From Blizzard's statements, it sounds like both software (code related) and architecture issues (hardware scaling). There are many, many factors when dealing with a project of this size, be it a game or not. Not all of these issues ever come to light during function & stress tests, though we would all like them to. Most of the time, a company will stress test against what they feel is the most reasonable user population estimate & also a "worst case scenario" estimate to find the absolute breaking point for an application under load.
Judging from their comments, they did this, however the game's popularity largely blew away anybodies prediction. I would suspect that played the largest part of the problem. Blizzard probably did their best to estimate the worst case scenario to make sure their architecture would scale for the holiday season (and on), but they fell drastically short. Could they have estimated better, probably. How, I am not sure.
Now, if their architecture scaled as it should, they should simply be able to add more hardware and boom, be off and running for another few thousand users. However, it sounds like this is where they found software problems that kept that from being so easy. Either straight out bugs or just bottlenecks, these software related issues can keep projects from scaling as they should. It does sound like they identified some or most of those issues, and are working on them as we speak.
Another possible factor, as that happened to Star Wars Galaxies during the first week or two after release, they may have come across a problem or bottleneck with a vendors product (be it hardware, or software like a database). I think Galaxies was using Oracle, and they came across a problem (either with their database configuration, or just a bug in Oracle's product, I don't remember) with the database that was not discovered (and probably could not have been) during the beta test. Most beta tests probably don't provided anything close to real load for this type of game, but they should provide enough accurate data to simulate and estimate as best as possible. Their is a known risk that bugs and issues will likely come up after testing, that almost always tends to happen.
I am pleased at how Blizzard seems to be addressing the problem and that they are working hard to fix the issues as quickly as possible.
I have been playing very happily for over 2 weeks with absolutely no problems. I have a few friends that are on the same boat (no problems for a few weeks). Luckily, we are not on one of the servers that Blizzard decided to migrate over last week. I do expect that I will have some problems in the future as they try to resolve these issues and/or more unhappy people try to move over to the lower population servers. Only time will tell.
I have to commend Blizzard on the game and their approach to handling problems. It is never easy to manage a MMOG, but I think they are doing the best they can.
I believe that everyone has at least some responsibility. I am definitely a fan of creating intuitive applications, in fact we always do the best job we can at work - and that work is always improving (trying to match the ever growing expectations of our users). However, as a programmer, I do find fault with those that think that something magical & immediate will remove all user responsibility.
As others have mentioned already, most everything a user does is learned behavior. Everything from turning the computer on, what a mouse is, what happens on the screen when the mouse is moved on the desk, etc, etc. It all has to start somewhere.
Remember action/reaction? For every action a potential risk or consequence (of something going wrong, expected or not)? I think there is some truth to that, even in this context.
As you mentioned
The computer is a virtual world. The real world has rules that cannot be changed and can be deadly (like gravity and inertia)
Computers are not a virtual world. They may represent a virtual world to the user, but they must obey the same laws of physics that I do. You raise a good point though, that programmers make the rules within the computer. To the users, we (programmers) do seem to make the rules. And we have a responsibility to the users to "give them what they paid for", within reason and constraints (some of which you raised)
I believe that everyone has some sort of responsibility (be it large or small) within the context of the issues raised by Tog. For example (just to list a few & definitely not an exhaustive list):
Hardware manufacturers should continue to improve hardware to meet the needs of users, plus continue to be innovative at the same time (because users don't always know what they want, or what is possible with new technologies etc). They need to provide manuals, instructions + specs, and make sure that their hardware works consistently & as described under said conditions. goals: to create robust and ground-breaking hardware that performs as desired risks: prohibitive cost & failure to go to market
Developers, like myself, need to do the same. We need to listen to, study, try to understand and help our users by making software more intuitive, robust and still try to maintain an innovative edge. goals: to develop intuitive, robust & flawless software risks: prohibitive development cost, increased likelihood of bugs in complex systems (potential damage costs increase as well)
Users need to read instructions, manuals and help when they need it. They must continue to relay their wants, desires and support questions to those that need to hear them. However, as I pointed out before, they are not totally free from all responsibility. When a beginning user buys their first computer, their responsibility is to learn how to use it (plug it in, turn it on, move the mouse, learn how & when & where to click etc). When experienced professionals buy an expensive/complex system, you better be sure that most companies will send their employees to training or something, to help ensure that they know how to use it properly (thus reduce risk and potential damage & costs). goals: obtain & use completely user-friendly software & hardware risks: prohibitive cost. Potential damage to self/property/lives if instructions and/or proper training was not followed. (for casual users, mostly just data loss)
You will notice that "prohibitive costs" kept on creeping up in all categories. Unfortunately this is very real and must be dealt with as we go. Nobody will develop/engineer for free or just throw money away, there must be an analysis and a decision somewhere.
So while Tog made some very valid points, I don't think they are all easily obtainable or even realistic to fix (at this stage at least). Keep in mind that the perfect world solution to all his problems span a
But when I read "Principle: The user is in charge and should be free to carry out any activity at any time without fear of reprisals" I just about lost my lunch.
