Thread by @thedextriarchy

Day 14 of Epic (but not a fortnight, because weekends) starts in five minutes. It’s all Apple expert witnesses: Dominique Hanssens, James Malackowski, and Aviel Rubin. Yesterday below:

https://twitter.com/thedextriarchy/status/1395029579484606464

To be perfectly blunt Epic’s expert witness portion was a little slow, but if we make it through today we get Tim Cook tomorrow.

Epic’s lawyer says their side might call some rebuttal witnesses in response to Apple’s testimony today, but it’s up in the air. They may also call a new witness to respond to “testimony that has only just come up."

Apple says it’s too late to call surprise witnesses. Epic says “this is an unanticipated event,” it would only take 10-15 minutes, and it’s “still TBD.” Judge notes there’s nothing to rule on yet, so we’re moving on.

Apple is calling Hanssens to the stand. Hanssens’ work has come up before — he surveyed iPhone/iPad users to see if they regularly used other devices that could access Fortnite.

A binder *and* a book are being referenced! I cannot see them.

Hanssens is a research professor at UCLA, focuses on researching marketing effectiveness.

Hanssens outlining his research: he conducted a survey of iOS device users 13+, and he surveyed the survey work done by Peter Rossi, one of Epic’s expert witnesses. (Who surveys the surveyors? Hanssens, apparently.)

The 13+ number (as opposed to 18+) is important because Fortnite is recommended for that age, and a major demographic is people under 18, says Hanssens.

Hanssens says he used “standard procedures” to obtain parental consent for < 18s. One of Epic’s expert witnesses (Rossi IIRC) contended that this wasn’t feasible, and Hanssens is undercutting that.

Same goes for Hanssens collecting survey data over a 12-month period — Apple’s lawyers argued that Epic’s antitrust analysis relied on a month-long study during a period right after the holidays where people had very specific expectations about pricing, and that skewed the study.

Hanssens’ survey assessed how much iOS users relied on non-iOS devices on a regular basis, and what devices they had “available” but didn’t necessarily use, like devices family members had.

To be clear his survey wasn’t just 13-19, it just *included* younger users than Epic’s surveys did.

https://twitter.com/90one80/status/1395398874718683141

Hanssens also conducted a survey of the behavior of iOS Fortnite players.

The first survey found 92% of iOS respondents regularly used one non-iOS device that could access digital gaming content too, 99% “could have” used one.

Hanssens checked robustness of results in a variety of ways, including throwing out anybody who said they regularly used a Windows phone or had access to one, since a “surprising” number of people claimed to do this for a phone platform that virtually no longer exists.

(Apparently 13% of iOS-using respondents said they had regularly used a smartphone running a Microsoft OS in the past year, and 30% said it was “available” to be regularly used. Hanssens finds this improbable.)

Hanssens says that Rossi, Epic’s witness, missed important elements of survey methodology, including interviewing respondents.

(Rossi’s survey, to recap, was examining whether iOS users would change their behavior based on a hypothetical price increase. His respondents mostly said no, and Epic used it to argue that Apple wasn’t subject to meaningful market pressure.)

Hanssens says he takes issue with Rossi setting his hypothetical as Apple *having* raised prices in the past. “When a consumer decides to react or not to a price change, they take the risk of a price increase … of not being satisfied with a product…"

“…if you move the price back 30 days, you already know what your experience has been, because you haave purchased that asubscription or that partilar app.” So two elements are conflated: price reaction and satisfaction with the product.

If people’s experiences with apps are pretty good, they’re going to be satisfied and more likely to say the increase was okay — making them more likely to be “stickers” who don’t change their behavior based on price increases.

Hanssens also says Rossi obtained biased results by phrasing the “sticking” questions as a “yes” rather than a “no,” since people are more likely to acquiesce in a survey.

Epic’s Lauren Moskowitz examining Hanssens now. Starts by asking if he considers himself an expert on “general survey design” — he says that’s correct. He can’t recall if he’s testified in an antitrust case, and he’s not positive he conducted a smartphone-related survey before.

Epic’s lawyer basically asking questions to establish that Hanssens doesn’t have the expertise to analyze what constitutes a potential monopoly market.

Hanssen’s survey didn’t specifically address the issue of “substitutability” — not just whether someone has access to a device, but whether it’s a meaningful substitute.

Hanssens also says his 12-month period doesn’t measure whether somebody’s talking about a phone they later got rid of (i.e., if they switched to iOS in the past 12 months.)

