Apologizing to ZOS after being educated from ESO LIVE 8/12

tinythinker

JD2013 wrote: »

Never raged. I see no point in raging on a forum.

However, I did find that segment particularly interesting. I know very little about how bug fixes work, so it was nice to see some insight into the process.

One thing that I will probably say trips ZOS up a lot is how differently a test server works compared to the live server. There's lots of variables that they can't or don't anticipate. For example, I can imagine that it is hard to recreate what happens on a live server in PVP, given that the server and indeed the client has to calculate so many different numbers at any given time.

And then, when they add thousands of lines of new code in with each DLC, that can create some havoc.

Is it an excuse? No. But sometimes things just can't be calculated no matter how much testing goes on. Could this game do with some shine and polishing? Sure could in some areas. But I can only imagine the stress that goes on in their offices.

That's one thing I've asked about and still don't get/never got an answer about.

I understand coding with so many interactive blocks of code can be tricky. I understand testing is hard. I understand bug chasing/squashing is hard. I understand that we each experience the game from different perspectives, have different encounters with performance, and want our own priorities moved to the top of the list. I also get that they say that they can only do so much on the PTS and also that buying more server space for live isn't a silver bullet. I'll grant each point.

Fine. OK. Sure.

But why *not* spend money increasing PTS server space to get it closer (not close, not the same, but closer) to conditions on live? I can only speculate that the bug/performance concerns aren't hitting the bottom line hard enough to justify the expense, but it could potentially save so much time and money spent going back to find and fix things after they hit live.

Lysette

tinythinker wrote: »

JD2013 wrote: »

Never raged. I see no point in raging on a forum.

However, I did find that segment particularly interesting. I know very little about how bug fixes work, so it was nice to see some insight into the process.

One thing that I will probably say trips ZOS up a lot is how differently a test server works compared to the live server. There's lots of variables that they can't or don't anticipate. For example, I can imagine that it is hard to recreate what happens on a live server in PVP, given that the server and indeed the client has to calculate so many different numbers at any given time.

And then, when they add thousands of lines of new code in with each DLC, that can create some havoc.

Is it an excuse? No. But sometimes things just can't be calculated no matter how much testing goes on. Could this game do with some shine and polishing? Sure could in some areas. But I can only imagine the stress that goes on in their offices.

That's one thing I've asked about and still don't get/never got an answer about.

I understand coding with so many interactive blocks of code can be tricky. I understand testing is hard. I understand bug chasing/squashing is hard. I understand that we each experience the game from different perspectives, have different encounters with performance, and want our own priorities moved to the top of the list. I also get that they say that they can only do so much on the PTS and also that buying more server space for live isn't a silver bullet. I'll grant each point.

Fine. OK. Sure.

But why *not* spend money increasing PTS server space to get it closer (not close, not the same, but closer) to conditions on live? I can only speculate that the bug/performance concerns aren't hitting the bottom line hard enough to justify the expense, but it could potentially save so much time and money spent going back to find and fix things after they hit live.

A live server like this is about 1.5-2 metric tons of hardware - this is not just a few blades in a rack. And it costs quite some amount of money as well. And even if they would use a copy of the live server, they do not have the load which is required to make sure it all will work, when the system comes under production load.

ArchMikem

I don't apologize, because I don't need to. I never once raged or even just complained about Zenimax and their game. Sure I was (and still am a bit) upset over Crashinium, but I never raged at ZOS for that, I sent in a Support Ticket when my character got trapped in a loop and then avoided the Zone as much as possible in hopes of a future fix. Other than that all the other things wrong have been so minute that I would have been a total *** if I decided to rage about them.

I understand completely that trying to make something like this work, and continue to work is a massive effort. I'm not gonna give them crap for there being some problems.

Thealteregoroman

yodased wrote: »

Rylana wrote: »

Almost three years, some bugs from beta still exist

Keep making all the excuses you want

ZOS meeting 101.

Dev: man we got this punch list of bugs.
Boss: how much revenue will we lose if we don't fix it?
Finance: Maybe 2%
Boss: How many people bought the new dlc?
Sales: 750k
Boss: make a new dlc
Dev: but the bugs, I can fix them given a team and enough time.
Boss: you are fired, to junior dev: Make a new dlc.

It's also why they don't fix pvp, if everyone stopped pvp obviously the cost benefit says that's ok.

OMG!!!

tinythinker

Lysette wrote: »

tinythinker wrote: »

JD2013 wrote: »

Never raged. I see no point in raging on a forum.

However, I did find that segment particularly interesting. I know very little about how bug fixes work, so it was nice to see some insight into the process.

One thing that I will probably say trips ZOS up a lot is how differently a test server works compared to the live server. There's lots of variables that they can't or don't anticipate. For example, I can imagine that it is hard to recreate what happens on a live server in PVP, given that the server and indeed the client has to calculate so many different numbers at any given time.

And then, when they add thousands of lines of new code in with each DLC, that can create some havoc.

Is it an excuse? No. But sometimes things just can't be calculated no matter how much testing goes on. Could this game do with some shine and polishing? Sure could in some areas. But I can only imagine the stress that goes on in their offices.

That's one thing I've asked about and still don't get/never got an answer about.

I understand coding with so many interactive blocks of code can be tricky. I understand testing is hard. I understand bug chasing/squashing is hard. I understand that we each experience the game from different perspectives, have different encounters with performance, and want our own priorities moved to the top of the list. I also get that they say that they can only do so much on the PTS and also that buying more server space for live isn't a silver bullet. I'll grant each point.