I haven't been able to read the article yet, since it appears to be/.'ed, however I would have to agree. As with driving a car, flying a plane, or just about operating anything, use comes with some responsibility.
As for the above, swap user with driver and you may see my point.
Slashdot Mirror
From a programmer's (and avid gamer) perspective:
From Blizzard's statements, it sounds like both software (code related) and architecture issues (hardware scaling). There are many, many factors when dealing with a project of this size, be it a game or not. Not all of these issues ever come to light during function & stress tests, though we would all like them to. Most of the time, a company will stress test against what they feel is the most reasonable user population estimate & also a "worst case scenario" estimate to find the absolute breaking point for an application under load.
Judging from their comments, they did this, however the game's popularity largely blew away anybodies prediction. I would suspect that played the largest part of the problem. Blizzard probably did their best to estimate the worst case scenario to make sure their architecture would scale for the holiday season (and on), but they fell drastically short. Could they have estimated better, probably. How, I am not sure.
Now, if their architecture scaled as it should, they should simply be able to add more hardware and boom, be off and running for another few thousand users. However, it sounds like this is where they found software problems that kept that from being so easy. Either straight out bugs or just bottlenecks, these software related issues can keep projects from scaling as they should. It does sound like they identified some or most of those issues, and are working on them as we speak.
Another possible factor, as that happened to Star Wars Galaxies during the first week or two after release, they may have come across a problem or bottleneck with a vendors product (be it hardware, or software like a database). I think Galaxies was using Oracle, and they came across a problem (either with their database configuration, or just a bug in Oracle's product, I don't remember) with the database that was not discovered (and probably could not have been) during the beta test. Most beta tests probably don't provided anything close to real load for this type of game, but they should provide enough accurate data to simulate and estimate as best as possible. Their is a known risk that bugs and issues will likely come up after testing, that almost always tends to happen.
I am pleased at how Blizzard seems to be addressing the problem and that they are working hard to fix the issues as quickly as possible.
I have been playing very happily for over 2 weeks with absolutely no problems. I have a few friends that are on the same boat (no problems for a few weeks). Luckily, we are not on one of the servers that Blizzard decided to migrate over last week. I do expect that I will have some problems in the future as they try to resolve these issues and/or more unhappy people try to move over to the lower population servers. Only time will tell. I have to commend Blizzard on the game and their approach to handling problems. It is never easy to manage a MMOG, but I think they are doing the best they can.
As others have mentioned already, most everything a user does is learned behavior. Everything from turning the computer on, what a mouse is, what happens on the screen when the mouse is moved on the desk, etc, etc. It all has to start somewhere.
Remember action/reaction? For every action a potential risk or consequence (of something going wrong, expected or not)? I think there is some truth to that, even in this context.
As you mentioned
Computers are not a virtual world. They may represent a virtual world to the user, but they must obey the same laws of physics that I do. You raise a good point though, that programmers make the rules within the computer. To the users, we (programmers) do seem to make the rules. And we have a responsibility to the users to "give them what they paid for", within reason and constraints (some of which you raised)
I believe that everyone has some sort of responsibility (be it large or small) within the context of the issues raised by Tog. For example (just to list a few & definitely not an exhaustive list):
Hardware manufacturers should continue to improve hardware to meet the needs of users, plus continue to be innovative at the same time (because users don't always know what they want, or what is possible with new technologies etc). They need to provide manuals, instructions + specs, and make sure that their hardware works consistently & as described under said conditions. goals: to create robust and ground-breaking hardware that performs as desired risks: prohibitive cost & failure to go to market
Developers, like myself, need to do the same. We need to listen to, study, try to understand and help our users by making software more intuitive, robust and still try to maintain an innovative edge. goals: to develop intuitive, robust & flawless software risks: prohibitive development cost, increased likelihood of bugs in complex systems (potential damage costs increase as well)
Users need to read instructions, manuals and help when they need it. They must continue to relay their wants, desires and support questions to those that need to hear them. However, as I pointed out before, they are not totally free from all responsibility. When a beginning user buys their first computer, their responsibility is to learn how to use it (plug it in, turn it on, move the mouse, learn how & when & where to click etc). When experienced professionals buy an expensive/complex system, you better be sure that most companies will send their employees to training or something, to help ensure that they know how to use it properly (thus reduce risk and potential damage & costs). goals: obtain & use completely user-friendly software & hardware risks: prohibitive cost. Potential damage to self/property/lives if instructions and/or proper training was not followed. (for casual users, mostly just data loss)
You will notice that "prohibitive costs" kept on creeping up in all categories. Unfortunately this is very real and must be dealt with as we go. Nobody will develop/engineer for free or just throw money away, there must be an analysis and a decision somewhere.
So while Tog made some very valid points, I don't think they are all easily obtainable or even realistic to fix (at this stage at least). Keep in mind that the perfect world solution to all his problems span a
As for the above, swap user with driver and you may see my point.