(Epic is also contending that switching costs between ecosystems are very high, so if lots of people are switching between iOS/Android that’s not necessarily a huge point in its favor either.)

Hanssens asked in the Fortnite survey if people specifically used the devices for gaming, but otherwise, he acknowledges he didn’t measure what people used each of their devices for, and therefore if those things were meaningful substitutes for each other.)

“You just didn’t care what respondents were using their devices for,” says Epic’s lawyer, just whether the devices were “available.”

Hanssens says yes, in the sense that this is what he was hired to do.

Judge says Epic’s lawyer is talking too fast, asks her to slow down. Lawyer keeps going through the details of what device usage really means in the survey.

Epic is still going through numbers, the idea being that Apple used Hanssens’ survey to argue that virtually all iOS users were regularly accessing other devices, and in reality its 12-month window didn’t account for important questions like whether users still had those devices.

Epic’s lawyer also objects that “you insist that we assume” that people who regularly use a game console use it to play games. They might just use it to watch Netflix, for instance. The survey doesn’t ask for these details.

I want to make fun of “do people *really* play games on their game consoles” but my PS4 is basically a streaming device for months out of the year.

Apple started the morning by focusing on how Epic’s expert made a bunch of structural errors in survey design. So far Epic is focusing on Hanssens’ design being unacceptably vague.

If you've ever wondered what the word “regularly” means, this trial is for you.

Lawyer confirms Hanssens didn’t do any interviews or focus groups to understand how people think or talk about issues in the survey. “You did not explore before dropping your survey whether people think about devices they use in terms of the operating system, for example."

“You think people understand the difference between a computer, a smartphone, and a tablet,” but even Hanssens didn’t, lawyer says. He couldn’t decide whether a Microsoft Surface was a tablet or a computer, apparently. (He says he determined it’s both. Which seems fair?)

Epic’s lawyer now drilling down on the surprisingly high reporting of Windows phones. “You know that Microsoft Windows smartphones are no longer being sold, right?”

Hanssens indicates he knows this.

(Some surveys use made-up brands or products to test if people are paying attention, but it appears this was not the case with the Windows Phone question.)

Lawyer says that throwing out people who reported Windows Phone use resulted in removing 151 of 500 respondents (in one specific robustness check, not the whole survey) and even more in the Fortnite survey.

For context the robustness check was designed to measure whether throwing out one potentially dodgy group dramatically changed the results, and Hanssens' conclusion was that it didn’t.

Questioning goes back to Apple. Hanssens’ clarifies (at lawyer’s prompting) that he isn’t objecting to all hypothetical surveys, he just believes Rossi missed important steps.

We are back onto the Microsoft OS smartphone question, and Apple’s lawyer is suggesting that maybe people have a newer Microsoft phone even if it doesn’t technically run a Microsoft OS. I'd guess she’s talking about the Android-powered Surface Duo.

I’m not a survey expert but I feel safe concluding that 13% of iOS users are not regularly using that.

In any case Hanssens said what I did a few tweets up — the point of the check is that removing alleged Microsoft Phone users didn’t change the percentages either way.

Hanssen recasts what Epic’s lawyer calls vagueness as an attempt to meet people where they are, basically — instead of trying to preemptively decide for them what terms like “regular use” mean.

Apple tries to raise the issue Epic did about substitution, but lawyer has trouble pointing to a part of Hanssens’ testimony where he discusses it.

“If he doesn’t use the word ‘substitution,’ it is what it is,” judge says, and they move on.

Epic’s lawyer appears to be now saying that because a lot of survey respondents were probably confused about the Windows Phone thing, Hanssens' should have done something to control whether everyone else was confused about the survey overall.

With that, we’re done, and expert witness James Malackowski is coming up.

Malackowski is CEO of Ocean Tomo LLC, a financial firm.

Malackowski works on intellectual property valuation, and he’s here to bolster Apple’s claim that its IP rights are valuable and Epic is seeking to free-ride on them.

“Academically, I enjoy teaching.” This is (unironically!) a good thing for an academic to enjoy.

Malackowski is discussing that it’s important for IP owners to have the right to determine how their intellectual property is used. If you lose control over rights, you don’t know if you’ll get a return on the investment you’ve put into developing the IP.

Malackowski says free-riding on IP is harmful to innovation. “If that were permitted, the inventor can’t reasonably predict a return on investmetnt,” so it can’t reasonably judge how to make that investment.