Fine. OK. Sure.

But why *not* spend money increasing PTS server space to get it closer (not close, not the same, but closer) to conditions on live? I can only speculate that the bug/performance concerns aren't hitting the bottom line hard enough to justify the expense, but it could potentially save so much time and money spent going back to find and fix things after they hit live.

A live server like this is about 1.5-2 metric tons of hardware - this is not just a few blades in a rack. And it costs quite some amount of money as well. And even if they would use a copy of the live server, they do not have the load which is required to make sure it all will work, when the system comes under production load.

Again, like I said, not close to, closer. I don't expect them to replicate live or something approximate to it. I have no doubt the PTS will always be tiny by comparison. But it can still be expanded. In any case, like I said, I'm sure they've done the calculations and know what is enough to get "good enough" out, so that will have to do.

Roechacca

I'm still waiting for my apology for a year plus of paid beta.

JD2013

tinythinker wrote: »

JD2013 wrote: »

Never raged. I see no point in raging on a forum.

However, I did find that segment particularly interesting. I know very little about how bug fixes work, so it was nice to see some insight into the process.

One thing that I will probably say trips ZOS up a lot is how differently a test server works compared to the live server. There's lots of variables that they can't or don't anticipate. For example, I can imagine that it is hard to recreate what happens on a live server in PVP, given that the server and indeed the client has to calculate so many different numbers at any given time.

And then, when they add thousands of lines of new code in with each DLC, that can create some havoc.

Is it an excuse? No. But sometimes things just can't be calculated no matter how much testing goes on. Could this game do with some shine and polishing? Sure could in some areas. But I can only imagine the stress that goes on in their offices.

That's one thing I've asked about and still don't get/never got an answer about.

I understand coding with so many interactive blocks of code can be tricky. I understand testing is hard. I understand bug chasing/squashing is hard. I understand that we each experience the game from different perspectives, have different encounters with performance, and want our own priorities moved to the top of the list. I also get that they say that they can only do so much on the PTS and also that buying more server space for live isn't a silver bullet. I'll grant each point.

Fine. OK. Sure.

But why *not* spend money increasing PTS server space to get it closer (not close, not the same, but closer) to conditions on live? I can only speculate that the bug/performance concerns aren't hitting the bottom line hard enough to justify the expense, but it could potentially save so much time and money spent going back to find and fix things after they hit live.

See, in essence I agree with both you and @Lysette both more maths calculating (which is I presume part of their jobs anyway) and extra server space/upgrading server space could help. Absolutely.

However, they still can't replicate it completely due to how many players are in each campaign at any given time, as I presume that there aren't enough staff in their offices, and bots, again I'm presuming, act differently compared to players. Add into that different passives that players have in their Champion Points, and all of those variables (there's a reason non CP pvp runs smoother.)

Now, I presume calculating things for PVE is easier than it is for PVP (again, all presumption and theoreticals) and there certainly could be more testing etc. This I agree with. Never said ZOS get any kind of free pass. But there will sadly always be variables in writing and implementing so much code.

tinythinker

Roechacca wrote: »

I'm still waiting for my apology for a year plus of paid beta.

*type*type*type*type*type*type*

TOS says no...



[reference for those who missed the joke in spoilers]

https://www.youtube.com/watch?v=WOdjCb4LwQY

mb10

yodased wrote: »

Rylana wrote: »

Almost three years, some bugs from beta still exist

Keep making all the excuses you want

ZOS meeting 101.

Dev: man we got this punch list of bugs.
Boss: how much revenue will we lose if we don't fix it?
Finance: Maybe 2%
Boss: How many people bought the new dlc?
Sales: 750k
Boss: make a new dlc
Dev: but the bugs, I can fix them given a team and enough time.
Boss: you are fired, to junior dev: Make a new dlc.

It's also why they don't fix pvp, if everyone stopped pvp obviously the cost benefit says that's ok.

So today I've seen an accountant for ZOS on the forums elsewhere and now a ZOS board member who attends their meetings and discussions.

Funny, these apparent ZOS could be you or me!

Lysette

JD2013 wrote: »

tinythinker wrote: »

JD2013 wrote: »

Never raged. I see no point in raging on a forum.

However, I did find that segment particularly interesting. I know very little about how bug fixes work, so it was nice to see some insight into the process.

One thing that I will probably say trips ZOS up a lot is how differently a test server works compared to the live server. There's lots of variables that they can't or don't anticipate. For example, I can imagine that it is hard to recreate what happens on a live server in PVP, given that the server and indeed the client has to calculate so many different numbers at any given time.

And then, when they add thousands of lines of new code in with each DLC, that can create some havoc.

Is it an excuse? No. But sometimes things just can't be calculated no matter how much testing goes on. Could this game do with some shine and polishing? Sure could in some areas. But I can only imagine the stress that goes on in their offices.

That's one thing I've asked about and still don't get/never got an answer about.

I understand coding with so many interactive blocks of code can be tricky. I understand testing is hard. I understand bug chasing/squashing is hard. I understand that we each experience the game from different perspectives, have different encounters with performance, and want our own priorities moved to the top of the list. I also get that they say that they can only do so much on the PTS and also that buying more server space for live isn't a silver bullet. I'll grant each point.

Fine. OK. Sure.

