Slashdot Mirror
Data Harvesting From a Developer's Perspective
cliffski raises some questions about the need for game developers to have some amount of data from the users who play their games. He says,
"PC Games connecting to a central server to send information (outside of MMOs) have gotten a (deserved) bad reputation in recent years. The huge outcry about Mass Effect and Spore are evidence enough of that. But in gamers' hurry to prevent intrusive DRM systems and dubious privacy-breaking data harvesting, are we throwing out the good with the bad?"
Clearly, some aspects of games could be improved by having a better knowledge of average PC specs or knowing which parts of the games are more entertaining to the users. Input from customers helps to improve almost any product, as indicated by the usage of countless surveys and focus groups. But where do we draw the line between being inquisitive and being intrusive? What can game developers do to prove that the collection techniques or the data themselves wouldn't be abused?
Shockingly, new studies suggest that people may be able to make decisions all by themselves without a company or a government or anything!
Spelling mistakes, grammatical errors, and stupid comments are intentional.
I recall seeing detailed info collected from the new Team Fortress on what classes were selected, "heat maps" of death locations, etc. Looked to me like it was all valuable info, especially for the game and map developers. I know Steam keeps a backend connection going, and it seems like this data could be really useful. While I'm definitely against collecting personal data, the aggregate stuff should be just fine from a privacy standpoint.
Jonah HEX
Horror & SciFi Erotic Nudes
I have no problem with software collecting data that is aggregated and not kept in an individually identifiable format.
There should also be a way to opt-out with no negative repercussions (feature disabling).
Those two simple guarantees and I'm comfortable. The problem is I don't trust a corporation to be honest and forthright with the handling of customer data.
I agree with this sentiment. Set up an opt-in program that allows gamers to share their information with game companies. If a player is truly invested in the game, they will share their data to support further improvements in the game. The players that are most invested are the ones for whom you want to tailor your games, no? Sounds like a win-win to me.
Seriously? Ask us first. Tell your users exactly - EXACTLY - what's being monitored, and 98% of the problem goes away. Users are sick and damn tired of being misled and lied to about stuff like this for our own good.
Maybe from the Developer's perspective having an intrusive all-seeing eye installed on everyone's computer which either can't be turned off or only via a default-selected checkbox in the disused lavatory tab of the options menu sounds like a good idea, but to anyone else it really doesn't. Don't do it.
Be honest with your users or they'll hate you whatever you do.
I hate to say anything good about Steam, but this is one thing they get right - they simply ask.
todo - The developer's equivalent of confession: "Forgive me Father, for I have sinned..."
They usually mention such things in the license agreement. Do you read the license agreement?
That's not asking is it? That's telling someone after they've purchased the product that aspects of their system will be monitored. Of course you can take the product back but that's inconvenient. Technically everyone should read the licence agreement but the plain fact is that nobody does, and while that's obviously got no legal standing, if the developers don't want to be hated by everyone using their product they should cut out the sly bullshit and ask in plain and simple English. If the developers really do just want certain info regarding gameplay or system specs I'm sure that enough people would say yes to get a representative sample of users.
Spelling mistakes, grammatical errors, and stupid comments are intentional.
You know, I'm not a big fan of cops, but it never ceases to amaze me, how software engineers on ./ rant and rave about everyone collecting information on other people, but make every exception for themselves.
IF civil rights is that important, that you want to go on and rail about Obama's FISA betrayal and horridly fill out online donations to the ACLU over the idea of your government collecting information to aid in counter-intelligence against not only the "terrohistas", but also the Chinese, Europeans and anyone else who might have their information collected by their governments, then that's worthy.
But, I would like to know, what exactly about a video game, shopping experience or some other fluffy adventure that entitles you as a software developer to violate people's rights to privacy, for your own ends, when you would deny that same efficiency to everyone else? You aren't elected to represent anyone, but our government is.
This is my sig.
Data collection should be considered intrusive unless the user is warned beforehand and/or has the option to disable it.
A good example is popularity-contest in Debian and I think it was Winamp that also asked if you wanted to let it send anonymous statistics.
the problem with relying on people to make decisions is that people might not do what I want them to.
There. Fixed that for you.
And I don't consider that a problem. If such a small amount of people do say yes then all that's doing is clarifying how many people you're potentially pissing off by forcing such decisions on users. And to be perfectly honest if you word the question correctly and explain how it can help development I think you could get considerably more than 1% of users to accept. It would still probably be a very small proportion of users, but it should be enough to gather useful stats on the kind of information TFA claims developers want.
Spelling mistakes, grammatical errors, and stupid comments are intentional.
Remedy:
Promise that amongst everyone who shares their system info, once a week/month/year someone wins a prize (no need to ask delivery info in advance, just that "If you win, you'll be notified via this program and then asked to enter delivery address"). The price can be anything cheap and token-ish (eg. in games, some bonus freebie item), as long as there's something.
