I wonder what the comfortable medium between privacy and letting developers get feedback about how well their code works is. It seems to me like Wacom just wants to know if their drivers work, so they can focus engineering efforts around fixing the issues that are affecting their users. "Oh hey, the new beta of Photoshop breaks our drivers!" They don't make a "cloud product" and they have an obligation to make their hardware work with any software the user might want to use, so they are kind of painted into a weird corner here. If they collect data to drive their engineering, they're spyware. If they collect no data, they're a bug-ridden disaster area whose product you would never buy.
I am guessing that the answer will be "they should test everything in house and tell users to complain via email when shit is broken"... but we all know that synthetic QA is never going to be as good as "ground truth", and that 99% of users will just silently be unhappy. So I wonder what the privacy balance is here.
I understand the need for developers to know more about the hardware and software running on their client machines. For example, I believe information like the hardware survey from Valve [1] are very valuable for the whole industry.
But there's a some kind of an etiquette you need to follow, if a company wishes to collect data:
- Be straightforward. Say what information you are collecting, at what time and what for.
- Tell me in what way this information will be stored and how will it be anonymized.
- Will the data be stored forever? And is there a way for me to request the data or it's deletion?
- Don't collect data per default. Make it opt-in.
- Publicize the data in a suitable way. It may be useful to others.
Wacom simply ignored all of that human decency. How can you ever trust this company again?
> I wonder what the comfortable medium between privacy and letting developers get feedback about how well their code works is.
I consider the nut of the problem to be informed consent. If you have user's informed consent to get the feedback, then there is no problem. If you don't, then the whole operation is unacceptable.
And no, mentioning it in the privacy policy or terms of use don't count as "informed consent".
This would be a real challenge for some companies. Having a clear privacy policy creates a hard dependency between it and the code. And developers are notorious for not even being able to keep comments updated along with their code changes.
It's not impossible at all, just in the current state of the industry there's a good reason we have vague agreements (also including good old-fashioned laziness, of course). It'd probably need to be developed ground up as an API with side effects, so when the code is compiled it spits out some details about how it's used.
> Having a clear privacy policy creates a hard dependency between it and the code. And developers are notorious for not even being able to keep comments updated along with their code changes.
That's just a small extra step in the QA pipeline.
Also: analytics and telemetry code doesn't just appear out of the blue. Someone makes an explicit decision to scoop error logs from users, or track clicks, or spy on system configuration. That someone is usually higher up the management or technical chain, and should know enough to recognize that sending anything collected on user's machine that is not crucial (in the most strict, technical sense) to performing the action user activated has privacy implications.
So what? Informed consent is also "a real challenge" for some medical studies, does that mean we should let doctors carry out unethical studies?
I'm actually pretty sympathetic to Wacom in this instance, more sympathetic than the blogpost author at least. But unethical actions are unethical regardless of whether acting ethically is "a real challenge" for some companies.
The deep problems of “informed consent” are apparent in medical studies/treatment. Few patients are equipped to be informed because they don’t have a med school degree.
Since users ”can’t be informed” about tracking, it doesn’t make sense to discuss whether they “should be informed”.
Doubtless there are deep problems with "informed consent", but saying they "can't be informed" is nonsense. Is your plan to not bother to inform people because they "can't be informed", and decide what's best for them without their knowledge or consent?
> This would be a real challenge for some companies.
Tough. If a company can't do it the right way, they shouldn't do it at all.
> in the current state of the industry there's a good reason we have vague agreements
Well, I guess that depends on your point of view. I see no good reason for this, but I have no doubt that the various companies do see a good reason by their definitions.
You're right about the current state of the industry, but the current state of the industry is a travesty.
Yeah, 'good' was the wrong word. Maybe 'understandable' but that's still is a bit too charitable.
I was mostly musing about how changing code can have legal/business as well as technical side effects, and we've seen that to some degree with mobile app permissions who just grab everything because it's seen as too much effort to do it right. So I'm curious if this is going to change for the better any time soon.
Thing that I know happens, from personal experience: you can put a giant modal alert, and write in blinking, all caps, 60pt bright-red font that you will do something unless the user presses a button, then draw a bright red arrow to the button. Users will still complain that they weren’t informed.
Users are lazy and dumb, and the most ideological users are often the laziest and/or the dumbest, because they have an agenda. They will go out of their way not to give you the benefit of the doubt (”why was the font not 80pt? Clearly, you’re trying to hide something from users on high resolution screens!”)
If your goal is to eliminate user complaints about this, justified or not, then just stop intrusive data collection entirely. Then you don't need to bother with obtaining consent.
The way Steam handles the occasional hardware survey requests, which are purely opt-in, seems appropriate here. “Please select which applications you use your Wacom with” with a permanent opt-out checkbox would be quite acceptable. (Steam knows what games we play but not beyond that; Wacom must ask us to specify which apps we ‘play’ since the OS can’t be more specific.)
FWIW, as a customer of Wacom's products they very much do not view themselves as
> [having] an obligation to make their hardware work with any software the user might want to use.
They update drivers for 4 or 5 years then tell you to buy a new product if you expect it to work with current-gen software. Despite the fact that none of their tablets have had a substantial new feature in 20 years beyond the wireless connection kit, somehow a driver for a "Intuos Pro 4" cannot be used with a functionally-identical "Intuos Pro 3".
To me the comfortable medium is 100% privacy. 0% feedback. There is no middle ground because feedback and privacy should not be conflated. Users already own the device and owe no data to their vendor.
I've started turning off analytics everywhere. I turn off reporting on Firefox, Atom, everywhere. No crash report. Nothing if I can untick the box. Windows Firewall Control or LittleSnitch for all the outbound traffic as well. I don't let most windows services talk to the Internet unless it's updates or Windows Defender.
Some stuff is going to get through, but it should just be because you missed it. I'm sorry FOSS people; everyone is collecting way too much and I don't want to give Mozilla my data either. No, not even crash reports.
I have a Wacom tablet. The drivers don’t install on macOS any more. There doesn’t seem to be any technical reason for this. It’s a USB device (“essentially a mouse”) and it worked fine for several major versions. Maybe it was a 32/64 issue.
You’d think if keeping users happy was their primary goal, they might start by keeping their existing USB drivers compiled for the current macOS.
They don’t need me to email them to tell them it’s broken under current macOS. They’re the ones who told me!
I am guessing that the answer will be "they should test everything in house and tell users to complain via email when shit is broken"... but we all know that synthetic QA is never going to be as good as "ground truth", and that 99% of users will just silently be unhappy. So I wonder what the privacy balance is here.