The future of hoodoo (and discussion about filesize)

  • If you're asking a question make sure to set the thread type to be a question!

What should I do?


  • Total voters
    69

YM

LVL100 YM
aa
Dec 5, 2007
7,135
6,056
There have been lots of discussions about the size of maps and a lot of people seem to suggest that anything bigger than a valve sized map (pl_goldrush is 74MB for reference) is too big.
Since pl_hoodoo is now at 103MB it qualifies as too large, until now I've just been very stubborn about the filesize but since a server owner has taken the time to email me about it I thought I should actually re-evaluate my position on the matter.

So I've come up with a few ideas of how to continue:
  • Stay stubborn and let filesize increase and simply hope valve want it in an update which of course will negate the worry of filesize
  • Go soft and try my hardest to reduce the filesize as much as possible by stripping out custom textures, removing some cubemaps and lowering lightmap resolutions
  • A mix of the two, create two maps, pl_hoodoo_*version* and pl_hoodoo_reduced. The first will be the unadulterated file with high filesize and the latter will have no custom content at all, lower res lightmaps and fewer cubemaps. The cut down version is released at _b3 level and not updated whilst the other version is tweaked and modified in the usual cycle

Discuss.
 

Icarus

aa
Sep 10, 2008
2,245
1,210
I'd much rather have a map that was more accessible to people than one with extra eyecandy.

If VALVe wants your map you could always just re-add the custom stuff, like how VALVe got Steel before the final was even out.

Having two versions out is rather confusing.
 
Last edited:

Uriak

L8: Fancy Shmancy Member
Apr 27, 2008
543
70
I'd say "stay stubborn". There are things that please me and some that don't about hoodoo, but, I'm really on the "don't bother with size" side. Of course, you should check for lightmaps as much as you can, and you will always be questionned about this, since, the cart aside, the custom content is not really obvious for untrained eyes.

You're on a stage where release are not very frequents, and you should rather try to make it as "classical" as you can. People will ultimately download it once and for all.

I've been myself investigating about map size, and I'm afraid I may hit the 100 Mo cap with snaketemple with beta release. I think the best answer is making the map more desirable than lowering expectations and "cost" for players.
 

Apom

L6: Sharp Member
Sep 14, 2008
366
65
I think you should stay stubborn. A map that's twice as good deserves twice the download time, doesn't it? And I believe yours fits that description compared to most custom maps around.

Anyways, people who won't stand an extra 30 seconds of download sound like spoilt children to me. I know that, as a mapper, you want your map to be played by as many people as possible - but certainly , the people who're going to be the most appreciative of your work aren't the ones complaining about map size.
 

CakesOnAPlane

L1: Registered
Aug 7, 2008
15
0
Yea I would say also stay stubborn.

You only need to download the map once, then you can enjoy all the goodness, and personally I would prefer more goodness with a longer download time, which I only need to endure once.
 

YM

LVL100 YM
aa
Dec 5, 2007
7,135
6,056
Anyways, people who won't stand an extra 30 seconds of download sound like spoilt children to me. I know that, as a mapper, you want your map to be played by as many people as possible - but certainly , the people who're going to be the most appreciative of your work aren't the ones complaining about map size.

Yeah well this particular server owner claims that the average user his country would have to wait 15 minutes even from a fast download redirect, which really is too much (I'm not entirely sure his figures are accurate since I downloaded _b3 from an american server at an abismal 200kB/sec and it only took 4 minutes)

Thats why I like the idea of the reduced version, it would take a day (including compile time) to make and publish and I'd not be updating it at all so if you want to play the most polished version you wait for longer but if you cant wait and want to play hoodoo you can play the reduced version, which instead of 100 uncompressed would be more like 60-70MB since the current version is 57MB without cubemaps or any custom content.
 

Dan2025

L2: Junior Member
Aug 29, 2008
67
4
I don't really know how servers work. Seriously. So here's a question I have, are custom maps downloaded every time their played or is it a onetime only thing--Do you download it once and then have it for life?

I mean, if something is really good (and it's my understanding that your map is really good) then even two hoursis an acceptable amount of time if you never have to go through it again.

So no. Keep stubborn.
 

Arhurt

L6: Sharp Member
May 7, 2008
315
140
Problem is that most servers woul "creep" to the featureless version and thus, it become standard.

I`d say keep it large and pretty, on the long run better looking and more polished maps are going to do YOU more good than a popular and featureless map.
 

laghlagh

L6: Sharp Member
Jul 15, 2008
389
53
I don't really know how servers work. Seriously. So here's a question I have, are custom maps downloaded every time their played or is it a onetime only thing--Do you download it once and then have it for life?