He says he “searched high and low” through Epic’s case documents, but he found nothing about this intellectual property concern.

He was hired to assess the “innovation footprint” of iOS — research and development investment and commitment, the IP that results, and the use of that IP by Apple or others who license it, including Epic Games.

Malackowski’s written expert testimony includes the conclusion that Epic is seeking "essentially a compulsory license to all of the IP necessary to distribute apps to iOS users."

Apple spent $500 million on R&D in 2005, and $18 billion in R&D in 2020, per public numbers. Chart shows a linear rise:

Malackowski also looks at the steady rise of patents applied for and granted — wants to note that Apple has a “significant and sustained commitment” to innovation based on these two charts.

He says Ocean Tomo reviewed thousands of patents to make sure Apple wasn’t just collecting patents — when someone comes up with an idea and they patent it “because it’s cool,” as opposed to because they plan to use it.

Apple lawyer asks if Malackowski is familiar with open source software — you might remember Epic counsel citing Apple’s use of open source software (in arguably one of the trial’s sketchier bits) as an argument that it wasn’t innovating as much as it claimed.

Malackowski says he’s focusing on proprietary tech.

We’re going to break for 20 minutes, back at 1:35ET.

We’re back with Malackowski on the stand.

Malackowski is running through a patent that he says is relevant to the iOS ecosystem question — it’s a patent relating to the Metal graphics API (incidentally the thing Epic once got on stage with Apple to talk about being awesome).

Without Metal, 3D graphics like Fortnite’s would be a much bigger drain on mobile device batteries, Malackowski says.

Another patent related to UIKit, designed to help developers create iOS user interfaces.

Malackowski being asked about the relationship between patents, copyrights, trademarks, and trade secrets. Says they’re distinct, but they can interact — so you have a specific innovation like Metal, he says, where all four come into play.

Malackowski clarifies that he had people manually go through “metal”-related patents to make sure that they related to Metal the software tool and not metal the material.

unclear whether some patents may relate to the gary numan song

Malackowski also says there’s evidence developers are benefiting from using APIs, because app development has increased steadily and the quality has also increased thanks to things like CoreML.

Says there’s also “consumer quality benefit” around security and privacy, citing earlier testimony in the trial supporting that.

Again Malackowski’s saying there’s no discontinuity between the rising lines between investment, market growth, and patent filings — implication is Apple invests money, it files patent at rates in line with its level of investment, and its business grows apace.

Epic has listed which APIs it uses in other court filings, and it’s talked publicly about how it uses the APIs — so there’s no question it relies on these technologies, Malackowski says.

We’re looking at the Apple developer agreement, which Malackowski describes as an agreement by which developers are allowed to use Apple’s IP. There’s no fee to signing the license, he says. (There is a separate fee for putting an app on the App Store.)

Malackowski says the license for the latter paid developer agreement is “limited” — it’s contingent on agreeing to go through App Review, among other things.

He concludes that logically, Epic is asking for relief that would “take away Apple’s control or Apple’s provisions in its license agreements,” requiring things like supporting a store-within-a-store.

“It’s not only that they would lose out on the compensation,” he says; Apple would have to support these apps going foward.

“It’s ‘you can use my stuff and I have ot keep performing for you now and evermore.’ It’s quite extreme."

Epic’s Lauren Moskowitz is now doing cross-examination. She’s citing earlier cases where a court found Malackowski’s testimony was unreliable.

Lawyer says he had some testimony excluded from the Oracle v. Google trial — a judgment on future lost profits, apparently.

There are a lot of cases where Malackowski's had testimony excluded in some form! I don’t know enough about courts to know if this is an unusual amount or something you’d expect from any long-running expert witness career.

I think we’re up to 6 cases in around 50 now

https://twitter.com/leah_nylen/status/1395443469938741249

Epic’s lawyer is noting that he didn’t do a direct financial appraisal of Apple’s IP, or how Apple’s R&D compared to its total assets or revenue.

Lawyer is asking Malackowski probing questions about whether Apple would really be giving up revenue from its IP under Epic’s demands, especially compared to, say, its hardware profits.

Lawyer notes there are no patents listed in the developer agreement. So if you’re a developer and want to know which patents you’re licensing, you wouldn’t see that.

“You could find it, but it wouldn’t be in this agreement, sure,” Malackowski says.

Shouldn’t licensing agreements actually show which patents are being licensed, lawyer asks?

Malackowski agrees in theory.

