Lady Vashj down

[nobbie]

... but the highend encounter is (again) very buggy: http://www.nihilum.co.uk/news/41/

[ima_nerd]

... but the highend encounter is (again) very buggy: http://www.nihilum.co.uk/news/41/

Quite the nice dagger.

[nobbie]

The comment of raid leader 'Awake' on this obviously very bugged encounter is now up as well.

EDIT: Removed copy & paste commentary as Awake decided to move his posting into a personal blog.

[Quark]

The comment of raid leader 'Awake' on this obviously very bugged encounter is now up as well:

Valid complaints aside, my opinion of Awake is about as low as it can get.

[nobbie]

A rant is a rant, and I can clearly understand these guys.

Btw, they have removed Awake's commentary and will put it up in a personal blog later.

[Artega]

Quite the nice dagger.

I'm dismayed that the final versions of the Blacksmithing weapons are still superior to something out of Serpentshrine Cavern.

[Tuftears]

Why shouldn't they be? You get the materials to craft them out of Serpentshrine... I'd expect crafting to yield a result superior to loot from the instance that was needed to get the materials.

[Concillian]

I'm dismayed that the final versions of the Blacksmithing weapons are still superior to something out of Serpentshrine Cavern.

err, don't the Vortices only drop in Serpentshrine Cavern (and beyond). Doesn't seem that out of place to me. I mean it takes a while to collect 15 vortices if you assume that most bosses probably only drop 1-2 and that has to be split among 25 people. You're probably looking at 15-25 SSC boss kills to get one T3 upgraded weapon in your entire guild.

[Brista]

err, don't the Vortices only drop in Serpentshrine Cavern (and beyond). Doesn't seem that out of place to me. I mean it takes a while to collect 15 vortices if you assume that most bosses probably only drop 1-2 and that has to be split among 25 people. You're probably looking at 15-25 SSC boss kills to get one T3 upgraded weapon in your entire guild.

I'm dismayed that non-raider crafting still isn't competitive. The current choices are 1) raid 2) raid but be a master crafter with it and 3) arena but only for 6-7 equipment slots. As more raid instances come out and further mudflate the gear we will be back to pre-TBC where groups of players from raid guilds absolutely destroy everyone in pvp. I'm not convinced that non-raiding arena farming is anywhere close to viable, I think rather if you want to win you have to do both

As for Nihilum there's an interesting interview with them here:
http://www.nihilum.co.uk/wcradio/nihiinterviewhq.mp3

It's a long interview but when asked about the bugginess they stated that they feel they are beta testing the game for Blizzard and that some encounters have been quite literally impossible. Interestingly they don't realise why. The reason I'm sure is that Blizzard QA simply don't have 25 people as good at playing WoW as these guilds so even with a free selection of gear and consumables they can't test bleeding edge adequately

[NiteFox]

The reason I'm sure is that Blizzard QA simply don't have 25 people as good at playing WoW as these guilds so even with a free selection of gear and consumables they can't test bleeding edge adequately

Apparently, Blizzard does have a dedicated raid test team; which does have the added benefits of unlimited access to consumables and near-0 lag. Apparently, by the time any dedicated raiding guild has trounced an area, they've already done it several times.

This, I gather, is Nihilum's point: Blizzard claims that they have a raid test group, but if they did then they really should have picked up on the problems with the encounter. The raid boss respawning near-instantly isn't a little hiccup in the encounter that needs some tuning. That sort of thing should have been hammered out before any live player got there, unless you really want to believe that Blizzard's raid team was lucky enough not to encounter this sort of thing in all the time they've allegedely been "testing" Lady Vashj.

[Quark]

That sort of thing should have been hammered out before any live player got there, unless you really want to believe that Blizzard's raid team was lucky enough not to encounter this sort of thing in all the time they've allegedely been "testing" Lady Vashj.

Something as simple as Blizzard's raid teams class balance can really mess with their testing. Take Shade of Aran, for instance. His death script is really buggy. But what's he commonly doing when he doesn't die properly? Casting Polymorph, casting Arcane Explosion, casting Blizzard (actually casting it, not just having it in the room).

The amount of time he's casting a Blizzard during the fight is very, very small. Chances of death there are pretty low for one QA department to pickup. Dying right after the Poly call goes out has an incredibly low chance of happening -> and effectively zero depending on what your group's DPS is (many groups beat it no sweat, many groups force a poly before he calls water elementals). Finally, casting Arcane Explosion does have a long time ... but DPS is effectively nil during this. This has the highest chance of happening if you're full of classes with DoTs.

