> As a partial mitigation, Apple is enabling an alternative way for advertisers to measure campaign success, with Private Click Measurement ad attribution now available in Safari Private Browsing mode. Private Click Measurement allows advertisers to track ad campaign conversion metrics, but does not reveal individual user activity.
While as a consumer I do objectively like the privacy measures Apple is adding, at end of the day they're simply consolidating all tracking power to themselves.
Private Click Measurement is a standard that Apple has proposed and is working with the W3C to standardize, as well as working with other browser manufacturers:
To my knowledge, Mozilla's design is the only one where someone other than the browser collects & reports on click activity, and with a fairly trustless anonymizing double blind strategy for those intermediaries.
Without the online advertising business Mozilla cannot survive. As such, they advocate for preserving the online ads business. They partner with an advertising company. And they derive profit in return.
Opinions may differ, but to me being an advocate for it is no better than being "in" the advertising business.
(Not to suggest Mozilla is worthless. Far from it. I use Firefox on mobile, although I use a browser sparingly.)
> Websites should not be able to attribute data of an ad click and a conversion to a single user as part of large scale tracking
I'm curious, do people generally care about this specifically, or is it the sharing of this information with third parties that's the problem?
I run Firefox with in-built privacy protection enabled and uBlock Origin. However, I'm not doing so because I want to stop the websites I frequent knowing which email of theirs I clicked. My concern is when they share my data with other parties whose services I'm not explicitly engaging with.
Attributing data to a single user is what allows 3rd parties to amass a detailed profile of you (by looking at all your data from many websites that all share with the same 3rd party).
I doubt people are educated enough to care about this specifically, but the tech industry should care.
Of course they can only do this because of their cult-like ecosystem and hugely inflated prices for hardware, software and services.
Without those sources of revenue, well, Apple would turn to data and whatnot too. Was only a while ago that Apple first switched on to the "private and secure" as a selling point, because much of their competition relies on it.
I think we can maybe expect other large tech companies to follow suit, not make money from data/advertising and instead raise their prices, wall their ecosystems as much as possible. Advertise tablets as "not computers". Come up with buzzy marketing stuff like "retina" for every little detail of their products.
Wait until Google implements something and shoves it down everyones throat in Chrome and then has all the Google fanboys claiming that its the best thing since sliced bread and thus should get implemented by every other browser just because Google did it?
That's how we ended up in the situation where Google shipped U2F, sites implemented their implementation and then when the standard WebAuthN was built it was not compatible so sites had to be updated to be standards compliant, and it took a while to do so.
Or when Google added WebP without clear consensus. Or when they added FLoC or Topics API, or whatever else they have cooked up. Or things like WebUSB, WebMIDI and others?
There's a glut of Chrome only sites out there, and it continues to grow as web developers test just on Chrome but not the other browser engines. It's turning into the next IE 6, I remember the time there were a lot of "Made for IE 6" logos and graphics on sites and they did not render well or at all in Netscape.
As much as the locking down of iOS is annoying for everybody technical, we should be somewhat thankful that Apple has ensured a large population of mobile safari users.
That's a pretty poor comparison, a lot of the Android vs iOS debate is more akin to buying a sports car and then finding out it won't go more than 50km/h outside of your own country.
Back before flagship prices inflated (due to Apple) people buying a $1k iPhone couldn't do something as simple as use a different keyboard. And of course the product is advertised as advanced tech, yadda yadda.
There's been a spate of things, custom keyboards, wallpapers, multi-tasking, installing apps from a store rather than built-in, home/lock screen widgets, default apps for certain files, PiP, app drawer, and so much more.
I don't consider companies being inspired by/adding features from the other, it's not theft when it's a good idea (unless it's Apple's, then you get sued). But for them to market the way they do while taking so long to add basic things like these, it's just ugh.
Keeping in mind, of course, there are plenty of people who won't turn off the TV because there's nothing on that they want to watch; they'll just pick the channel that is the least bad.
Except in every thread about iOS or Safari are a ton of people crying for real Chrome on iOS because Safari is the new IE 6 holding the web back by not implementing the new WebDogCam4 “standard” Google pushed out 2 days ago.
^ This, 100%. Chrome is by far the bad actor, and not only that, Safari is arguably the better browser even if strictly looking at "support for web standards" (but in many other ways as well).