Speaking from experience I can tell you that an "opt-in" program would never collect enough data to be useful.
I'd suggest an "opt-out" system along with restrictions on *what* data was sent. At least I'd say that nothing personally identifiable can be sent, there's no need for it. There may be other restrictions I can't think of right now.
IMHO, this issue is about what data gets sent, not that data gets sent at all. It should be clear and verifiable what data is being sent, so that users who are that way inclined can check to ensure that nothing untoward is being sent to the developers.
I hate printers.
Valve already does monitoring with their games, and I don't think anyone complains about it. For example, I know in Team Fortress 2 they keep track of which team wins the most, where people die the most, how heavily certain classes are used, etc.
They don't know where to draw the line already with invasive DRM that locks us out of our own games. Why would it be any different with private data collection?
Twinstiq, game news
What can game developers do to prove that the collection techniques or the data themselves wouldn't be abused?
That's easy. Just give me a checkboxed list of all the data items from my computer that you propose to send to your server. Then provide an "UNCHECK ALL" button so I can still maintain my privacy.
The higher the technology, the sharper that two-edged sword.
Of course you can take the product back but that's inconvenient.
It's more than inconvenient, it's usually impossible. Most retailers refuse to take back computer software, especially opened software, as a matter of course. And you would have to open the software to get into the installer to read the EULA (Which, in some cases, you "agreed to" before even seeing it, with phrases on the CD case like "By opening this package you agree to be held liable to the End-User License Agreement contained therein", a so-called "Shrink-wrap" license).
So saying that 'taking back software is an option' is, for most cases, wrong.
Maybe that should tell you that most people *don't* want to share with you with no compensation?
a handful of selfish greedy people are no match for millions of selfish, greedy people -u4ya
Precisely the point.
At our local mall, there's a survey and opinion company in the corner. They ask mall-goers for surveys based upon demographics and other information told by their clients (like Coca-cola, Pepsi ola, and others).
I've been asked about 8 times. I cannot discuss what was reviewed by myself, because of NDA. However, I received payment from 25$ to 75$ for said reviews. I also provided accurate demographic information, along with the proper write-ups.
I sold my privacy for a pretty penny. In some cases, I later bought some nice hardware for my computer. Why should I give it away when it is seeked and compensated for fairly?
That game developer has no clue about privacy.
First, if the game has online registration, that's the one time to collect, with the user's permission and knowledge, basic system configuration info. That's useful to have if they call for support. It doesn't require a continuous connection to a server.
Second, if more data is required for game tuning, that's what play testers are for. Or free beta users. It's reasonable to have a free beta that sends back play data, if the developer is up front about it. It's not reasonable to have it in a paid product.
Third, if you can't meet basic EU privacy regulations, your market is much smaller.
Valve would disagree.
http://www.steampowered.com/status/survey.html
http://www.shacknews.com/onearticle.x/52707
1,728,662 Steam users have voluntarily agreed to participate in their semi-annual hardware survey by having detailed specification of their PC hardware cataloged.
Game-related bonuses could be offered in exchange for the data. Additional network subscription time, screensaver images, free game area/map (selected from several which are sold, so identity of survey participants is hidden).
While we're talking about opt-in and opt-out systems, how about a preferences system for Slashdot which allows you to decide whether you'll allow morons to post mangled quotes of your posts followed by inane phrases like "fixed that for you"? There could be three options: "no", "yes", and "yes, as long as he acknowledges that by doing this he is a complete asshole."
If you mod me Overrated, you are admitting that you have no penis.
Is data from over a million and a half samples not useful?
Seriously though, Steam's hardware survey is the first thing that came to mind when I saw this story. It's non-intrusive, it clearly asks you before sharing any information, and it keeps the summary information available for all to see. I probably wouldn't mind sharing technical information if it worked similarly to this.
LegendMUD
There. Fixed that for you.
I like it when my quotes are used out of context!
There, fixed that for you.
"He who can destroy a thing, controls a thing." --Paul Atreides, Dune
The problem with an opt-in approach is that you'll only hear from the vocal minority. Most of the time, that's the *worst* demographic to make decisions based on. Blizzard has done an excellent job of not falling into that trap as compared to, say, SOE.
Interested in open source engine management for your Subaru?
Companies like Valve and Microsoft have already adopted this mentality — they don't just capture information about how you play the game, they store it in an online profile, and let you unlock achievements, compare your data with others, or view a chart of your own scores to monitor your improvement.
Ironically, by making this kind of data public, you'll cause players to start putting less value on their own privacy. It's the Alcoholics Anonymous effect in action — when other people disclose private information, you're more likely to disclose private information too.