But why *not* spend money increasing PTS server space to get it closer (not close, not the same, but closer) to conditions on live? I can only speculate that the bug/performance concerns aren't hitting the bottom line hard enough to justify the expense, but it could potentially save so much time and money spent going back to find and fix things after they hit live.

See, in essence I agree with both you and @Lysette both more maths calculating (which is I presume part of their jobs anyway) and extra server space/upgrading server space could help. Absolutely.

However, they still can't replicate it completely due to how many players are in each campaign at any given time, as I presume that there aren't enough staff in their offices, and bots, again I'm presuming, act differently compared to players. Add into that different passives that players have in their Champion Points, and all of those variables (there's a reason non CP pvp runs smoother.)

Now, I presume calculating things for PVE is easier than it is for PVP (again, all presumption and theoreticals) and there certainly could be more testing etc. This I agree with. Never said ZOS get any kind of free pass. But there will sadly always be variables in writing and implementing so much code.

With a bit of knowledge in graph theory one can see, that this will not work in this way - it does not even require testing, it is math which tells, no, this will not work out and it will lead to lag - it is that simple.

tinythinker

mb10 wrote: »

yodased wrote: »

Rylana wrote: »

Almost three years, some bugs from beta still exist

Keep making all the excuses you want

ZOS meeting 101.

Dev: man we got this punch list of bugs.
Boss: how much revenue will we lose if we don't fix it?
Finance: Maybe 2%
Boss: How many people bought the new dlc?
Sales: 750k
Boss: make a new dlc
Dev: but the bugs, I can fix them given a team and enough time.
Boss: you are fired, to junior dev: Make a new dlc.

It's also why they don't fix pvp, if everyone stopped pvp obviously the cost benefit says that's ok.

So today I've seen an accountant for ZOS on the forums elsewhere and now a ZOS board member who attends their meetings and discussions.

Funny, these apparent ZOS could be you or me!

Don't blow my cover, man, or Rich will get so mad.

(No really.)

ArchMikem

Roechacca wrote: »

I'm still waiting for my apology for a year plus of paid beta.

That's funny. You're a funny person. Owing you an apology for giving them money.

JD2013

Lysette wrote: »

JD2013 wrote: »

tinythinker wrote: »

JD2013 wrote: »

Never raged. I see no point in raging on a forum.

However, I did find that segment particularly interesting. I know very little about how bug fixes work, so it was nice to see some insight into the process.

One thing that I will probably say trips ZOS up a lot is how differently a test server works compared to the live server. There's lots of variables that they can't or don't anticipate. For example, I can imagine that it is hard to recreate what happens on a live server in PVP, given that the server and indeed the client has to calculate so many different numbers at any given time.

And then, when they add thousands of lines of new code in with each DLC, that can create some havoc.

Is it an excuse? No. But sometimes things just can't be calculated no matter how much testing goes on. Could this game do with some shine and polishing? Sure could in some areas. But I can only imagine the stress that goes on in their offices.

That's one thing I've asked about and still don't get/never got an answer about.

I understand coding with so many interactive blocks of code can be tricky. I understand testing is hard. I understand bug chasing/squashing is hard. I understand that we each experience the game from different perspectives, have different encounters with performance, and want our own priorities moved to the top of the list. I also get that they say that they can only do so much on the PTS and also that buying more server space for live isn't a silver bullet. I'll grant each point.

Fine. OK. Sure.

But why *not* spend money increasing PTS server space to get it closer (not close, not the same, but closer) to conditions on live? I can only speculate that the bug/performance concerns aren't hitting the bottom line hard enough to justify the expense, but it could potentially save so much time and money spent going back to find and fix things after they hit live.

See, in essence I agree with both you and @Lysette both more maths calculating (which is I presume part of their jobs anyway) and extra server space/upgrading server space could help. Absolutely.

However, they still can't replicate it completely due to how many players are in each campaign at any given time, as I presume that there aren't enough staff in their offices, and bots, again I'm presuming, act differently compared to players. Add into that different passives that players have in their Champion Points, and all of those variables (there's a reason non CP pvp runs smoother.)

Now, I presume calculating things for PVE is easier than it is for PVP (again, all presumption and theoreticals) and there certainly could be more testing etc. This I agree with. Never said ZOS get any kind of free pass. But there will sadly always be variables in writing and implementing so much code.

With a bit of knowledge in graph theory one can see, that this will not work in this way - it does not even require testing, it is math which tells, no, this will not work out and it will lead to lag - it is that simple.

Then, playing devils advocate here, if it is that simple, why haven't ZOS done that?

Elsonso

Rohamad_Ali wrote: »

But why *not* spend money increasing PTS server space to get it closer (not close, not the same, but closer) to conditions on live? I can only speculate that the bug/performance concerns aren't hitting the bottom line hard enough to justify the expense, but it could potentially save so much time and money spent going back to find and fix things after they hit live.

It has more to do with the number of people on the server than the size of the server.

If you don't have the players, the server space is just wasted.

TequilaFire

Lysette wrote: »

tinythinker wrote: »

JD2013 wrote: »

Never raged. I see no point in raging on a forum.

However, I did find that segment particularly interesting. I know very little about how bug fixes work, so it was nice to see some insight into the process.

One thing that I will probably say trips ZOS up a lot is how differently a test server works compared to the live server. There's lots of variables that they can't or don't anticipate. For example, I can imagine that it is hard to recreate what happens on a live server in PVP, given that the server and indeed the client has to calculate so many different numbers at any given time.

