Here are the main points taken out
Quote
I don't have a transcript, but I do have a crappy summary that misses some stuff.
Best update? Probably Spy vs. Sniper due to how it all came together both in and out of game and because the Spy can be changed up a lot by items like the Dead Ringer compared to, say, the Heavy's narrow focus. The Spy has probably the widest scope of possible changes of any class.
Weapon stats are thought up and tested by the entire TF team - though one guy usually "owns" a weapon and drives it's development - it's hard to really point fingers.
Lots of feedback from e-mail, forums, stream games and playing the game themselves. They try and find common threads between all the conversations they find and reconciling them, it's not "who should we listen to more" due to biases and tradeoffs (see: the different opinions between a games forum talking about TF and a TF-dedicated forum). No bias towards the Steam forums either, they're only around 3% of the player base. They try to focus on problems, not solutions, until they're sure they've identified the problems correctly. THEN they're interested in what solutions the community proposes.
The original closed beta didn't die due to some sort of internal problems, it was simply "solving a different problem" from the current one, that problem being getting feedback from the hardcore comp players. Now that they get more regular feedback from that group and they had this whole Replay thing they'd been working on for a year, it made sense to open the beta to the public to make sure huge code changes didn't render the game unplayable for a couple of days.
They knew the whole Unusual Hat sales on eBay were going to happen, just not how much they'd go for. That's why they dragged their feet on adding trading for so long, because they were worried about having the backend, enough customer support to deal with this sort of thing. Whether he LIKES it or not is irrelevant, it'll happen regardless. If anything, they'd like to make it less risky, to stop scamming.
They're working on giving people something to do with Crates if they don't want to open them. They don't really want to force the issue because people are interested in different things with the store. They're doing data analysis to see whether crates are dropped enough so that they have trade value. They're trying to find out how they can give everyone something they're interested in (hence the newer crate series.)
Hats and artstyle... The original TF2 art design didn't have anything for character customization at all. Just adding that means changes to the design and art direction. So the art direction's different now because its trying to do different things from the original art direction. It would have been nice to have known at the time what they'd do later on so they could better prepare for it, but hindsight...
"Thousands" of community item submissions, narrowed down by a quality bar and what fits whatever design problems/ideas they're thinking of addressing at the time. "We're pretty bad at this part", need to provide better feedback to contributers on why some items are selected and some aren't. The primary problem is volume. They're working on some stuff to make things better in that regard and make things more transparent.
Promotional items... "An incredibly small amount of work for us" when it's just a hat/misc item - they usually work with raw source files from the collaborating game, so it only takes a day or two while working on all the other updates at the same time. They're happiest with the Shogun pack because they fit the artstyle so well and it had class-update levels of attention paid to weapon design. They want to do more promo updates like that, where you'd barely tell it was a promo and it's a true collaboration with the other developer and the community. Though, of course, they take more work since they ARE another update.
Promo items were also used early on to test just how far they could bend the artstyle before it just plain looks bad. They hit that point (the Dangeresque glasses being specifically pointed out) so they won't go there again.
"We're not ready to talk about the Portal 2 stuff yet."
Paints! Some of it is testing the edges, as said above. The swatch on the Contribute site was for the original content and didn't remotely allow for customization. Paints are a challenge because if there's not enough changes from the original skin, nobody would want to use them.
Clientside hat/paint disabling: "We've talked about it, but haven't come to any decision yet."
Class distinction: There's other ways to heal yourself, but nobody's really gonna argue that that means Medics are obsolete. "To be honest, any issues we have there today are probably ones we had when we shipped."
The future: They don't plan too far ahead, they perfer to react to customers. They have ideas, but those are just ideas!
More "big" updates: They just tend to build stuff, then talk later about how to package it. You can get bigger updates slightly slower. It's a tradeoff, though community assets can help with this. They've thought of doing a second round of Class Updates, there's all the Beta stuff, and there's constant community suggestions... there isn't a lack of stuff to do.
Community perception seems to be that an update with three or so Medic items would be less than a "Medic Update". It's something they need to think about.
Story-heavy updates: Maybe? The writing team's been busy on Portal 2. Most changes in update content are less due to philosophy changes as it is team members running off to help ship things on other teams.
Will there be a "final" update? "I have no idea what we'll be doing six years from now." Nothing's off the table, though. They certainly didn't think they'd still be working on this game! Customer opinions make the decisions.
More out of the game stuff like graphic novels, post-Meet The Team shorts, maybe? They've talked about it. The desire for a Raising The Bar-style art book is "absolutely there", but it's a whole lot of work and nobody's gone off to actually do it.
Console Orange Box Updates: "I don't even know how we'd pull that off." Optimising 3+ years of updates that didn't consider, say, strict memory budgets or controller support for the new menus is a huge amount of work that nobody wants to do. It's not IMPOSSIBLE, but there's just been more important things to do. Like the PC updates. PS3 Steamworks helps, but doesn't solve the aformentioned problems.
"The hard part isn't doing the work, it's figuring out, what's the right work to do?"
Edited by Drag#!, 22 March 2011 - 18:49.