I mean, if something is really good (and it's my understanding that your map is really good) then even two hoursis an acceptable amount of time if you never have to go through it again.

So no. Keep stubborn.
One download, then never again. and it's not 2 hours lol, more like 3 minutes.
 

TheDarkerSideofYourShadow

L10: Glamorous Member
Apr 12, 2008
792
286
I would keep the version you've got, ie stay stubborn. Mapping is an art, and its not as if you're being limited by the boundaries of the medium, but rather by the price individuals are willing to pay to enjoy that art. I see how that can be a problem, since of course you want as many people to enjoy it as possible, but you shouldn't be willing to sacrifice a masterpiece to get a few more people to own a copy.

Make the best damn map you can. Put it out there. Some people may not like it because of its size, but thats okay. Hell, the better the map is, the more likely valve would make it official anyways.
 
Oct 6, 2008
1,947
445
I think you should make the map whatever size you like and screw the others - that's what broadband is for - LOL
 

Nineaxis

Quack Doctor
aa
May 19, 2008
1,767
2,820
I'd say be stubborn but do your best to reduce it. Don't remove custom content just because, but do all the fancy lightmap optimisation, etc.
 

What Is Schwa

L6: Sharp Member
Jan 13, 2008
375
445
People are throwing out all sorts of numbers regarding download time. Is it quick? Is it fast? Who cares how long it takes to download from here or FPSBanana, that doesn't matter. The majority of players DO NOT GET THEIR MAPS THIS WAY.

Players get maps from servers they join. Therefor, download time depends entirely on the server and the environment at the time.

For example, if 75% of the people on your server don't have hoodoo at mapchange you have 14-16 players all hitting one source at one time, which DOES slow the download significantly.

Hoodoo is a server killer under these circumstances. When players see the little progress bar not budging for 30+ seconds they disconnect.

----

Youme, make the map you want to make.

However, if you want to make a map that other players play on, you're going to have to concede elements of design to the audience. This isn't a creative challenge unique to you, every artist in his medium must decide if his vision is more important than its reception. Sometimes people are lucky and the vision is exactly what the audience wants. Unfortunately this is not the case for you.

The mere fact that you query a poll from your audience suggests you are interested in its reception. If so, change it to suit the people who are going to play it. If not, than quit asking what we think and make what you want.
 

Chilly

L6: Sharp Member
May 3, 2008
326
127
Preface: I'm a server admin with a fast redirect setup and we playtest MANY custom maps that come out.

I'll list out my points...
1. A decent fast download server (like mine) can easily deliver a map fast enough to max out my Internet connection (I top out around 700kb/s. This means that I downloaded pl_hoodoo_b3 last night in under 2 minutes. If 20 people are all downloading the map at once, it slows down the redirect, but it gradually speeds up as each person finishes their download. This means that it might double the amount of time it takes each person to download the map.

2. Some players play on servers where they don't get good connections to the fastdl server. A good example is one of our clan members that lives in Japan. It takes him about 15 minutes to download hoodoo. He's smart, though... when we start loading a new map he goes and grabs it off fpsbanana.com to get full download speeds.

3. Server admins should take steps to make downloads easier on their players. I'm setting up a mailing list for ours, plus I have this webpage for people to check daily: http://yinyarr.com/Server_Maps.aspx.

3. Map updates can be a pain in the butt to server admins. Frequent updates of even great maps absolutely suck if the map is large. This discourages players from downloading the map since they just downloaded a previous version a week ago. A good rule of thumb: as the map size gets larger, increase the amount of time between each release. I think youme did this perfectly, with more frequent updates during the alpha releases and more time/changes during each beta release.

4. Players will take the time to download good maps. They will not take time to download poor maps.

5. For reasons 3 and 4, mappers should make as many improvements as possible between each release, while limiting their total number of releases as much as possible. If you're unable to do that, consider hooking up with a clan that beta tests a lot (like mine) or with an admin of a very popular server (like FLOOR_MASTER) to do closed testing of your map. With a closed test you don't have to worry about download speeds because the players can download the map in advance. Additionally, you limit the number of versions floating around and make huge advances between each public release.

6. People will take time to download a final whereas they might not take that time with a beta.

7. Multiple versions of the same map will confuse people. They'll wonder which version is more recent. An example is the turbo edition of pl_snaketemple. A couple players asked me what the _te was at the end and said it'd be confusing jumping between versions on different servers. Making an analogy to the NFL... if you have two quarterbacks and you're not sure which one should be the starter, it generally means that neither one is very good. I'm not saying hoodoo isn't good (it's my favorite custom by far)... I'm saying that if you do another version it'll dilute the "specialness" of the original.

