WalliD

wallid

PROMISES FULFILLED.

The Walled City

EPIC GAMES, INC., Plaintiff, v. APPLE INC., Defendant Court Filing, is part of . You can jump to any part in this filing . his is part 2 of 11. Aug 24 2020 HackerNoon’s Legal PDF Series here Feature Image: , Prompt “landlord looking happy in his walled garden” HackerNoon’s Stable Diffusion AI I. BACKGROUND The Court summarizes the limited record before it on this motion for preliminary injunction. To expedite issuance of this Order, the Court cites to some of the basic background from its prior order====[2]==== as the background is equally relevant here. The Court notes disputes in the evidence where otherwise appropriate. A. The Players With respect to Epic Games: Epic Games is a United States-based tech-company that specializes in video games, including, as relevant here, the popular multi-platform==[ ]== game, is structured around "seasons," whereby narratives, themes, and events are introduced for a limited time. Cross-platform play is enabled for all users so long as those users remain on the same version of the game.... Epic Games International, S.a.r.l ("Epic International") is a related company based in Switzerland and hosts, among others, the Unreal Engine. The Unreal Engine is a graphics engine created by Epic International to assist in its development of video games that it later began licensing to other developers. The Unreal Engine 4, the current version of the engine on the market, is used by third-party developers for the development of video games for both console and mobile platforms, including for games currently offered in the iPhone App Store. These third parties range from smaller game developers to larger corporations, such as Microsoft Corporation. The Unreal Engine has also been used by third parties for architecture projects, film and television production, and medical training. 3 Fortnite. Fortnite Epic Games has released twenty-five (25) updates to Unreal Engine since 2014, and anticipates releasing future updates to ensure that Unreal Engine remains compatible with new versions of Apple's software, such as the now released iOS 14. Developers can use Unreal Engine commercially on a royalty model or negotiated license, but it is otherwise free for non-commercial use. Although more applications on the iOS platform are powered by a rival game engine, Unity, a significant number of iOS applications are constructed based on Unreal Engine, including competitor (" "). Fortnite PlayerUnknown's Battlegrounds PUBG Epic Games also maintains or controls other affiliates including: Epic International, Life on Air, Inc. (both in San Francisco, California and Austin, Texas), KA-RA S.a.r.l., Psyonix LLC, and Quixel AB (collectively, "Epic Affiliates"). The Epic Affiliates maintain control over certain applications and software within the Epic Games business. These identified applications include: Unreal Engine, Unreal Remote 2, Unreal Match 3, Action RPG Game Sample, Unreal Remote, Live Link Face, and House Party, among others. Meanwhile, Epic Games itself controls , Infinity Blade Stickers, , and, as relevant here, Battle Breakers Spyjinx Fortnite. Beyond these games and applications, Epic Games also operates a digital marketplace to sell game software called the . As pled in the operative complaint: the Epic Games Store was created to compete against the leading multi-publisher digital video game marketplace on computer platforms, Steam, which is operated by Valve Corporation. The Epic Games Store provides access to more than 250 games from more than 200 developers. Like other video game digital distribution platforms, the Epic Games Store offers personalized features, including friends list management and game matchmaking services. As alleged, absent Apple's alleged anti-competitive conduct, Epic Games would also create an analogous Epic Games Store for the iOS platform independent of Apple's digital marketplace. Epic Games Store With respect to Apple: Apple is a ubiquitous tech-company that makes products ranging from hardware to