The future of hoodoo (and discussion about filesize)

[What Is Schwa]

Why would people download it eventually if servers aren't putting it up because initially it will have the effect of emptying their server?

Youme is getting direct feedback from server admins saying pretty much what I'm saying. It seems amazing that a bunch of people with no experience running a server all act like the ones running the servers are some sort of creativity squashing overlords.

Admins want great maps just like we do. Further, admins know better than us what makes or breaks a server. When one comes to a mapper and offers direct feedback we should be thanking them. Instead the tone is usually "omg they are stupid I can download that map in 30 seconds lol!"

 

[Chilly]

To me, this statement is nonsense. How will a player know a map is good unless he plays it? Word of mouth? I don't believe anything anyone tells me anymore, especially after hearing how awesome mariokart and orange_x are. I think most players are skeptical of all custom maps.

Players don't know if a map is good unless they PLAY it. Getting it into the hands of players requires servers to run the map. If server mods aren't happy that they lose half their player population each time the map runs the map won't get played.

Server admins are the gatekeepers of what maps get played. Either make the map how you want to make it, or concede to their whims to get it played. It is a crappy situation but that is just how it goes.

You really don't think people know in advance? Word tends to spread about maps in different communities which is why a handful of large maps have had good success. If a map is good, there's a good chance it will get replayed on servers. If a player sees a map starting to download once, he might not download it. However, if he sees a map starting to download on 10 different occasions he'll be more likely to wait for the download. Additionally, when map votes come up people often ask about them. A good map will get good word-of-mouth promotion and people will be more likely to download it. Finally, you state that admins are the gatekeepers, which is definitely true. Being a server admin I can tell you that many of us look at screenshots of maps first and then test the map out on a server.

I, being a server admin of a server that can be full on any given night, am a gatekeeper. Many admins will end up looking at screenshots of maps before loading it onto our servers. Some will test the maps in advance as well. Good maps, as a result, will get more play time and will get more downloads.

 

[MrAlBobo]

Admins want great maps just like we do. Further, admins know better than us what makes or breaks a server. When one comes to a mapper and offers direct feedback we should be thanking them. Instead the tone is usually "omg they are stupid I can download that map in 30 seconds lol!"

This /\

Anyways...my opinion...
im probably much closer to being a gamer then a mapper, and i have a server, so im kinda seeing all the sides.

In my opinion the file size should be efficient, not small, but efficient. If you can cut out useless things then do it, do it to just about anything you can, but do not cut details.
The tf2 custom community is basically in its infancy, there are far more people playing only valve maps, then there are people playing any customs period. And mostly its due to that one bad experience where the 1st custom they play is just terrible. So there needs to be more maps out there that can reach hoodoos quality, while at the same time being conscience of the filesize. Being that my server is rarely full I tend to catch alot of people new to the game that will download for a long period of time simply because they think thats what you have to do. Then once in, they usually will like the map and look for it on occasion. The new users are easy to sway towards custom maps, its the people that have played for awhile and know that most customs just aren't good that tend to leave the server while waiting for a really long download. Its for these people that the filesize needs to be lower. Like personally ill leave a really long download if I don't know what the map is.

...kinda trailed off there...anyways...
but yeah, like grazr said, do what you can to clean it up
if the quality warrants the filesize then fine

and the average user, based on the people ive asked will download that much if someone recommends the map

 

[Altaco]

Eh, I'd say just offer a .bsp.bz2 along with it for server owners, bsps compress amazingly well.

 

[III_Demon]

the argument in my head goes something like this:

are you trying to make a standard TF2 map, or do something different?

if you're trying to make a normal (possibly recognized by valve) TF2 map, then you must adhere to the conventions established by valve. no ladders, no branching cart paths, etc.. and no files significantly bigger than any of the other maps.

if you're trying to do something different, then you can deviate, and do whatever the hell you want, file size be damned.

it seems to me like you're trying to make a normal TF2 map, with hoodoo. there are no wacky gimmicks, the look and feel is similar and fits within the established TF2 themes nicely. therefore, you are restricted by the established valve standards, and i'd say you want to keep the file size down.

look at it as part of the art. its easy to get caught up in the 'art' of the thing, and feel like limits like polygon counts and file sizes are holding you back, but thats part of the point. the idea is to make it look as good as possible within the limits. ignoring the limits is a bit like cheating, in this case.

i think hoodoo would still be an excellent map at half the filesize.