The vitriolic hate it gets in many threads are completely misguided and likely the result of years-old opinions on it. In the last 3 years Safari dramatically accelerated development, leapfrogged Chrome in performance to a staggering degree, and basically became close to an ideal browser.
And nearly every so-called standard people point to to "prove" Safari is lagging behind is almost always just something Chrome pushed out without any consensus.
It's funny because I think the hate comes from Webkit being forced on iOS, but it often comes out as "Safari sucks it's the new IE" which is pretty much the opposite of true and undermines the point.
Safari used to be "the new late-IE" a few years ago. It lagged significantly behind other browsers and it kept intentionally holding back support for open standards and codecs, forcing websites to make Safari-specific workarounds whenever you wanted to do basic things (I had to write scripts to transcode Vorbis to MP3 when deploying a web game just so it could have sounds on Safari, for example).
These days Safari gets better indeed (through it's still a PITA in some areas), while Chrome is clearly "the new golden-days-IE" - which long-term is probably much worse than Safari could ever be.
As someone who grew up with mp3s, your example seems interesting because it seems like an example where you had to encode from something obscure (ogg) to the closest thing to a “universal” sound format as wav could be: MP3.
In that case doesn’t it seem like Google took an ideological stance by choosing a non-patent-encumbered codec instead of supporting mp3? And they could do it because of their dominance? Or is that not accurate?
Vorbis was ubiquitous on the Web way before Google had such influence over it as it has today, and MP3 was absolutely nowhere near universal. Firefox did not support MP3 for a while, and even once it did it relied on system codecs - and many distros didn't ship with MP3 support by default until its patents expired a few years ago.
Choosing non-patent-encumbered codecs for open standards is as much ideological as practical.
I’ve never seen Vorbis and ubiquitous in the same sentence. Besides Mozilla, who had Vorbis native support in their browser?
In practice, MP3 has had support on every platform since the mid 90s. I could probably count the number of times I’ve come across an ogg file being distributed on one hand.
> Besides Mozilla, who had Vorbis native support in their browser?
Pretty much everyone who implemented HTML5 audio for a few years, while MP3 support in the browser was initially barely heard of? Mozilla, Google, Opera, KHTML, WebKit... Safari and IE were the only outliers, and had tiny minority of market share for a while (not counting versions that didn't support HTML5 audio at all). Today only mobile Safari is still an outlier.
You couldn't use MP3 without Vorbis fallback on the Web for about a decade. Whenever you played audio in the browser without using Flash on anything else than iPhone, chances that it was Vorbis were very high for a good while. YouTube used it in WebM before switching to Opus. In fact, even AAC gained reasonable support on the Web earlier than MP3 did.
In practice, although very popular, MP3 was nowhere near universally supported until just a few years ago. I know that pretty well - I have posted patches to some projects that enabled MP3 support once its patents expired myself; I've also used to maintain websites based on HTML5 audio since 2009.
Non-patent-encumbered and standard are not the same thing at all. Platforms do not want to ship support for esoteric formats, and will fight back against adding anything not needed for long-term interoperability - for examples, see JPEG XL drama lately in Chrome.
This is for maintainability/security surface reasons as well as patent risks, as open/closed and standardized/bespoke axes have nothing to do with whether or not something is patent encumbered.
I haven't seen anyone re-evaluate it since Safari added Web Extensions on Mac and iOS a few years back.
Most likely not for same reasons ad blockers were freaked about Chrome's Manifest v3 push - browsers are trying to optimize away the latency from a massive synchronous javascript-based list check on page load, and the privacy risk that comes from these extensions having exposure to every page (and injecting their code into every page). Conversely, the web extension authors don't see the set of limitations as feasible.
But it is odd that uBlock Origin doesn't seem to have even issued a public-facing statement of even evaluating the functionality that is there in Safari.
No. Apple doesn’t let plugins hook in so much because the plug-in ends up seeing a ton of data about what the user browses, and they can slow the browser down if badly written.
Apple lets plugins provide lists of elements/css/IPs/etc to block. Safari processes them and is able to block stuff based on that extremely fast and power efficiently.
There is flexibility lost. Plug-ins can’t see which rules are/aren’t hitting. You can’t filter based on the content of a request.
So there are ad blockers, and they work well. But uBlock, as it works elsewhere (I know it’s considered the best), isn’t possible due to the trade offs Apple chose.
This isn’t true anymore, there are numerous ad blocking plugins that do the extensive blocking just like uBlock. May not be as good due to less active community, but the capabilities are there and a few of them are quite good.