Of course, this doesn't mean that you as a developer should be collecting any sort of truly private data. If you can't explain to players in detail what data you're getting and why you want it, you shouldn't be collecting it.
Also, provide a simple way for players to provide spur-of-the-moment feedback on your game. For example, add a simple text box to the game's pause screen that lets users zip off a note to the game developers, along with data about where they are in the game and their current status. I can think of a hundred times when I would have given the developer feedback but was stymied by the hassle of finding the proper web site, setting up an account, explaining the situation in detail and not even knowing that anyone on the development team actually read the message boards. A quick message system built in to the game would be much handier to players, would collect raw off-the-cuff impressions, and best of all, would be entirely opt-in.
How can I believe you when you tell me what I don't want to hear?
Instead of trying to transform the gamers community into labs rats in order to find new ways of selling us more crappy half finished entertainment (games in this case), try to focus on creativity and innovation.
... Make more money with lower costs and very little concern about your product other than if it will sell.
How many crappy games are released today because they are unfinished, bugged and unplayable? Of course data mining is the better economic plan since it allows to collect money by releasing the game early and pretend you care about your customer base instead of beta testing your product (This actually cost more money and delays the production cycle).
Data mining is flawed: It collects data about what we like. The result is a massive amoung of clone games with very little creativity other than mind blowing GFX. Focus on something new instead, something we havent seen yet, something original, something that will be a surprise instead of the version 65 of a "well selling title".
Beta test your product, (data mining isnt a cheap way of doing it) Ask for feedback if you like (There is plenty of discution forums the gamers will be happy to contribute). Funny part about this, most gamers have the impression you never read those since you rarely answer them, yet you pretend you want to collect data using a sneaky method while you ignoring most of our suggestions/feedback on discussion boards. Makes me wonder about your real motives
We have lots of cases where companies have collected this information and then done Evil Things with it, so people are reluctant to provide it.
So --
-- stop being evil. Start using information only for benign purposes, and then people will trust you in time. ... in time. You screw people over, you have to *stop* screwing them over first, and only then figure out how to regain their trust.
It is reasonable to have it in a multi-player game where players are logged into a central online service. Ladders and ranked play require some transmission of game statistics to function at all. Since we require a mechanism to transmit that data, we can gain a lot by extending it to capture detailed game statistics as well. The types of games that require public betas are usually online multi-player games that fit the above criteria, and the need to continue monitoring that data won't go away once the game ships. I'm not sure of a scenario where a developer would publicly beta a game that wasn't an online title, so I'm not sure where your above mentioned scenario fits in the real world.
Glad you brought this up - I only dropped into this thread to point out the rather excellent Valve Hardware Survey.
The fact it is self-selecting does make it a shade biassed towards the high-end, but it is amusing to see the sheer amount of laptop hardware out there with Steam installed.
It is always funny to smirk at the glacial pace of Vista migration too.
The 40 EUR I spent on ESET's Smart Security package is probably the best money I've spent on software in the last 5 years. First time I started up Mass Effect I was greeted with a warning from the ESET firewall about the game trying to access my internet connection. Check "Make a rule", click "Deny". Problem solved.
As for how companies should approach information gathering, I'm with most everyone else here. Simply explain to the users exactly what information is being collected and give them the option to opt out. I say exactly because a lot of it depends on how you ask. If I'm greeted with a simple question like "Allow Mass Effect to send anonymous usage statistics to Bioware?" I'll probably click no, because I have no idea what "anonymous usage statistics" entails. Ask the same question and give me a list of exactly what information will be sent out, how often and to where, then I'll be more inclined to agree to it. Best case scenario, actually show me the information being sent and let me click the send button. Just don't do it so often.
Companies think they have to sneak this phone home stuff in because people don't like it, they don't realise that most people don't like it precisely because they try to sneak it in. The rest just don't like it at all, so let them opt out. Everyone's happy.
Murphey's fighting Occam, and we're in the stands.
Echoing through the cave, you hear a voice in the distance call out "I before E except after C".
I'm an American. I love this country and the freedoms that we used to have.
Is data from over a million and a half samples not useful?
Who's the wiseguy that sent in the half a sample?
http://crummysocks.com
I believe that shrink-wrap licenses have been found to be nonbinding in some jurisdictions, though I don't remember the court cases that back that up.
Perhaps, or even probably, but that's not the point. Companies offer money back guarantees and mail in rebates because they know most (or at least a statistically significant proportion of) consumers don't find it worth their while to bother. Hiring a lawyer and loosing time off work (or just plain opportunity costs) for a sixty dollar game isn't worth it. Of course you are assuming that the consumer even believes he can get his money back. In the era where laws and information campaigns are constantly reminding consumers that their rights take a back seat to that of the copy right owner, I doubt if too many (consumers) would even consider the law, much less attempt to try and enforce said law. It's not so easy to just call the police and tell them to press charges against a software company because their EULA or business practices are illegal. Believe me I've tried :)
Chances are however that if you call up the actual game company and they have some reasonable customer service rep on the line then they may refund your money if you pay to have the disks, etc sent back to them. Maybe, maybe not. It's your bet. In the end it's still a hassle. It's often easier to buy something than to return it.