And then, when they add thousands of lines of new code in with each DLC, that can create some havoc.

Is it an excuse? No. But sometimes things just can't be calculated no matter how much testing goes on. Could this game do with some shine and polishing? Sure could in some areas. But I can only imagine the stress that goes on in their offices.

That's one thing I've asked about and still don't get/never got an answer about.

I understand coding with so many interactive blocks of code can be tricky. I understand testing is hard. I understand bug chasing/squashing is hard. I understand that we each experience the game from different perspectives, have different encounters with performance, and want our own priorities moved to the top of the list. I also get that they say that they can only do so much on the PTS and also that buying more server space for live isn't a silver bullet. I'll grant each point.

Fine. OK. Sure.

But why *not* spend money increasing PTS server space to get it closer (not close, not the same, but closer) to conditions on live? I can only speculate that the bug/performance concerns aren't hitting the bottom line hard enough to justify the expense, but it could potentially save so much time and money spent going back to find and fix things after they hit live.

A live server like this is about 1.5-2 metric tons of hardware - this is not just a few blades in a rack. And it costs quite some amount of money as well. And even if they would use a copy of the live server, they do not have the load which is required to make sure it all will work, when the system comes under production load.

1.5 to 2 metric tons?
I see the problem now they are still running IBM 360s. lmao
/joking

Lysette

JD2013 wrote: »

Lysette wrote: »

JD2013 wrote: »

tinythinker wrote: »

JD2013 wrote: »

Never raged. I see no point in raging on a forum.

However, I did find that segment particularly interesting. I know very little about how bug fixes work, so it was nice to see some insight into the process.

One thing that I will probably say trips ZOS up a lot is how differently a test server works compared to the live server. There's lots of variables that they can't or don't anticipate. For example, I can imagine that it is hard to recreate what happens on a live server in PVP, given that the server and indeed the client has to calculate so many different numbers at any given time.

And then, when they add thousands of lines of new code in with each DLC, that can create some havoc.

Is it an excuse? No. But sometimes things just can't be calculated no matter how much testing goes on. Could this game do with some shine and polishing? Sure could in some areas. But I can only imagine the stress that goes on in their offices.

That's one thing I've asked about and still don't get/never got an answer about.

I understand coding with so many interactive blocks of code can be tricky. I understand testing is hard. I understand bug chasing/squashing is hard. I understand that we each experience the game from different perspectives, have different encounters with performance, and want our own priorities moved to the top of the list. I also get that they say that they can only do so much on the PTS and also that buying more server space for live isn't a silver bullet. I'll grant each point.

Fine. OK. Sure.

But why *not* spend money increasing PTS server space to get it closer (not close, not the same, but closer) to conditions on live? I can only speculate that the bug/performance concerns aren't hitting the bottom line hard enough to justify the expense, but it could potentially save so much time and money spent going back to find and fix things after they hit live.

See, in essence I agree with both you and @Lysette both more maths calculating (which is I presume part of their jobs anyway) and extra server space/upgrading server space could help. Absolutely.

However, they still can't replicate it completely due to how many players are in each campaign at any given time, as I presume that there aren't enough staff in their offices, and bots, again I'm presuming, act differently compared to players. Add into that different passives that players have in their Champion Points, and all of those variables (there's a reason non CP pvp runs smoother.)

Now, I presume calculating things for PVE is easier than it is for PVP (again, all presumption and theoreticals) and there certainly could be more testing etc. This I agree with. Never said ZOS get any kind of free pass. But there will sadly always be variables in writing and implementing so much code.

With a bit of knowledge in graph theory one can see, that this will not work in this way - it does not even require testing, it is math which tells, no, this will not work out and it will lead to lag - it is that simple.

Then, playing devils advocate here, if it is that simple, why haven't ZOS done that?

They would have to touch the combat system mechanics - and if you listened to what they said in the live show, they are afraid to touch such a system - not to talk about that they would have to convince Mr. Wrobel, and he loves AoE because they are cool. With people like that it is hard to fix anything.

There would be a quite simple solution though, and I have suggested in many times - reverse scale the efficiency of spells and attacks with the effort required to resolve the interaction in a zerg - this would make it highly undesireable to form a zerg and things would get better. Meaning the more effort is required to resolve the interaction graph, the less effective spells and attacks coming from the zerg will get. Simple solution, zergs would be highly undesirable, and the lag would be significantly less. Of course this is based on that they have actual interaction graphs and do not just compute it in a brute force way. It has to be computed separately for any interaction group which is large enough to be a zerg.

Mojmir

Lol

tinythinker

lordrichter wrote: »

Rohamad_Ali wrote: »

But why *not* spend money increasing PTS server space to get it closer (not close, not the same, but closer) to conditions on live? I can only speculate that the bug/performance concerns aren't hitting the bottom line hard enough to justify the expense, but it could potentially save so much time and money spent going back to find and fix things after they hit live.

It has more to do with the number of people on the server than the size of the server.

If you don't have the players, the server space is just wasted.

Yeah, that's why I used to make calls for out there testing ideas, like having a week on live of some major PvP changes just to see if any of it worked and gets lots of helpful data even from the stuff that didn='t. Honestly regular PvPers would probably love it. But, that means taking the severs down for everyone twice in one week, so it's a non-starter. And the campaigns have to run with the same basic architecture. Only minor variables can be changed, so no "test campaign" on live. Basically, the conclusion I came to a while ago is that we're pretty much stuck with the testing/bug fix system that already exists for the entire game. The only thing I can see at the moment is hiring more coders and Q&A people in terms of serious changes to how this whole process works.