Nope. Not at all. I think they’re doing great. Most of the stuff I see complain about here on HN are features I’m not sure should exist (web push notifications, hardware access), uBlock (there are other options), and some PWA stuff they’re doing but I do t think there is anywhere near the call for from users some developers think.
If Safari was as bad as so many claim it would have next to no desktop market share. But despite Google pushing Chrome at every opportunity tons of people like my self prefer Safari.
Apple has different priorities for Safari than Google does for Chrome. That’s fine. My priorities match Safari far better, I’m perfectly happy with how they’re doing things.
This is something that lots of people complain about but somehow I never experience. Not to doubt it—I’ve seen the complaint enough that I believe it—mostly I’m just confused as to how I’ve managed to dodge the problem.
Firefox and mobile Safari, so I guess I should experience it…
Hmm, many google web features not-so-subtly nudge you to install Chrome. Some outright block usage of anything other than Chrome. This has been going on for well more than a decade.
I regularly see small/medium websites which state they only work with Chrome, but I feel they do so at their peril.
I see some of those sites push people to install a "desktop app" if you do not use Chrome, which is of course an Electron-based app.
I also regularly see services that just fail for long runs of time on non-Chrome browsers due to complete lack of regression testing, or (slightly more generously) because they aren't testing their services against current releases of Firefox/Safari. Safari is more sensitive to this, both because of a much more active development tick compared to Firefox, and because it is leveraging system frameworks rather than a relatively static compatibility layer 'buffer'.
That's how nearly everything that is now a web standard came about. The web standards groups generally don't want to even consider something for standardization until someone has actually implemented it and deployed it.
It doesn't require them to be inflicted on all of the internet. And it most definitely doesn't require you to go ahead and redirect a huge chunk of ad income to yourself while "processing" the standard.
This is somewhere between anti-consumer and stealing.
> It doesn't require them to be inflicted on all of the internet. And it most definitely doesn't require you to go ahead and redirect a huge chunk of ad income to yourself while "processing" the standard.
I can't quite figure out who or what in particular you are talking about.
If the original topic, Apple/Safari do not charge for private click measurement, and Apple does not have any significant advertising platform for the web.
There are other more controversial uses of the technique - mobile game SDKs used tracking identifiers and other techniques to try to measure conversion and associate it to a persona, which Apple shut down with ATT. Apple added a PCM-like technology at about the same time so that advertisers could get broad metrics on advertising programs and conversion.
This is different/separate from more controversial uses, such as blocking in-app advertising/install conversion metrics and persona building. There, Apple _does_ have some competing interest in terms of the App Store ad platform.
No they wouldn't. I can't think of a single thing in any browser that was implemented after a standard was created. It's always been driven by one browser just doing a thing, then other browsers do it slightly differently, then the standards body comes together and they settle on the-one-true-way and everyone updates their support to match the standard.
It very much used to work like this, pretty much exclusively.
More recently, though (especially, the last couple of years), browser vendors work very closely with standards groups, contributing there, and looking for feedback from other browser vendors. At least in the CSS and JS space, the extensions to those standards have proceeded largely as a group effort rather than as you described.
There's pretty much always implementations, but it is a huge headache when people rely on behavior which is not yet stable.
The browsers have started to first ship things behind feature flags, and in Chrome's case also behind "Origin trials". We just have less need now for one browser to go off and define their own way for borders to be drawn as rounded rectangles.
I suspect we'll see some funky CSS extensions ship pre-standardization for VR headsets, though. Things like controlling the z-axis height of <dialog> and other elements.
I don't understand what's preventing me from manually removing these URL parameters as well, just like I currently do with UTM params when I copy/paste something into a chat app.
It's about clicks instead of sharing it. So it'll strip it when you click it, instead of copying the link, pasting it, then stripping and hitting enter. Workflow optimization basically :)
Also, more privacy by default seems like a good thing, not everyone understands URLs.
Not saying it's practical but you could add your own CA on each client device and the router MitMs.
Or, e.g. you can set a flag when building Firefox that will store the secrets necessary to decrypt those packets, and the client sends the secrets to the router which sniffs and decrypts on the fly.
You can, but these sorts of setups have historically been bad about evaluating the upstream certicate/CA chain for validity, and for things like proper certificate transparency.
This might be something a web extension could do, though.
That would be neat. I suspect the browser and/or OS would have to be aware of it though, in order to cooperate, in which case why not just have the browser/OS implement it?