Lawyer is casting doubt on the methods that Malackowski and his firm did to search for relevant patents, including the possibility that titles referenced “App Store” but the actual claims didn’t.

Also casting doubt on whether Apple is licensing *trademarks* to developers — Malackowski says sometimes yes, and that developer relations folks work with developers to coordinate when necessary.

Epic is back on talking about which patented Apple APIs use open source technology and whether Malackowski did any work to disaggregate how much of the work is Apple’s investment.

Epic’s lawyer is also bringing up the cases in which using copyrighted code can be treated as fair use. We’re getting a mention of Google v. Oracle, which hinges on this exact issue (and found in Google’s favor.)

Malackowski says he hasn’t tried to disaggregate any usages that might count as fair use.

Malackowski also doesn’t necessarily know if Apple has directly said it doesn’t allow stores-within-stores to protect its IP, as opposed to other reasons (like protecting security, etc.)

Lawyer notes that courts can conclude IP protection doesn’t outweigh anticompetitive behavior, if the IP claims are just being used as a pretext for shutting out competition.

With that, we’re going back to Apple questioning.

Malackowski says the developer agreement is a broader portfolio agreement — so people don’t have to note every single new patent Apple files and puts under it, etc.

Apple lawyer is also coming back to the cases where Malackowski’s opinion was excluded. Emphasizes that he’s testified in 50 cases, and that he testified in some of the cases Epic’s lawyer cited, just not about the specific piece of testimony that was discounted.

Now Malackowski is stepping down, and we have Aviel Rubin, the last witness listed today. Rubin is here to testify about iOS security — a counterpoint to James Mickens, whom Epic called as one of its last witnesses.

Side note on the companion case Epic filed against Google:

https://twitter.com/BobbyAllyn/status/1395461115325210624

Rubin determined that Apple’s App Store manual review offers “significant security benefits to iOS and results in lower infection rates on phones as well as a low volume of malicious and untrustworthy apps in the App Store."

Rubin is defining what he means by security. He looks as security as “dealing with an adversary. That means there’s somebody bad that wishes to cause harm.” That includes safety, privacy, reliability, and trustworthiness.

Now we’re going to break until 4:15ET.

We’re back in court, talking about the witness Epic wants to call. So apparently Epic wanted to call a current Apple witness, then a third-party witness, and Apple says it was misled and that it’s too late to call witnesses.

Epic says that Apple’s findings of fact show IAP were introduced in 2009, but a back-and-forth with Phil Schiller raised questions about in-app commerce opportunities before they technically launched.

If these opportunities existed, then introducing IAP would have increased prices, Epic’s lawyer says, so they want to introduce more information about that — either by calling a rebuttal witness or filing documents.

Apple says that this is “the first time” they’ve seen any justification, and that Epic had its chance to submit documents already. “Now is not the time for them to be attempting to kick this door open."

Judge Rogers says Epic counsel can give them a page outlining the issue and she can look it over — notes that Schiller testified three days ago. Epic’s lawyer Forrest said Epic was interviewing developers who might testify.

Forrest calls the issue of in-app purchases “critical."

“If it was so critical, you should have been prepared for it,” judge says. “Clearly you knew it was an issue on the 14th and on the 17th. It is an issue that I highlighted that I have to figure out."

No rebuttal witnesses, Judge Rogers says. “At most,” she’ll allow filing of documents.

Forrest says Schiller denied what Epic believed was an “obvious point,” and Epic was scrambling to grab documents. “It came as a surprise to me that he would deny it so vehemently.” (“It” being that people were processing payments before IAP.)

Rubin is back up now for questioning by Apple.

We’re talking about app distribution models — including centralized systems like Apple and direct distribution where developers transmit apps to users.

Phones have to be configured differently for different models, says Rubin. (Also, previous tweet should say ‘iOS’ and not ‘Apple’, obviously.)

When you have multiple app stores, says Rubin, a malicious developer can submit to all those stores and hope to get into one.

Malicious devs can also post a “good” version of their app in an app store that has good review, they wait for people to adopt it, then they post the same app with malware on a store that won’t do the same review, coasting on their established reputation.

Devs can also create imposter apps that impersonate better-known apps in more trustworthy stores, or maybe even impersonate a whole store.

“The weakest link is really the problem here,” Rubin says, if there are many places to get software.

“In a centralized app store, you have an opportunity to do as good a job as possible and not worry that another app store is going to cause a security problem,” Rubin says.