Roechacca

tinythinker wrote: »

Roechacca wrote: »

I'm still waiting for my apology for a year plus of paid beta.

*type*type*type*type*type*type*

TOS says no...



[reference for those who missed the joke in spoilers]

https://www.youtube.com/watch?v=WOdjCb4LwQY

Hahaha Exactly a blanket response I'd expect from them too.

The punchline is I'm happy I don't sit around waiting for anything to get fixed. Game markets booming right now with solid games well worth the price tag. Just looking at these forums once in a while is a great reminder of what a famous employee here said. I don't have to be here! Best advise for anyone confused and complaining. Loving the break.

JD2013

Lysette wrote: »

JD2013 wrote: »

Lysette wrote: »

JD2013 wrote: »

tinythinker wrote: »

JD2013 wrote: »

Never raged. I see no point in raging on a forum.

However, I did find that segment particularly interesting. I know very little about how bug fixes work, so it was nice to see some insight into the process.

One thing that I will probably say trips ZOS up a lot is how differently a test server works compared to the live server. There's lots of variables that they can't or don't anticipate. For example, I can imagine that it is hard to recreate what happens on a live server in PVP, given that the server and indeed the client has to calculate so many different numbers at any given time.

And then, when they add thousands of lines of new code in with each DLC, that can create some havoc.

Is it an excuse? No. But sometimes things just can't be calculated no matter how much testing goes on. Could this game do with some shine and polishing? Sure could in some areas. But I can only imagine the stress that goes on in their offices.

That's one thing I've asked about and still don't get/never got an answer about.

I understand coding with so many interactive blocks of code can be tricky. I understand testing is hard. I understand bug chasing/squashing is hard. I understand that we each experience the game from different perspectives, have different encounters with performance, and want our own priorities moved to the top of the list. I also get that they say that they can only do so much on the PTS and also that buying more server space for live isn't a silver bullet. I'll grant each point.

Fine. OK. Sure.

But why *not* spend money increasing PTS server space to get it closer (not close, not the same, but closer) to conditions on live? I can only speculate that the bug/performance concerns aren't hitting the bottom line hard enough to justify the expense, but it could potentially save so much time and money spent going back to find and fix things after they hit live.

See, in essence I agree with both you and @Lysette both more maths calculating (which is I presume part of their jobs anyway) and extra server space/upgrading server space could help. Absolutely.

However, they still can't replicate it completely due to how many players are in each campaign at any given time, as I presume that there aren't enough staff in their offices, and bots, again I'm presuming, act differently compared to players. Add into that different passives that players have in their Champion Points, and all of those variables (there's a reason non CP pvp runs smoother.)

Now, I presume calculating things for PVE is easier than it is for PVP (again, all presumption and theoreticals) and there certainly could be more testing etc. This I agree with. Never said ZOS get any kind of free pass. But there will sadly always be variables in writing and implementing so much code.

With a bit of knowledge in graph theory one can see, that this will not work in this way - it does not even require testing, it is math which tells, no, this will not work out and it will lead to lag - it is that simple.

Then, playing devils advocate here, if it is that simple, why haven't ZOS done that?

They would have to touch the combat system mechanics - and if you listened to what they said in the live show, they are afraid to touch such a system - not to talk about that they would have to convince Mr. Wrobel, and he loves AoE because they are cool. With people like that it is hard to fix anything.

See, if they really would have to dig that deep into the engine and mechanics, then maybe they do need to look at that. (FYI I did listen to what was said, however it's quite late at night after a long day of work when I get to watch ESO Live, so admittedly all the info doesn't always compute )

However, such an overhaul may cause even more rage than already exists around here. They could certainly look at changing some of the abilities in different ways.

Lysette

TequilaFire wrote: »

Lysette wrote: »

tinythinker wrote: »

JD2013 wrote: »

Never raged. I see no point in raging on a forum.

However, I did find that segment particularly interesting. I know very little about how bug fixes work, so it was nice to see some insight into the process.

One thing that I will probably say trips ZOS up a lot is how differently a test server works compared to the live server. There's lots of variables that they can't or don't anticipate. For example, I can imagine that it is hard to recreate what happens on a live server in PVP, given that the server and indeed the client has to calculate so many different numbers at any given time.

And then, when they add thousands of lines of new code in with each DLC, that can create some havoc.

Is it an excuse? No. But sometimes things just can't be calculated no matter how much testing goes on. Could this game do with some shine and polishing? Sure could in some areas. But I can only imagine the stress that goes on in their offices.

That's one thing I've asked about and still don't get/never got an answer about.

I understand coding with so many interactive blocks of code can be tricky. I understand testing is hard. I understand bug chasing/squashing is hard. I understand that we each experience the game from different perspectives, have different encounters with performance, and want our own priorities moved to the top of the list. I also get that they say that they can only do so much on the PTS and also that buying more server space for live isn't a silver bullet. I'll grant each point.

Fine. OK. Sure.

But why *not* spend money increasing PTS server space to get it closer (not close, not the same, but closer) to conditions on live? I can only speculate that the bug/performance concerns aren't hitting the bottom line hard enough to justify the expense, but it could potentially save so much time and money spent going back to find and fix things after they hit live.

