because then there would be a reddit thread where eighty people coordinate to play one copy of a game and then what text_fish said.

Okay, maybe there would have to be a limit on your time played. Similar to getting a refund, but more lenient. 

if I'm being honest I think I would prefer the Unreal engine 1 implementation of pcx textures. Which is basically that every texture has its own unique palette. File size would be much smaller than using .tga files, it would be self-contained with the .bsp file for neatness. 

That would require modified compilers, a new bsp format, and new engine support. 

File size? Really? 

I'm sure EricW would be happy to implement all three. ;)

He's a god-damn wizard you know. 

 

Yeah but I just think its dafter than a box of weasels to have to modify the code of a number of tools, engines and file formats, to support baked-in unreal-style pcx textures, when external TGA support is overwhelmingly superior AND backwards compatible AND available right now :}

Filesize is a non-issue imo. 

We'll agree to disagree. ;) 

Fitzquake / QS already supports actual pcx textures, you could put them in a oak file with your map if you don't like loose files. 

Oak = pak, thanks iphone 

Fifth, been there, done that - halflife bsp.
palettes still suck, as do standard texture resolutions.

storing image data inside your bsps themselves is not good for file sizes, even if they're RLE pcx files.

either way, (gl) engines generally already support both pcx and tga external textures.
If you want to strip the textures from bsps all you have to do is to set all the offset values of your mipmap images to 0 and strip the image data. Its not a problem unless an engine can't load the correct external texture, in which case you'll just get corrupt textures rather than anything crashy. 

palettes still suck, as do standard texture resolutions

Regarding texture resolution, I beg to differ. Quake's pixel vision is a seriously legit art style that would still be relevant if someone wanted to make more colourful, but still retropixelly textures.

If you are one of those heathens that plays quake in GL blurry mode, then ugh. 

hey let's have this argument again 

I play Quake with linear filtering and model interpolation disabled. 

with TGAs, you can always palettize and maintain pixel resolution yourself. at least the option is there either way. 

 

yeah tgas can also be 8-bit if you want to save those precious kilobytes, or want to maintain the retro cred.

Interestingly, initial tests indicate an 8-bit tga is slightly smaller in filesize than an 8-bit pcx. Heh. 

I suggested the compiled in custom palette mostly for the sake of neatness. I never have my maps as separate .pak files because I find it a bit unnecessary to devote a whole folder for 1 map. 

HYPOTHESIS
After my post in the Katedra thread I was curious what actually makes a good baseline ammo HP to monster HP ratio, so I cheesed some debug info into progs.dat and loaded all the stock maps with it, as well as an utterly unscientific list of random custom maps I had at the ready.

EXPERIMENT
My metric was to add up monster health, except for zombies, then add up all ammo HP, subtracting the number of zombies from the number of rockets before figuring them in.
- Multipliers: 25 per shell, 9 per nail, 30 per cell, 180 per rocket.
- Starting shells were not counted.
- Totals include all ammo in secrets (there's no way to programmatically filter them out).
- No rockets vs Shamblers math was done. We can assume players don't use rockets on shamblers anyway.
- Chthons, Shubs, and crucified zombies were excluded.
- Numbers for E1M7 are inaccurate anyway due to meat fireworks.
- All numbers are skill 2.

http://lunaran.com/quake_mapammo.txt

Obviously there's a lot of assumptions inherent in this, and how hectic any given map's combat is (hecticness? hecticity? hectitude? let's go with hectitude.) the hectitude of any given map's combat is going to affect how these numbers actually feel in practice, but this is a good learning exercise to establish some boundaries.

CONCLUSIONS
- the first map of an episode always has way more ammo than necessary. (if I had counted starting shells these numbers would be a little higher even.) e2m1 wins the award for most excessive ammo placement, with enough ammo to kill every monster in the map 7 times. however, we can consider these numbers flawed, because enforcers' cells are all dropped in the first map but can't be used until the player has a lightning gun.
- e1m3 wins for most rockets, with 101. (you can only carry 100.)
- ratios otherwise hover around the 2.0-2.5 mark. episode 1's are the highest, episode 4 the lowest, predictably, although the plethora of quads in e4 probably offsets that.
- even though players of one-off custom maps don't have the benefit of bringing all the previous map's ammo with them (because there was no previous map), the handful of custom maps I checked were still a little more scarce with ammo than the id maps.
- the ammo to monster ratio in Katedra is just under 3:1. 

CORRECTION
- e2m4 wins for most rockets with fucking 118 

smqe08d_lun.bsp 

Very interesting. 

for science! 

As fascinating to study these numbers are (and I admire this kind of academical approach to Quake's gameplay), ultimately they're not very useful as benchmarks, because of how much maps vary.

For instance, jam2_lunaran's ratio is 1.9, and could's is 2.7, which would suggest that the former is less "balanced" than the latter, and therefore harder. And in reality it most certainly isn't.

Perhaps this code could be expanded to calculate the maximum monster damage/health pickups ratio? 

If there are tarbabies, it needs more ammo.