Comments on: It Really IS Like An Expansion Launch!

By: Alarik

Alarik — Mon, 11 Dec 2006 17:59:33 +0000

"As much as the beta server is supposed to find all issues in a game before ship, it never does. The path that Blizzard has taken has increased the odds that most data issues with the game, such as class imbalances and whatnot, will be fixed before the expansion goes live in a month."

In this case it wasn't even really an issue of finding the bugs. The worst of the bugs that went live have been known for weeks, and fixed on the beta servers for nearly as long.

I can understand not being able to keep the public test server/live patch up to date with the most current beta, but there are bugs that were literally fixed on the beta server days after the PTR went live, and as far as anyone I've heard can tell, the version of the patch that went live is the same version that was on the PTR when it first opened weeks ago; nothing at all was fixed except for possibly server issues from the public testing period.

By: Damion Schubert

Damion Schubert — Mon, 11 Dec 2006 16:43:12 +0000

As much as the beta server is supposed to find all issues in a game before ship, it never does. The path that Blizzard has taken has increased the odds that most data issues with the game, such as class imbalances and whatnot, will be fixed before the expansion goes live in a month.

What Blizzard has managed to do is quite clever, actually. Whether they meant to do it is a matter of some debate, but the net result is that they probably aren’t going to lose any of their hardcore subscribers, at least until they launch the expansion, while getting the patch out early increases the odds that most of the data issues are found and fixed before BC launches, and most flaws will be restricted to content issues – which is a pretty good place to be.

By: Knurd

Knurd — Mon, 11 Dec 2006 13:53:53 +0000

Lietgardis: Cheers. I do agree with what you’re saying. I guess I’m looking at it from a production angle that would say, “get the easy crap out of the way early, so we don’t have to deal with it later.” Bliz has already seemed to indicate that providing a foundation patch for the expansion is going to crop up some wierd inconsistencies for game play, but larger considerations might make that necessary. At least for the few weeks that they have planned their real release. I guess if you want to call that ‘collateral damage’, I wouldn’t argue that point.

By: Lietgardis

Lietgardis — Mon, 11 Dec 2006 00:28:20 +0000

Knurd: generally, you want to implement the stuff that’s going to need the most testing first. I liked to implement the new classes first, because all those new spells take a while to test. You also want them ready for day one of beta, because you need as much time as you can get to iterate on balance.

I would argue that treasure data needs to go in early too, because the data entry is prone to error and the testing is time-consuming — literally, the QA guys are spawning hundreds of mobs and killing them to see what they drop. Is this new thing spawning 1% or 2% of the time? If it’s on something that spawns a lot, little differences really add up, and it’s not always going to be apparent in internal testing (another reason to have it ready to go on day one of beta). Treasure table fuckups can really hurt the economy, and economy fuckups are hard to recover from without embarassing rollbacks.

That’s an argument to not mess with treasure data that you’re confident in, and not to bother branching bits of it even if the data layout makes it possible. They’re high-risk changes.

When I commented earlier, I wasn’t actually thinking about the unique nature of treasure data. I was just remembering times when I got frustrated that I had to patch stuff that shouldn’t have been patched yet because of data design problems. On second thought, yeah, I have to say that I’d rather not touch treasure data that’s passed testing, even if early release makes me look bad.

By: Knurd

Knurd — Sun, 10 Dec 2006 13:05:10 +0000

So, in your folks estimation, is it better to have loot table data adjusted in a pre-patch or in a release canidate? Is it better to put the man-hours in earlier for that, or later?

By: Andrew Crystall

Andrew Crystall — Fri, 08 Dec 2006 23:32:42 +0000

Liet – AutoAssualt’s *2 hour* application times for 100MB patches come to mind. No, really.

By: Rand 'al Thor

Rand 'al Thor — Fri, 08 Dec 2006 22:20:10 +0000

"Assuming that it was a known issue: if the data isn\'e2\'80\'99t very well compartmentalized, you can\'e2\'80\'99t really pick and choose what you\'e2\'80\'99re going to patch if you\'e2\'80\'99ve already changed a lot.

In other words, if all the treasure data is in one big file, and there\'e2\'80\'99s some stuff in there you want to patch and some stuff you don\'e2\'80\'99t, it\'e2\'80\'99s a pretty painstaking process to go through and clean it up, especially if you have crap source control and/or crap data design."

Thanks for the response to that question, which I figured if it was an accidental side effect it was going to basically come down to the complexity of the bits of code working otgether.

Since I asked that question though it's been stated that it was an intentional change which ,as has become apparent, has baffled alot of people as to why they would intentionally make the change so far before the expansion launch.

The post from Tseric that I quoted above doesn't really give a reason and basically in a nutshell tells everybody to "suck it up your opinions don't matter, because the game isn't about you (customers)".

By: Lietgardis

Lietgardis — Fri, 08 Dec 2006 18:41:02 +0000

"Maybe one of you design/dev types could explain the reasoning behind letting loot no one can use for a month or so start dropping."

Assuming that it was a known issue: if the data isn't very well compartmentalized, you can't really pick and choose what you're going to patch if you've already changed a lot.

In other words, if all the treasure data is in one big file, and there's some stuff in there you want to patch and some stuff you don't, it's a pretty painstaking process to go through and clean it up, especially if you have crap source control and/or crap data design.

By: Andrew Crystall

Andrew Crystall — Fri, 08 Dec 2006 16:06:37 +0000

Jeff, to be fair with WoW it’s pretty easy to see what rises to the top because, well, lots of people use it. More people than PLAY most MMO’s, for the most successful.

By: sinij

sinij — Fri, 08 Dec 2006 15:57:19 +0000

I know most game designers come from IT field and tend to think (or not think) in terms of code stability, maintainability and similar equally useless things when it comes to UI design.

When you design UI you should concern yourself with following things:

1) Does interface does what players need to do at all times with at most 2 clicks or 2 key presses?
2) Does interface is easy to understand and easy to master?
3) Is it consistent at all times - i.e. does what it suppose to do at all times in the same way
4) Is there too much clicking involved, what are most click-intensive tasks that should be automated?
5) What is longest click/button press chain for single action? How can I shorten it?