A live server like this is about 1.5-2 metric tons of hardware - this is not just a few blades in a rack. And it costs quite some amount of money as well. And even if they would use a copy of the live server, they do not have the load which is required to make sure it all will work, when the system comes under production load.

1.5 to 2 metric tons?
I see the problem now they are still running IBM 360s. lmao
/joking

This is really what something like this is - CCP has just ordered the hardware for the new TQ live server - which has to deal with about a similar amount of concurrency - and they said, the new hardware was 1.5 metric tons in weight.

tinythinker

Lysette wrote: »

There would be a quite simple solution though, and I have suggested in many times - reverse scale the efficiency of spells and attacks with the effort required to resolve the interaction in a zerg - this would make it highly undesireable to form a zerg and things would get better. Meaning the more effort is required to resolve the interaction graph, the less effective spells and attacks coming from the zerg will get. Simple solution, zergs would be highly undesirable, and the lag would be significantly less. Of course this is based on that they have actual interaction graphs and do not just compute it in a brute force way. It has to be computed separately for any interaction group which is large enough to be a zerg.

So basically if you cause lag your abilities do diddly?

Lysette

JD2013 wrote: »

Lysette wrote: »

JD2013 wrote: »

Lysette wrote: »

JD2013 wrote: »

tinythinker wrote: »

JD2013 wrote: »

Never raged. I see no point in raging on a forum.

However, I did find that segment particularly interesting. I know very little about how bug fixes work, so it was nice to see some insight into the process.

One thing that I will probably say trips ZOS up a lot is how differently a test server works compared to the live server. There's lots of variables that they can't or don't anticipate. For example, I can imagine that it is hard to recreate what happens on a live server in PVP, given that the server and indeed the client has to calculate so many different numbers at any given time.

And then, when they add thousands of lines of new code in with each DLC, that can create some havoc.

Is it an excuse? No. But sometimes things just can't be calculated no matter how much testing goes on. Could this game do with some shine and polishing? Sure could in some areas. But I can only imagine the stress that goes on in their offices.

That's one thing I've asked about and still don't get/never got an answer about.

I understand coding with so many interactive blocks of code can be tricky. I understand testing is hard. I understand bug chasing/squashing is hard. I understand that we each experience the game from different perspectives, have different encounters with performance, and want our own priorities moved to the top of the list. I also get that they say that they can only do so much on the PTS and also that buying more server space for live isn't a silver bullet. I'll grant each point.

Fine. OK. Sure.

But why *not* spend money increasing PTS server space to get it closer (not close, not the same, but closer) to conditions on live? I can only speculate that the bug/performance concerns aren't hitting the bottom line hard enough to justify the expense, but it could potentially save so much time and money spent going back to find and fix things after they hit live.

See, in essence I agree with both you and @Lysette both more maths calculating (which is I presume part of their jobs anyway) and extra server space/upgrading server space could help. Absolutely.

However, they still can't replicate it completely due to how many players are in each campaign at any given time, as I presume that there aren't enough staff in their offices, and bots, again I'm presuming, act differently compared to players. Add into that different passives that players have in their Champion Points, and all of those variables (there's a reason non CP pvp runs smoother.)

Now, I presume calculating things for PVE is easier than it is for PVP (again, all presumption and theoreticals) and there certainly could be more testing etc. This I agree with. Never said ZOS get any kind of free pass. But there will sadly always be variables in writing and implementing so much code.

With a bit of knowledge in graph theory one can see, that this will not work in this way - it does not even require testing, it is math which tells, no, this will not work out and it will lead to lag - it is that simple.

Then, playing devils advocate here, if it is that simple, why haven't ZOS done that?

They would have to touch the combat system mechanics - and if you listened to what they said in the live show, they are afraid to touch such a system - not to talk about that they would have to convince Mr. Wrobel, and he loves AoE because they are cool. With people like that it is hard to fix anything.

See, if they really would have to dig that deep into the engine and mechanics, then maybe they do need to look at that. (FYI I did listen to what was said, however it's quite late at night after a long day of work when I get to watch ESO Live, so admittedly all the info doesn't always compute )

However, such an overhaul may cause even more rage than already exists around here. They could certainly look at changing some of the abilities in different ways.

I edited my post and gave a solution, like I would implement it in their place. Have a look.

Lysette

tinythinker wrote: »

Lysette wrote: »

There would be a quite simple solution though, and I have suggested in many times - reverse scale the efficiency of spells and attacks with the effort required to resolve the interaction in a zerg - this would make it highly undesireable to form a zerg and things would get better. Meaning the more effort is required to resolve the interaction graph, the less effective spells and attacks coming from the zerg will get. Simple solution, zergs would be highly undesirable, and the lag would be significantly less. Of course this is based on that they have actual interaction graphs and do not just compute it in a brute force way. It has to be computed separately for any interaction group which is large enough to be a zerg.

So basically if you cause lag your abilities do diddly?

Yes, this would solve the problem - the role play reasoning would be - if people are too close together, they interfere with each other and are less effective.

It is not a problem to compute all this without lag in a lot of smaller groups, but zergs with AoE spam are a nightmare and create lag.

JD2013

Lysette wrote: »

JD2013 wrote: »

Lysette wrote: »

JD2013 wrote: »

Lysette wrote: »

JD2013 wrote: »

tinythinker wrote: »

JD2013 wrote: »

Never raged. I see no point in raging on a forum.

However, I did find that segment particularly interesting. I know very little about how bug fixes work, so it was nice to see some insight into the process.