software. Apple, as relevant here, maintains an App Store for the iOS platform that is geared for its mobile devices, the iPhones [and iPads]. The App Store allows third-party developers an opportunity to create and thereafter sell applications to iPhone [and iPad] users. Apple's agreements with developers and the App Store guidelines do not generally permit third-party developers to circumvent the IAP system. Apple generally takes 30% of the sale of the application or of the IAP made within the third-party application itself. In addition to preventing developers from circumventing the IAP system, developers are also prohibited from distributing applications outside of the App Store on the iOS platform==[ ]==. 4 In short: . In order to access the App Store and to obtain developer tools, developers are required to comply with Apple's rules and regulations through a web of agreements and guidelines: Apple maintains the iOS platform as a walled garden or closed platform model, whereby Apple has strict and exclusive control over the hardware, the operating system, the digital distribution, and the IAP system As relevant here, Apple maintains separate developer agreements and developer program licensing agreements between Epic Games, Epic International and four other affiliated entities. Apple also maintains a separate agreement, "Xcode and Apple SDKs Agreement," regarding its developer tools (software development kits, or "SDKs"). , 2020 WL 5073937 at *1 (Dkt. No. 48 at 2-3). These agreements have broad language including terminable at-will clauses. Epic Games [5] The relationship between Epic Games and Apple dates to at least 2011, when Epic Games released its first game on the iOS platform. Epic Games and Apple collaborated for several Apple events, showcasing Epic Games’ iOS games and the earlier iterations of the Unreal Engine running on the iOS and macOS platforms. Following the success of on other video game platforms, Epic Games launched on iOS in April 2018, where it remained on the platform until, as discussed below, August 13, 2020. Infinity Blade Fortnite Fortnite During this time period: (i) 116 million iOS device users accessed , spending more than 2.86 billion hours in the game; Fortnite (ii) the daily average users numbered approximately 2.5 million daily iOS players, representing nearly 10% of ’s total average daily players; and Fortnite (iii) 63% of iOS players on have only accessed from an iOS device. Fortnite Fortnite Finally, iOS users accounted for more IAPs within than those on the Android platform, but iOS users spend less on IAPs than those on the console platforms, including the Sony PlayStation 4 and Microsoft Xbox One. Fortnite Continue Reading . Here [ ]See Epic Games, Inc. v. Apple Inc. , 4:20-cv-05640-YGR, 2020 WL 5073937 (N.D. Cal. Aug. 24, 2020) (Dkt. No. 48). 2 [ ]These platforms include Android, iOS, macOS, Windows, Sony PlayStation 4, Microsoft Xbox One, Nintendo Switch. Fortnite is also available for download through the Epic Games Store, as discussed herein. 3 [ ]For purposes of this motion, the parties refer to the operating system for both iPhones and iPads as iOS. (See Opp'n at 4, n.2 (Dkt. No. 73 at 10).) Moreover, Epic Games pleads that there are no differences between iOS and iPadOS to the allegations in the complaint. (Compl. ¶ 39, n.1 (Dkt. No. 1).) Similarly, this Order refers to iOS to refer to both the iPhone and iPad platforms, and references to iPhones generally also apply to iPads. 4 [ ]The record also contains two enterprise account agreements for Epic Games and YEVVO Entertainment, Inc. The parties do not otherwise discuss the significance of these agreements. 5 About We bring you the most important technical and insightful public domain court case filings. HackerNoon Legal PDF Series: This court case 4:20-cv-05640-YGR on Oct 9, 2020, is part of the public domain. The court-created documents are works of the federal government, and under copyright law, are automatically placed in the public domain and may be shared without legal restriction. published