I could easily see 4 groups constantly testing Aran and not once experiencing this. By the same token, having the entire playerbase inside Karazhan and the issues become readily apparent.

My point being, at some time we have to stop blaming the QA team for bugs that are really hard to catch and start blaming the raid team for having them in the first place.

[LochnarITB]

My point being, at some time we have to stop blaming the QA team for bugs that are really hard to catch and start blaming the raid team for having them in the first place.

However, if you are going let QA off the hook, you have to do so for the raid team also, for the very same reasons. If you want complex encounters, you have complex code, too complex for anyone to think of all the interactions that can occur to give you bugs. If bugs, be they outright breaks or just "it shouldn't do that", could always be caught in the coding, there would be no need for QA. While it would be ideal to not be paying beta testers, it is just never going to be the case, unless you purposely hang back several months until the other subscribers can do their thing.

[Zarathustra]

I could easily see 4 groups constantly testing Aran and not once experiencing this. By the same token, having the entire playerbase inside Karazhan and the issues become readily apparent.

My point being, at some time we have to stop blaming the QA team for bugs that are really hard to catch and start blaming the raid team for having them in the first place.

Valid point. In response, how do we address the fact that KNOWN bugs that are caught in beta tend to persist for over a year?

[nobbie]

If you want complex encounters, you have complex code, too complex for anyone to think of all the interactions that can occur to give you bugs.

Agreed, but how comes that i.e. Vashj respawns shortly after her apparent death and kills the loot master? I mean she was dead, and dead is dead, right?

[LochnarITB]

I mean she was dead, and dead is dead, right?

Apparently you don't watch a lot of horror movies, right?:P

But, yeah, that is the kind of thing I am talking about. In a complex encounter, there are a large number of conditions, "if this then that". Each of those conditions is like a digit in a binary number. The more conditions, the more the digits and the larger the number. One of those possible numbers indicates a heal event that brings the HP back to positive, bringing life to the dead.

Making this point, however, only gives the teams an out until the bug has been observed. Once it has been seen, then there is a basis for hunting down the interactions that lead to that condition. Not fixing it at that point is where they can be faulted. There is no excuse for having bugs that have existed forever. Of course, fixing one bug changes conditions that interact in other places as well and it is to be expected that more bugs will appear from the new combinations, possibly even harder to anticipate because the new code is interacting with old code that is not even being looked at.

In the past, I have earned a living as a programmer, but only in a small department. Even with smaller pieces of code, I have learned that there are two types of programmers. There are those that grind out large volumes of code that works but contains bugs. Then there are those that are good at determining the cause of, and fixing, the bugs that are found in testing. I tended to be the fixer and I can't even imagine how much there is to dig through to discover the cause of some of the things that can happen in something like WoW. I also know that the glory tends to go to the guys that pour out the code and fixing their code tends to be fairly thankless. Some praise should go to the bug teams for the things that do get fixed. Blizzard is obviously not dedicating the resources (man hours) needed to do all the bug fixing. Either they are being tight fisted with the dollars to pay them or it is hard to fill the desired number of positions, but there apparently are not enough people working on the bugs to add up to the number of man years that are needed.

[Monkey]

Valid point. In response, how do we address the fact that KNOWN bugs that are caught in beta tend to persist for over a year?

Well, not all bugs are created equally. I'm sure there are some known bugs which:
1. Are difficult to replicate (especially lag-related bugs)
2. Are difficult to fix (again, especially lag-related bugs)

A known bug isn't an easy-to-replicate or easy-to-fix bug. Meanwhile (back at the ranch) there's constant pressure on Blizzard to triage their efforts and get something substantial out to their paying subscribers. Not to mention that while some bugs (Consortium gem payment) are annoying, other bugs (leaving the 'bubble' before an EotS match) are flat-out gamebreaking.

[Taelas]

Valid point. In response, how do we address the fact that KNOWN bugs that are caught in beta tend to persist for over a year?

The greatest reason for not fixing bugs is simply that they are not serious enough.

And most bugs are fixed.

[Malakar]

Not only is their project pretty large scale, a lot of their code was written a long time ago. It's quite possible that even the original writer of some piece of code has forgotten some caveat of his own code, which has a non-existant, incorrect, or out of date comment.

I've never worked in a large scale dev team; I can only imagine how much it must complicate things. Big projects can be hard enough on your own, when you yourself did every section of code.

Well, not all bugs are created equally. I'm sure there are some known bugs which:
1. Are difficult to replicate (especially lag-related bugs)
2. Are difficult to fix (again, especially lag-related bugs)

QFT


Bugs are forgiveable, the huge amount of trash mobs in TBC is not.