One thing that I will probably say trips ZOS up a lot is how differently a test server works compared to the live server. There's lots of variables that they can't or don't anticipate. For example, I can imagine that it is hard to recreate what happens on a live server in PVP, given that the server and indeed the client has to calculate so many different numbers at any given time.

And then, when they add thousands of lines of new code in with each DLC, that can create some havoc.

Is it an excuse? No. But sometimes things just can't be calculated no matter how much testing goes on. Could this game do with some shine and polishing? Sure could in some areas. But I can only imagine the stress that goes on in their offices.

That's one thing I've asked about and still don't get/never got an answer about.

I understand coding with so many interactive blocks of code can be tricky. I understand testing is hard. I understand bug chasing/squashing is hard. I understand that we each experience the game from different perspectives, have different encounters with performance, and want our own priorities moved to the top of the list. I also get that they say that they can only do so much on the PTS and also that buying more server space for live isn't a silver bullet. I'll grant each point.

Fine. OK. Sure.

But why *not* spend money increasing PTS server space to get it closer (not close, not the same, but closer) to conditions on live? I can only speculate that the bug/performance concerns aren't hitting the bottom line hard enough to justify the expense, but it could potentially save so much time and money spent going back to find and fix things after they hit live.

See, in essence I agree with both you and @Lysette both more maths calculating (which is I presume part of their jobs anyway) and extra server space/upgrading server space could help. Absolutely.

However, they still can't replicate it completely due to how many players are in each campaign at any given time, as I presume that there aren't enough staff in their offices, and bots, again I'm presuming, act differently compared to players. Add into that different passives that players have in their Champion Points, and all of those variables (there's a reason non CP pvp runs smoother.)

Now, I presume calculating things for PVE is easier than it is for PVP (again, all presumption and theoreticals) and there certainly could be more testing etc. This I agree with. Never said ZOS get any kind of free pass. But there will sadly always be variables in writing and implementing so much code.

With a bit of knowledge in graph theory one can see, that this will not work in this way - it does not even require testing, it is math which tells, no, this will not work out and it will lead to lag - it is that simple.

Then, playing devils advocate here, if it is that simple, why haven't ZOS done that?

They would have to touch the combat system mechanics - and if you listened to what they said in the live show, they are afraid to touch such a system - not to talk about that they would have to convince Mr. Wrobel, and he loves AoE because they are cool. With people like that it is hard to fix anything.

See, if they really would have to dig that deep into the engine and mechanics, then maybe they do need to look at that. (FYI I did listen to what was said, however it's quite late at night after a long day of work when I get to watch ESO Live, so admittedly all the info doesn't always compute )

However, such an overhaul may cause even more rage than already exists around here. They could certainly look at changing some of the abilities in different ways.

I edited my post and gave a solution, like I would implement it in their place. Have a look.

I'm certainly all for making large Zergs undesirable.

tinythinker

Lysette wrote: »

tinythinker wrote: »

Lysette wrote: »

There would be a quite simple solution though, and I have suggested in many times - reverse scale the efficiency of spells and attacks with the effort required to resolve the interaction in a zerg - this would make it highly undesireable to form a zerg and things would get better. Meaning the more effort is required to resolve the interaction graph, the less effective spells and attacks coming from the zerg will get. Simple solution, zergs would be highly undesirable, and the lag would be significantly less. Of course this is based on that they have actual interaction graphs and do not just compute it in a brute force way. It has to be computed separately for any interaction group which is large enough to be a zerg.

So basically if you cause lag your abilities do diddly?

Yes, this would solve the problem - the role play reasoning would be - if people are too close together, they interfere with each other and are less effective.

It is not a problem to compute all this without lag in a lot of smaller groups, but zergs with AoE spam are a nightmare and create lag.

With this system if I run into them as a proxy det bomber or hit them with siege, those calculations would also be impacted, so my damage to them would be diminished as well then. So would a group next to them, like a group of five enemies who they are swarming over. So, then, they would be shielded in a way, and take anyone interacting with them down the diddly squat rabbit hole? Or could this still work with the graph being Alliance specific? So I can then hit them hard while they fumble about?

EDIT: Or by interaction group do you just mean formally grouped? Then I can see people still swarming but just in smaller formal squads. If it's any interaction by members of an Alliance who are in close proximity, though, I can see that working.

Reykice

I pay so i expect a working product. Imagine if cars worked like this:

"Hey guys we can`t track the software bugs that feed to much fuel and set the car on fire, too much code to go through".

Or well of anything worked like that really.

They have no excuse except that bug fixing can`t be sold in the store... and they can do it due to few regulations for the gaming industry. Imagine if they had to refund + give extra money to the people experiencing critical bugs, i assure you that then there would be a good Quality Assurance department tracking and fixing every bug before release because then it would cost them.

So tldr, as long as they don`t care there won`t be any major bug fixing because well, they don`t care all that much to dedicate resources(people) to actually do it. If you can take it and have no problem knowing this then all is cool, if not just vote with your wallet.

Lysette

tinythinker wrote: »

Lysette wrote: »

tinythinker wrote: »

Lysette wrote: »

There would be a quite simple solution though, and I have suggested in many times - reverse scale the efficiency of spells and attacks with the effort required to resolve the interaction in a zerg - this would make it highly undesireable to form a zerg and things would get better. Meaning the more effort is required to resolve the interaction graph, the less effective spells and attacks coming from the zerg will get. Simple solution, zergs would be highly undesirable, and the lag would be significantly less. Of course this is based on that they have actual interaction graphs and do not just compute it in a brute force way. It has to be computed separately for any interaction group which is large enough to be a zerg.