It's a convenience feature. Manually cutting parameters out of a huge URL is a pain, and this feature might help to remove that pain. Nothing stops users from continuing to do it manually when they cut and paste URLs.
This is a very important revelation for people to have: the deal with Apple is they have complete control over your identity and data. It's slightly better than the deal with Google, FB, & Microsoft where they both control and sell your data to the highest bidder.
I still think Apple is doing the best in the marketplace with respect to security and privacy, but if we're being honest they're playing the role of benevolent dictator.
The thing is, the terms-of-service they give you that you agree to. That thing everyone skips. In it, Apple specifically says they don't track you or sell your data (but as you say, that could change). This is why when they do have any breach of that agreement...like when they said that some humans listen to Siri requests to make sure it's being accurate, they were sued for it. People hold their feet to the fire over anything they may flub. And since Apple doubles down on saying they're the best at privacy, more and more people are chomping at the bit to sue them or call them out on it. They have to tread carefully.
Google and Microsoft on the other hand blatantly say "yeah, we look over your shoulder at everything you do on the Internet...you know, to "help" you find what you're looking for or to feed more of it to you. And also, our advertisers would be very interested too". I mean, read THEIR TOS and marvel at it.
The amount of false information in this message is staggering. Apple never said they won't track you. Apple specifically says they collect this data[1]:
> Usage Data. Data about your activity on and use of our offerings, such as app launches within our services, including browsing history; search history; product interaction; crash data, performance and other diagnostic data; and other usage data.
> Health Information. Data relating to the health status of an individual, including data related to one’s physical or mental health or condition.
> Financial Information. Details including salary, income, and assets information where collected, and information related to Apple-branded financial offerings
Also, Siri hadn't been sued for breaking TOS AFAIK, but they were sued for leaking health records which they couldn't even after getting permission from the user. Also google selling user data is largely a myth.
> Apple never said they won't track you. Apple specifically says they collect this data[1]:
Indeed. Apple's concerns are about "third-party" tracking, e.g. multiple parties sharing information about you and building a persona without your awareness or consent.
They fully _expect_ users to build first-party relationships, such as having Apple understand, save, and perhaps provide insights or fraud detection around the transaction log of your usage of your Apple Card.
This is why Google's reaction to App Tracking Transparency was to effectively say "we'll be fine, we have lots of services, and we might push people to log in for features." This is also why Facebook's reaction was to freak out - because they had no legitimate relationships on which to base their web surveillance advertising product on, especially when it was being used to select advertisements for non-Facebook-users.
For first party relationships, the App Store has "privacy labels", to document the data you collect, save and share with processors. Behavioral information is around sharing correlating factors that would be used for third-party tracking.
Of all of them isn't Google the only one that is actually incentivized to keep the data they get on you to themselves, because of their business model? It sounds bad for Google when they sell their data to others, so they'd feed a competitor for personalized ads. I might be wrong tho.
It's obvious, but I want to make clear, that this doesn't make Google less scary or more trustworty. Avoiding Google is still advised imo.
Right as I understand it at least. Google's business model has never been to sell your data, despite how persistent that idea seems to be. Your data is Google's most valuable resource and they should be extremely motivated to protect it.
Google's primary interest is in building a relationship with you so it understands the kinds of information you want, and can sell people on preferential order of getting information to you.
They are primarily an information marketplace, and offer people free services so that they'll participate in that market.
Their secondary interest is in things like Adsense and in products like Doubleclick, which are where you worry about tracking and them building a behavioral persona of what you do across the entire internet.
However, Google doesn't need to even see this persona themselves - they just want to get the best ads in front of you to make marketing departments happy. Ideas like "privacy sandbox" are partially driven by this desire - give Google the value of gathering and correlating existing data and more, without the brand impact or risks of holding that data themselves.
The thing they give up is the ability to do cluster analysis across people/demographics to understand where and why an ad is performing. Targeting of advertisements remains a manual process and not a machine learning driven one.
Can you clarify “don’t track you”? Apple charges its advertisers for ad clicks and reports the conversion rate, so whether you make a purchase after clicking one of their ads is definitely tracked (conversion is the main big money tracking data point that Facebook and Google care about).
Reporting a conversion rate is not the same as reporting that _you_ as a particular user converted.
For advertising to remain a viable way to support portions of the web, advertisers need to understand the effectiveness of their programs and marketplaces need to understand the exposure.