 

[MrAlBobo]

Eh, I'd say just offer a .bsp.bz2 along with it for server owners, bsps compress amazingly well.

anyone that does not know how to bz2 zip does not deserve to own a server
and the extra time it would take the person uploading it and the people downloading it, is not worth the shorter time it would take to just bz2 it themselves

now only uploading the bz2 might be a good idea, but most server ops would not know just what filetypes its zipped in by simply looking at the file and would just grab the map and zip it themselves

 

[Chilly]

I'm curious how much snipergen's custom cart model adds to the size. I absolutely love the model, but I've not seen info on how much space it takes.

 

[Nineaxis]

There's also the ambulance, which I'd have to say doesn't look amazing and is probably adding quite a bit of size.

 

[UKCS-Alias]

I voted for go soft. Especialy since the server i play on most doesnt want maps larger than ~50MB (unzipped). The problem is mainly that when a new map comes not many are in the server (which can have a massive negative impact on the fun the map gives the first time). I had it once that a 30 mb (unzipped) map in css resulted in a 10 vs 10 for over 10 minutes just because over 30 people were downloading it (cousing it to slow down alot). Normaly people download the map in less than 2 minutes but i find peak times important as they can increase the time massively as what happened with my map before.

FYI, the server owner he talked about is probably the one i play on. Note that in css we still allow maps that are <30mb zipped time to time if the quality is worth it. Just try to reduce it as much as you can. Quality is more important than the filesize after all. if you realy cant get below the 70mb without losing some quality just keep it that way.

Still, my map was ment for 60+ player in css and <30 made it a bit unbalanced. Im sure its not that bad in tf2 as many times a 5vs5 already plays well.

 

[Uriak]

The custom cart is around 1.5 Mo compressed (if I can trust the dowload section)

For instance the double cart model I use if 4 times heavier, because it features more high rez textures. So the custom cart shouldn't be discarded, imho, it's the most visible custom element and doesn't count much toward the total.

 

[Vander]

I think 2 maps with similar names is a bad idea. I say finish up Hoodoo how you intended to originally and then make your next map with better optimization and reduced filesize in mind.

 

[Laz]

try setting your resolution lower before you build cubemaps
since I run my game windowed, my map's size got halved. so perhaps resolution while building cubemaps realyl does matter? dunno.

 

[Apom]

Cubemaps are built at 512x512 resolution no matter what. That's why they only take the upper left corner of your screen when you build them.

 

[YM]

Cubemaps are built at 512x512 resolution no matter what. That's why they only take the upper left corner of your screen when you build them.

Yeah but you can downsize the resolution it saves them to, anything from 16x16 to 128x128 iirc. I don't think reducing the resolution down from the standard 32x32 is worth it, but certainly cutting a few out and spreading the remaining around might help out.

 

[Laz]

then how come my map decreases in size by half when I compile it windowed then when I do it fullscreen? :s Im confused :s

 

[Apom]

Yeah but you can downsize the resolution it saves them to, anything from 16x16 to 128x128 iirc. I don't think reducing the resolution down from the standard 32x32 is worth it, but certainly cutting a few out and spreading the remaining around might help out.

I knew you'd know Cubemaps aren't what makes a map huge, though. One cubemap essentially weights resolution² × sides × color depth, that's like 25kb each with default parameters (uncompressed, and ignoring the marginal non-image information).

then how come my map decreases in size by half when I compile it windowed then when I do it fullscreen? :s Im confused :s

If cubemaps are half of your map size, there is a serious problem somewhere. That or your map is an empty cube.

 

[Randdalf]

I'd stay stubborn, mainly because this map is a prime candidate for Valve selection (although the next update will be in 2009 according to Doug Lombardi...), and there is no need to squander the file size making the map as small as possible, possibly losing some quality, as that many reduces chances of being chosen. Plus, file size will be irrelevant if it is chosen.

 

[YM]

(although the next update will be in 2009 according to Doug Lombardi...)

D: links before i cry!

 

[UKCS-Alias]

http://www.shacknews.com/onearticle.x/55364

 

[Apom]

Well, they essentially say that because the guys at Valve are working on L4D. Making a "community maps update" is still quite possible within these boundaries, as it doesn't involve much work on Valve's side. At least I think so.