plr_pipeline interview

YM

LVL100 YM
aa
Dec 5, 2007
7,135
6,056
We've heard back from valve with responces to our questions about pipeline. Fascinating stuff if you ask me, which I know you didn't

TF2Maps.net: The new HUD for plr_ is very clear, when, and why, was the decision made to swap to an entirely new HUD for this?
Dario Casali: We knew we were going to need a new HUD from the very beginning. Because there are two sets of tracks, players can’t just follow the tracks to find their goal. They must now look at the HUD to decide where their or the enemy’s cart is.

TF2Maps.net: Can you explain the decisions behind making the cart only roll backwards on slopes, even if unattended? Was having rollback decided on after the sloped structure to the third stage or at the same time?

Dario Casali: The cart rolls backwards in pl maps to act as a win condition for the red team. It gives them breathing space and provides positive feedback to them if they are doing well. In Payload Race, both teams have a win condition pushing their cart. Also, with the mirrored setup of the map, having the carts recede would inevitably lead to them being in the same place for most of the time – right in the middle! Stage 3 was conceived as freestyle space that had a constant battle between the two teams. At any one moment you may have to prevent the enemy moving their cart or help your team move yours, all in the same arena. The ramps were added not only to reduce the occurrence of steam roll endings, but also to add sense of accomplishment to the cart pushers who made it to the top of the ramp. Of course, they also give a considerable sense of satisfaction to cart defenders who manage to win back all that ground made if the cart hadn’t quite reached the top of the ramp.

TF2Maps.net: Why were two narrow bridges used? In theory, and to a certain extent in practice, it leaves the player pushing the cart crouched behind the cart desperately trying to avoid a fatal headshot from a sniper on the other side of the map.

Dario Casali: Control of the area or a large number of people pushing the cart are essential to winning that stage. If you have 4 or 5 people pushing the cart, you can afford to lose a couple to headshots. We found that it was more exciting to expose the cart paths so that they’re vulnerable from many angles and threats. That means that teams have plenty of defense options despite this area not being sentry friendly.

TF2Maps.net: The head start feature has been described as minimal and almost pointless by many, were there alternate choices for an advantage in latter rounds?

Dario Casali: Initially, the winner of round 1 or 2 got their cart 30 seconds before the other team did. After trying this for a while, it was clear that the team who was waiting for their cart never noticed when it was available to them. They were too caught up in the final battle. Instead of giving a timed headstart, it was much clearer to simply move the carts along the track according to who won the prior round. The advantage effect is nominal, and in the map update we may very well add points to each stage rather than just one point at the end. Stay tuned!

TF2Maps.net: Members often use vmex to decompile official maps to use as a learning reference but pipeline has some nodraw 'fins' alongside the track in stage three, one would assume they're an entity of some sort but vmex hasn't decompiled them properly, what are they there for?

Dario Casali: They are blocklight brushes – they make shadows under the ramps.

TF2Maps.net: What specific design challenges do you feel were most interesting, or were most challenging? Were there any a-ha! moments when a specific problem was unravelled? Knowing the answers, what question would you ask yourself?

Dario Casali: I went through tons of whiteboard designs before ending up with what we have today. Some of the design problems encountered were 1) a linear race away from your spawn means you have to run too far to get to the action 2) a linear race means that one team may pull away from the other and be impossible to catch 3) a race that occurs entirely in neutral territory (like stage 3) means traditional defense setups would be difficult to pull off 4) giving both teams the identical goal can lean to stalemates

Initially stage 1 and 2 were a single stage, and because the tracks wrap around on each other, the spawns would start as they do now, and when either team got the first cap, spawns would swap to the second stage, and the side route doors would open up. This was confusing to players and the size of the battlefield (stage 1+2) meant that it was hard to find the action. Also, doors that open at certain points in the game confused people. I tried simplifying the system a few times and still people were confused as to where they were supposed to be. Eventually I found that there were neat places to split the areas into two different rounds, so that solved the complexity problem. Of course, then there were two mostly separated areas with two convoluted paths between them to get to the other side. Initially your cart started in the heart of the enemy base (where it ends in round 1) and you had to push it to where round 2 ends now. People didn’t like being separated from their cart at the beginning of the round, so I punched a tunnel through the neutral space to get the carts starting on their respective sides.


TF2Maps.net: Pipeline has a very straightforward path if you run along the track, the upper areas, particularly in the second stage, however are quite convoluted and easy to get lost in the first several times playing, a stark contrast to goldrush and badwater where you always ended up somewhere you were expecting to. Do you agree or disagree with this?

Dario Casali: I’d agree with this. I didn’t want to exceed a 20 second demoman run time from spawn to the cart (if it was in enemy territory), so that kept the map fairly compact. In order to pack neutral areas and safe routes through that compact space, I have to convolute the paths somewhat. It’s not an ideal way to move through the space, but most playtesters learned the routes pretty quickly after being initially confused.