So basically if you cause lag your abilities do diddly?

Yes, this would solve the problem - the role play reasoning would be - if people are too close together, they interfere with each other and are less effective.

It is not a problem to compute all this without lag in a lot of smaller groups, but zergs with AoE spam are a nightmare and create lag.

With this system if I run into them as a proxy det bomber or hit them with siege, those calculations would also be impacted, so my damage to them would be diminished as well then. So would a group next to them, like a group of five enemies who they are swarming over. So, then, they would be shielded in a way, and take anyone interacting with them down the diddly squat rabbit hole? Or could this still work with the graph being Alliance specific? So I can then hit them hard while they fumble about?

Nah, you can compute the center (location) of the interaction group - weighted - and see all those who are in a certain range of this center as effected by the reduction and those outside have normal conditions. This would not effect the siege attack, because it is further away from the zerg. And a weighted center will give a better center than a normal average - and it is as well no extra computation, because the squared distances required to do that are easier to calculate than the actual distance.

To do that alliance specific would be counter-productive - because the moment 2 zergs encounter each other, there is a sudden increase in effort required to resolve the interaction - the whole combined group creates the lag.

Elsonso

Lysette wrote: »

TequilaFire wrote: »

Lysette wrote: »

tinythinker wrote: »

JD2013 wrote: »

Never raged. I see no point in raging on a forum.

However, I did find that segment particularly interesting. I know very little about how bug fixes work, so it was nice to see some insight into the process.

One thing that I will probably say trips ZOS up a lot is how differently a test server works compared to the live server. There's lots of variables that they can't or don't anticipate. For example, I can imagine that it is hard to recreate what happens on a live server in PVP, given that the server and indeed the client has to calculate so many different numbers at any given time.

And then, when they add thousands of lines of new code in with each DLC, that can create some havoc.

Is it an excuse? No. But sometimes things just can't be calculated no matter how much testing goes on. Could this game do with some shine and polishing? Sure could in some areas. But I can only imagine the stress that goes on in their offices.

That's one thing I've asked about and still don't get/never got an answer about.

I understand coding with so many interactive blocks of code can be tricky. I understand testing is hard. I understand bug chasing/squashing is hard. I understand that we each experience the game from different perspectives, have different encounters with performance, and want our own priorities moved to the top of the list. I also get that they say that they can only do so much on the PTS and also that buying more server space for live isn't a silver bullet. I'll grant each point.

Fine. OK. Sure.

But why *not* spend money increasing PTS server space to get it closer (not close, not the same, but closer) to conditions on live? I can only speculate that the bug/performance concerns aren't hitting the bottom line hard enough to justify the expense, but it could potentially save so much time and money spent going back to find and fix things after they hit live.

A live server like this is about 1.5-2 metric tons of hardware - this is not just a few blades in a rack. And it costs quite some amount of money as well. And even if they would use a copy of the live server, they do not have the load which is required to make sure it all will work, when the system comes under production load.

1.5 to 2 metric tons?
I see the problem now they are still running IBM 360s. lmao
/joking

This is really what something like this is - CCP has just ordered the hardware for the new TQ live server - which has to deal with about a similar amount of concurrency - and they said, the new hardware was 1.5 metric tons in weight.

For reference, here is the PC/Mac EU megaserver circa 2014

tinythinker

Lysette wrote: »

tinythinker wrote: »

Lysette wrote: »

tinythinker wrote: »

Lysette wrote: »

There would be a quite simple solution though, and I have suggested in many times - reverse scale the efficiency of spells and attacks with the effort required to resolve the interaction in a zerg - this would make it highly undesireable to form a zerg and things would get better. Meaning the more effort is required to resolve the interaction graph, the less effective spells and attacks coming from the zerg will get. Simple solution, zergs would be highly undesirable, and the lag would be significantly less. Of course this is based on that they have actual interaction graphs and do not just compute it in a brute force way. It has to be computed separately for any interaction group which is large enough to be a zerg.

So basically if you cause lag your abilities do diddly?

Yes, this would solve the problem - the role play reasoning would be - if people are too close together, they interfere with each other and are less effective.

It is not a problem to compute all this without lag in a lot of smaller groups, but zergs with AoE spam are a nightmare and create lag.

With this system if I run into them as a proxy det bomber or hit them with siege, those calculations would also be impacted, so my damage to them would be diminished as well then. So would a group next to them, like a group of five enemies who they are swarming over. So, then, they would be shielded in a way, and take anyone interacting with them down the diddly squat rabbit hole? Or could this still work with the graph being Alliance specific? So I can then hit them hard while they fumble about?

Nah, you can compute the center (location) of the interaction group - weighted - and see all those who are in a certain range of this center as effected by the reduction and those outside have normal conditions. This would not effect the siege attack, because it is further away from the zerg. And a weighted center will give a better center than a normal average - and it is as well no extra computation, because the squared distances required to do that are easier to calculate than the actual distance.

So siege would wipe them (heals so weeeak) but if I or my small group ran into them we would get nerfed too. Hmm, well, it would definitely produce an interesting new dynamic. Siege+ranged attacks = upper hand unless the swarm reaches you and sucks you into their nightmare of weak sauce.