Specifications like PCM are competing mostly around how to provide that, although there are also efforts to sandbox local personas to make ad presentation decisions on-device. Personally I'm not so much against those efforts, as it doesn't seem like a betrayal for my web browser to try to select information it thinks I'd more likely care about - as long as I can also choose how/if it does so.
While not technically correct "sell your data to the highest bidder" is close enough to what Google and FB are doing, and the distinction is irrelevant for most people.
Apple already does the same thing. Apple Search Ads is not limited to the same restrictions that Facebook and Google are with regards to iOS tracking and reporting for advertising attribution.
Remember that Apple's "debacle" there was to comply with US law. Their issue was that (at the time) they were responsible for encrypted data backup's contents. So they could either scan on your device before backing up, or scan on the cloud. Scanning on your device, while it sounds scarier, actually offered more privacy protections, because otherwise their cloud needed to see your unencrypted data. And it only scanned on your device if you wanted to move things to the cloud.
This is false. Apple already scans everything on iCloud for banned material serverside, as iCloud Photos and iMessage are, for most people, not e2ee, and never will be so long as e2ee is opt-in.
Even if you enable the e2ee features rolled out in the
last 12 months, your iMessages are still not e2ee unless all of your conversation partners have as well.
Also there is no US law demanding scanning of user data, your opening assertion refers to nonexistent requirements.
* iCloud Photo Library was not scanned for CSAM content at the time of the announcement, which Apple confirmed at the time.
* iMessage E2E encryption is not opt-in. There isn't even an opt-out.
* The "E2EE features" you might be referring to is Advanced Data Protection for iCloud Backups, which is not related to the iMessage protocol at all. You don't have any guarantees about what your recipients are doing with the data you send them, ever.
There is US law which is ambiguous about the requirement on data providers to check content for CSAM material, which many have interpreted to require a check. This is why every other major cloud provider does scan for the content.
> iMessage E2E encryption is not opt-in. There isn't even an opt-out.
This is incorrect. iCloud Backup escrows endpoint keys for "Messages in iCloud" to Apple every night in a non-e2ee fashion, which means that a non-endpoint has the keys, which means that iMessage is not e2ee.
Apple has real-time access to plaintext of almost every single iMessage that transits their service. The only case in which they don't is where both users either don't have iCloud Backup enabled or both users have enabled e2ee iCloud Backup.
The PSI system was pretty cool in my opinion. It was a very neat algorithm for obtaining information about set intersection in a privacy preserving way.
This is completely wrong. They are saying they only don't strip PCM parameters because these are anonymous and somewhat privacy preserving. Apple is still uninvolved in the link attribution or other tracking here.
As a consumer. I don't care. I'd rather Apple be the only ones with my data. Advertisers handed Apple this power by spending the last decade being as abusive as technically possible. I hope Apple shuts down the entire industry.
This was the same attitude that led Google where it is now. "Don't be evil" was really believed and accepted by huge part of their users, and compared to the sheer evil of Microsoft they sure looked like the better alternative.
> I hope Apple shuts down the entire industry.
On Apple's relationship with the ad industry, I have bad news for you.
> ... at end of the day they're simply consolidating all tracking power to themselves.
1. I'm still learning about 'private click measurement'. Does it reveal extra information to Apple? I would hope it is designed as a truthworthy protocol where cheating, even by Apple, is hard.
2. Welcome to another arms race: detecting and removing identifiers versus hiding them or using alternative mechanisms.
If Apple is taking on this challenge, do they have theoretical reasons to be optimistic? Practically, how well are they doing?
Yes they are. And they know it. Apple knows it. Microsoft, Google and Amazon all know it.
Privacy and security is the most scalable, reliable way to locking users in and stifling competition behind a lot of marketing about "protecting" users.
Anything an advertiser can track, they will use where they can to build a profile of you to identify you uniquely for targeting. Apple allowing ANY tracking measurements through incognito that isn't already naturally happening by the nature of incognito is too much.
Here is my take: Apple is being pro-privacy to lock-in consumers to their products
But they are nowhere as untouchable as they seem to think they are wrt to anti-trust laws. So, let them, in their arrogance, destroy surveillance-capitalism and get destroyed themselves for being uncompetitive. Let the beasts of greed eat each other.
While as a consumer I do objectively like the privacy measures Apple is adding, at end of the day they're simply consolidating all tracking power to themselves.