TF2Maps.net: Knowledge of pipeline came in part with the scout update, why was this when if we look back at the images it was clearly unfinished?

Dario Casali: That was a teaser image of what was to come :) We wanted to show the ramps and the two carts in the background even though the map was not finished by that point.

TF2Maps.net: Was the change in lighting between the scout update and final something decided after the first images were shown or had pipeline just not reached the lighting/art pass in its development until after that point?

Dario Casali: It had not been art passed by that point

TF2Maps.net: In what order to new assets for a new theme come, there's many new textures and props for all three maps in this update, how do you decide what you do and don't need?

Dario Casali: The new art assets are usually created during the art pass. Level designers usually give their best approximation of what we want, and the artists make the cool looking stuff to replace our placeholders.

TF2Maps.net: Are there specific places you and the artists working with you draw inspiration from?

Dario Casali: Inspiration comes from many different places. A road trip through with Utah and Nevada inspired Goldrush, and for Pipeline, we really wanted to do a night theme because we hadn’t done one yet.

TF2Maps.net: You've said before you're not familiar with paint overs, what we were referring to were shots from Hydro's early development that an artist had literally painted over to give the blocky geometry more of a finished map look. (Specifically this article http://features.cgsociety.org/story_custom.php?story_id=4338&page=3). If your artists don't do that for your maps, how do you work with them to achieve the same goals?

Dario Casali: I must have misinterpreted a question in the past. We do use paintovers. In the case of Pipeline, Dhabih painted over key areas for the transformation into nighttime and I built the brush geometry to support the paintovers. That geometry was then playtested, and finally converted to models during the art pass.

TF2Maps.net: The new factory props have a grey-black gradient for a texture, would a normal texture not have worked? Was it necessary to fake the dark lighting on them?

Dario Casali: I think they were built that way because they remain in darkness. We wanted to only light the gameplay space and mostly have the rest of the geometry silhouetted against the night sky

TF2Maps.net: To date pipeline is the only Valve map without an intro video, why is this?

Dario Casali: I’m not entirely sure about that, you’ll probably see one pop up sooner or later.

TF2Maps.net: How do you come up with names for maps, are you very careful not to pick anything been done in the community already?

Dario Casali: We do try to avoid naming maps the same as popular community maps, but with so many out there, that’s becoming harder and harder! The name of the map usually follows from the location it is evoking.

TF2Maps.net: Pipeline, like the first release of badwater basin, suffers from "oh, was that the explosion?? :(" syndrome, badwater was updated at a later date, will pipeline?

Dario Casali: That’s the plan!

TF2Maps.net: How many other mappers are working on TF2, do you work on other games at the same time or just TF2?

Dario Casali: I am personally working on L4D2 at the moment. There are five other mappers on Tf2 currently I believe.
 
Last edited by a moderator:

Rasputin

L3: Member
Sep 20, 2008
107
20
Yipee!

Pipeline, like the first release of badwater basin, suffers from "oh, was that the explosion?? " syndrome, badwater was updated at a later date, will pipeline?

That’s the plan!

Yipee!

The head start feature has been described as minimal and almost pointless by many, were there alternate choices for an advantage in latter rounds?

Initially, the winner of round 1 or 2 got their cart 30 seconds before the other team did. After trying this for a while, it was clear that the team who was waiting for their cart never noticed when it was available to them. They were too caught up in the final battle. Instead of giving a timed headstart, it was much clearer to simply move the carts along the track according to who won the prior round. The advantage effect is nominal, and in the map update we may very well add points to each stage rather than just one point at the end. Stay tuned!

Yipee!
 

JosephOhSnap

L1: Registered
Jan 6, 2009
36
2
These are always very interesting to read. No mention as to why nobody noticed the no collisioned barrels/tires.
 

Fearlezz

L10: Glamorous Member
May 4, 2008
787
476
5 other mappers aswell. Wow.
 

Psy

The Imp Queen
aa
Apr 9, 2008
1,706
1,491
Good read. :)

I might add some points to my own map but deciding where to place them on such a short track may be something of a challenge.
 

Tom Foolery

L1: Registered
Apr 13, 2009
48
4
Very interesting. :D
 

Ezekel

L11: Posh Member
Dec 16, 2008
818
245
aww, my questions weren't asked or answered :(
i really wanted to hear more about the time and processes involved for designing and for implementing a new game mode (after a number of gamemodes already exist), querying the setup time decisions and lack of stalemate conditions, oh and why getting a cart up the rollback hills in the 1st 2 rounds doesn't result in an instanct victory.

aww well, at least there was some interesting info in the response.