My opinion of hoodoo is that it seems close to being done. If you're expecting many more changes, try doing them all in one release and possibly do a closed test of that release (contact me if you want to set this up, otherwise FLOOR_MASTER might be a good one to try). Once you think it's ready you can do a public release. If you allow extended closed testing, you might be able to take it final which will get it downloaded a lot more.

When we did beta testing for Furyo's cp_studio he made some comments about file size. Basically his opinion was that he wanted to release a map that he was 100% proud of... something that people would look at and see his abilities regardless of the file size. I don't think your goal should be to get the map played as much as possible... it should be to make the map as good as possible. If you do that, then gradually more and more people will try it and it'll get played more.
 
Last edited:

YM

LVL100 YM
aa
Dec 5, 2007
7,135
6,056
The mere fact that you query a poll from your audience suggests you are interested in its reception. If so, change it to suit the people who are going to play it. If not, than quit asking what we think and make what you want.

I'm simply conducting research into what people actually think, I'm trying to think a way I can basicly ask the same question in the valve tf2 forum without having a mod move it to the map section that no one looks at. I want to ask the same question there because those guys are gamers more than we are, almost all of us are tainted by the views of mapping whilst most of them haven't the foggiest about mapping.

I'm amazed so few think two is a good idea, since with the reduced version I can maximise airtime for hoodoo and then hopefully if valve ever want to make it official all I have to do is give them the proper version and then everyone gets forced to download that. The watered down version wasn't a way of allowing impatient gamers to play, its an underhand tactic to increase my chances of getting it officialised (which is of course, the ultimate goal)

Still, results are very interesting to see, I'm sure even others are benefiting from this.
 

Apom

L6: Sharp Member
Sep 14, 2008
366
65
Yeah well this particular server owner claims that the average user his country would have to wait 15 minutes even from a fast download redirect, which really is too much (I'm not entirely sure his figures are accurate since I downloaded _b3 from an american server at an abismal 200kB/sec and it only took 4 minutes)
This actually highlits the uselessness (?) of optimization. If the "quality" version takes 15 minutes to download, the "basic" version would only be half its size, hence still 7 minutes. I mean, I don't know what country that is, but if people are going to make some tea while the download happens, then reading newspapers over tea will not make much of difference.


For example, if 75% of the people on your server don't have hoodoo at mapchange you have 14-16 players all hitting one source at one time, which DOES slow the download significantly.
That argument is only valid when a new version was just released. With dilution happening fairly fast, I wouldn't take it into consideration.
 

What Is Schwa

L6: Sharp Member
Jan 13, 2008
375
445
I wasn't trying to be an ass Youme, just giving you my honest opinion. I like hoodoo and think it is a good map. I've lobbied for it on my server, but it hasn't received the same love.

Again, you said it took only four minutes on a connection that was most likely serving just you. Consider what would happen if 20 other players all hit the link at the exact same time as you. That is where the bottleneck happens and where you lose players.

Chilly said:
4. Players will take the time to download good maps. They will not take time to download poor maps.
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.
 

grazr

Old Man Mutant Ninja Turtle
aa
Mar 4, 2008
5,441
3,814
Again, you said it took only four minutes on a connection that was most likely serving just you. Consider what would happen if 20 other players all hit the link at the exact same time as you. That is where the bottleneck happens and where you lose players.

People will download it eventually. They'll loose a few players here and there but as Apom said; this effect gets diluted over time.

That said. You should try your best to optimise the file size whilst not significantly impacting the obvious details. It's the custom content that adds those extra MB's to your map like the pounds you gain from a hamburger.. with cheese, bacon, egg, fried onions and all the rest.

Reduce your cubemaps, for the most part it's only the snipers scope that uses them, and the water.. (which you have none?). You don't have to be really sensitive with them (But you know this much, it's just a matter of implementing it). Also, as you said, increase lightmap scales. One of the easiest things to get away with. Secondly, this is probably the hardest one to implement, reduce displacement scales. Sometimes the floor is a little too rigid in maps, so it'd probably be a good thing in more respects than file size optimisation for hoodoo.

I havn't decompiled your map. I don't know how well it's hint optimised.. but reducing leaf nodes is a good thing for file size also. good hint optimisation will reduce file size. Less maths to store.

Anything.. that doesn't detract from it's beauty.

If that doesn't work. Keep it the way it is. Anyone who's into online gaming and downloading that takes a day to download this.. shouldn't be on the internet at all.