Craig Federighi said a lot of what Rubin did yesterday, but Rubin is talking a bit more about multiple app stores specifically rather than just the benefits of a centralized model.

Like Federighi, Rubin is bringing up a Nokia security report — Apple’s lawyer briefly stops to let Rubin explain that Nokia isn’t just a consumer phone company but sells infrastructure security tools.

We’re going into quite a bit of detail about Nokia’s security and analysis techniques. I am a little impatient here because look it’s been a long week, I don’t need to know this much about malware honeypot methodology.

Judge asks whether there are just more Android phones and that’s producing numbers like this.

https://twitter.com/mslopatto/status/1395482932819623936

“The truth of the matter is, in the studies I read, the indications is globally there’s about 3x as many Android devices as iOS devices, but here we see the ratio is much bigger for infections on Android,” Rubin says.

Judge notes an Internet of Things chunk of the pie and asks what it is. He says things like smart home devices, and he notes they’re often very poorly secured and updated.

Now we’re looking at a RiskIQ study about threats in the mobile landscape — press release about the latest version of the study here riskiq.com/press-release/…

Rubin’s reading out this paragraph on “blacklisted” apps now, defining phishing.

Apple didn’t make the list of stores that allowed “blacklisted apps” — the researchers considered it but there weren’t enough apps to register.

The report says that "Apple treats its App Store like Fort Knox and rarely hosts dangerous
apps.” (Contrast with Epic’s repeated claims that lots of offending content can get through to the store.)

Rubin says concentration of dangerous apps is more important than just numbers of bad apps on the store. "If you were to throw a dart at the app store, what are the odds that you’ll hit a blacklisted app?"

“How can I be sure it’s not biased?” Judge Rogers asks of RiskIQ. Rubin says he thinks it’s a pretty widely trusted source.

Judge asks if it gets funding from Apple; Rubin says he doesn’t think so, although he doesn’t know for sure.

Now he’s citing a Purplesec report available here: purplesec.us/resources/cybe…

The report says that "98% of cyber attacks rely on social engineering,” i.e. tricks that exploit psychology to make users give access to protected info. Rubin cites a piece of Android malware that pretended to be a system update and got people to give it permissions that way.

This security testimony is addressing the most extreme Epic request — that Apple allow sideloaded apps. Curious if we’ll get any discussion of whether developers independently processing in-app purchases is safe — Apple’s gestured at dangers but hasn’t really delved in.

Rubin notes that some bad stores have “adult content,” and Apple’s lawyer asks if adult content has security implications — Rubin says there’s a correlation where malware is more likely to be present in stores with adult content.

Rubin being questioned by Epic’s Brent Byars now.

Epic’s Byers is talking about the importance of sandboxing and running down reasons Rubin argues Android is insecure, talking about how they don’t necessarily apply to iOS — like Android fragmentation or external storage.

Epic’s lawyer asks (rhetorically) whether Epic is asking Apple to license iOS to other phone manufacturers.

“I have been very confused about what Epic is asking for,” Rubin says.

Epic’s strategy here is essentially pointing out differences that could make Android more vulnerable beyond the simple fact that it allows sideloading — the things I just mentioned, as well as, say, Android’s authorized app store allowing more malware through.

Epic’s lawyer is bringing up Shlayer, the first Mac malware known to be accidentally notarized by Apple, found last year techcrunch.com/2020/08/31/app…

We’re reading from this wired article now: wired.com/story/apple-ap…

Epic’s lawyer asks if this article describes how notarization works at Apple — Rubin says he thinks it’s more about how recovery when notarization goes wrong works.

And now, questions about whether people use the Mac in sensitive moments.

“I think that it’s pretty clear that your phone is with you in more sensitive times than your Mac is,” Rubin says.

But he agrees people probably use Macs at work, in the living room, and in the bedroom.

Rubin’s now being asked about whether other stores could implement a manual review similar to Apple’s — Rubin quips that maybe yes, if you “cloned Apple."

Epic’s lawyer says he’s still got 20-25 minutes of questioning Rubin, but we’re breaking for the day, and we’ll pick up first with Tim Cook tomorrow morning.

We’re discussing a few last procedural things.

We’re still debating whether Epic will get to add more testimony about in-app purchases, but Judge Rogers shuts it down. “Everybody knew what was going on, and the fairest thing is to leave the playing field as the playing field."

Epic will still get to ask questions of Tim Cook, though.

And now we’re officially in recess till tomorrow morning — the last day of testimony.