No not really.
If you for instance get invited into the community theres a big button telling you press here to enable this feature. No where does it explicitly tell you that hitting that button will add all sorts of tracking information to your account freely available to any one else - and no way of opting out again. (This might have changed after I pointed out to them that their practice was in fact illegal and I would take it up with local consumer agencies if they failed to remove this information for my account - I can't actually check that because their current method of removing your information is to ban your account from the steam community (no I'm not kidding))
1. Ask. Simple as that. Ask. Don't just go "and now we'll transfer your demographics to our maker, hit ok", or don't say a thing altogether. People love the feeling of being in control. And they will much more readily provide you with information if you take the time to tell them what you need it for, i.e. making the product more suitable for your customers.
2. Let me review the information before it is sent. Let me see just what information you want from me. I'm uneasy when I'm asked to let a program gather information from my computer and send it to you. Let me see what information you want, if you want to be sure I let it pass, give me a reason why you want information aside of my hardware specs, because I can't see how my name, the number or ID of connected machines or the directory structure of my hard drive(s) could possibly help you develop a better game.
3. Don't wrap it in legalese junk. KISS is the key here. If you want to cram license agreements down my throat that require me to get a law degree and read for three hours, I will not send you any information whatsoever. State that the information is going to be used anonymously, that you will not store the IP address it was sent from, that you won't bombard my mailbox with junkmail and that you will not distribute the information. After all, you only want it to improve the games you make, right? So it should be no problem for you.
You can without a problem do all this as part of your installation routine, completely automatized, and if someone doesn't care about any of those things he can easily bypass the agreement, the list of information gathered and the terms of usage for the data collected. If he cares about it, he can read it.
Where's the problem with that?
We used to have a Bill of Rights. Now, with the rights gone, all we have left is the bill.
if it worked along the lines of Microsoft's 'customer improvement program' or whatever it is called for office and stuff it would probably work
null
> Not even my employer has any right to demand that I keep a secret
No one can demand that you keep a secret: you can decline when they offer to pay you money to do so.
> ...especially if one day I might be in court to testify about something that happened
> in the workplace.
No contract keep you from answering any questions you are ordered by a court to answer nor can it penalize you for doing so.
> NDA is meaningless.
A properly drafted and executed NDA is an enforceable contract.
Warning: this article may contain humor, sarcasm, parody, and perhaps even irony. Read at your own risk.
Yes,I honestly care. I care that they didn't ask. I don't care about the specific things that were monitored. I worry that, knowing it's okay to spy on me here without asking, they'll slowly start gathering more and more personal information until they finally do take things I wouldn't want to spread around. So long as I am asked, and have the ability to say "that's going too far," I'd be happy. If you just give them a carte blanche someone will abuse it eventually.
"A witty saying proves nothing." - Voltaire
Blizzard has done an excellent job of not falling into that trap as compared to, say, SOE.
Well ... I'm not sure what you mean by "Blizzard has done an excellent job of ...", but a number of months ago, my WoW client, which I run under WINE, started crashing right after an update. I went to their forums to discover that many others running it under WINE were crashing after the latest update. Blizzard's rep basically toted the company line that they don't support running WoW under anything other than Windows and Macintosh. Fair enough.
After poking around for while, I discovered that it was crashing on an ActiveX call. Why would that be? I was running it using the OpenGL API setting; there shouldn't be any ActiveX calls. I changed the config file to use ActiveX, instead, and ran the client. The message popped up "Submitting non-personal system specifications" (or something like that), and they the client started right up without crashing. It turns out that before running the updated client they were executing an attached program that submitted your system specs to Blizzard - without asking or even informing the end user. The submission program was written using ActiveX calls, so running it in OpenGL mode in WINE caused the Windows client to crash.
Admittedly, Blizzard never claimed the WoW client would run fine under WINE, although Blizzard must be aware that lots of people out there are doing it. I suppose this is a great way to "not fall into that trap" of getting feedback from a skewed minority of users, but adding ActiveX-only calls to a mandatory program that retroactively informs the user they just submitted their systems specs did not endear them to their WINE fanbase, and probably didn't thrill the rest of their users either (Note that the WINE fanbase may be a comparatively small number, but if even 0.1% - 1 in a thousand - of their 10 million user subscribership is running the client under WINE, that's 10,000 users).
"Can't you see that everyone is buying station wagons?"