establish the baseline). A similar baseline was established playing with each vessel for approximately ten hours each to give me a fair idea of how each vessel fits in standard deviation. Notes of known issues, their frequency of incidence and thus a regular tolerance between in-game handler crashes and API crashes allowing me to plot stability data with any degree of reliability.
The internal console tracker mod, external konsole output and debug packages set to verbose traceback by default as well as six separate stress tests (one on one 256^2, 512^2 and 1024^2 res models, three on three 512^2, up to 30 minutes at DS9 playing as the 2048^2 USS Challenger from the DJGalaxy pack against Dominion waves from DS9FX and no less than 12 hours at warp, usually Earth to Romulus as I've calibrated the distance) done several times over a week will typically give me quite a great deal of information. It grants me a relatively dependable judge of stability (as very few crashes go completely unexplained, and I'm easily able to identify internal issues and not count them against the wine version) with a tolerance around ±15%. That's a large tolerance, I know, but what it allows me to do is report only significant changes in general stability with as little influence from typical variance during a too short sample period as possible.
While I can post the large amount of raw data that leads me to believe there is a significant delta stability, it's not needful to do so. I'm not going to report higher or lower stability within the baseline tolerance of ±15%, and as always 'Your Mileage May Vary' due to differences in base operating system, RAM, v-card, s-card, their drivers, other stuff you have running at the time, and on and on and on. I do try my best with what I have, as a game with such a community as this one's deserves nothing less.