First thing first: Apple takes 30% of all sales made by developers via its walled garden app store

Once upon a time, the personal computer stored the user data. Then, the Web came along. What made the Web a mighty force was that it was free. Anyone could access and build applications for the Web. In those days, the Web was . Nobody, not one person, could say they own, control, or switch off the Web. decentralized Then, smartphones came along. Software companies started to have access to users’ data. User data became a currency, so it made sense for software companies to make it as difficult as possible for users to migrate their data to other services. In short, this is the concept of vendor lock-in. Vendor lock-in builds dependency: once you start using a vendor’s services, it becomes harder to move to a concurrent service. For example, if all your family are is on Facebook, you are locked into Facebook. If your family only uses WhatsApp, you are locked into WhatsApp. If your company uses Microsoft Office, you are locked into Microsoft Office. The result is that we moved from the decentralized Web to the centralized Web, where applications own, control, switch off, surveil or censor their ecosystem. . We also moved from the free Web to the “free” Web Unfortunately, in English, there is no distinction between “for free” and “liberty”. Thus, the software community adopted two phrases: “free as in freedom of speech” ( , free software) and “free as in free beer” ( , freeware). Free software or “libre software” (borrowing the French or Spanish word for “free” as in freedom) means that users have the freedom to distribute, copy, change the software that is not necessarily gratis. Gratis software has a price of zero. libre gratis With this distinction in mind, is using social media applications gratis? No, of course not. There is no monetary price to be paid because we pay social media services with our data, time, and attention. Every time we use the Web or mobile applications, we leave a data trail behind, ready to be used, abused, or sold to the highest bidder. Take unroll.me, an email decluttering service that unsubscribes its users from email lists. This service has access to email subscriptions, purchase receipts, travel reservations, or shipping-related emails to know which subscriptions to remove. so that Uber can gain insights about how many of their users were using their competitor, Lyft (remember, unroll.me had access to receipts, including Lyft receipts). Unfortunately, this is only one tiny fragment that we know of from the millions of other user data exploitation. Unroll.me had been selling user data to Uber Occasionally, application or cloud startup services offer their service for free. They try to sign up as many users as possible, hoping for a charge or a subscription later. Be wary when using such a service. Leading companies (Facebook, Apple, Google, Microsoft, etc.) can buy services you are using and shut them down with few chances of recovering your data. If startups go out of business, they can sell their user database to return money to their investors. Unfortunately, paying for a service does not guarantee user data will not be used or sold. As I mentioned earlier, “free” social media applications take hostage not only our data but our attention and time as well. Social media algorithms want us to stay longer and engaged in their ecosystems as these algorithms hoard and exploit additional bits and bytes of our identities in the background. That is why we see content that social media algorithms assume we will click on, share, like, or discuss. What content is guaranteed to always spread like wildfire? Appalling and surprising misinformation (defined by Nina Schick in her book as “bad information with no malicious intent behind it”) and disinformation (information meant to deceive) always gather user traction and engagement. Then, alongside click-baiting titles, our social media history (what content we interacted with in the past) considerably determines what content we will see in the future. Deep Fakes and the Infocalypse Why is that? Whenever we find out a new piece of information, we tend to judge it. If it is consistent with our world model, we are inclined to accept it. If it is inconsistent, then we are tempted to refuse it. The algorithm plays this routine repeatedly, learning about our preferences and keeping us engaged for longer and longer in a walled garden where the stress and unpleasantness of different opinions and mental models are almost non-existent. Enter virtual social bubbles and echo chambers where our biases are distorted and amplified because engaging with social media algorithms is autocatalytic (the process feeds itself). For example, social media can make you feel jealous, angry, hateful even. You start feeling bad about yourself, maybe experiencing a nagging sensation that you are not enough. So, you check more social media to find a tribe where you belong, a community where you feel at peace scattering negative messages about other tribes that made you feel conflicted about yourself. A shift is coming. The inventor of the “ ” and decentralized web, Sir Tim Berners-Lee, is currently involved with a new project that aims to change the way applications work today, . Suppose a user keeps their digital data in decentralized servers, called Pods. No organization can longer control their data, but they need to ask for the user permission to access relevant data in a Pod. Another project is , a blockchain-powered Internet in which users have complete control over their digital assets. free as in free beer The Solid Project Elastos Although social media applications were built on top of the free Web, these services didn’t repay the kindness (Wikipedia is the most significant exception). These applications gathered massive and unacceptable control over our lives as algorithms determine the size of our social bubbles, reconstructing a reality based on our preferences, likes, and biases. The Web should become re-decentralized and truly free again, where we own our data, and we control the depth and breadth of our mental models through repeated exposure to news and views from outside our echo chambers. . Previously published at https://www.roxanamurariu.com/the-perils-of-living-inside-the-walled-gardens-of-social-media/

The Walled Gardens of Social Media: From the Free Web to the “Free” Web

Comment comes as part of big tech CEO grilling by antitrustcommittee In a historic hearing questioning the heads of four of the nation’s five largest technology companies, the chair of Congress’s antitrust subcommittee, David N. Cicilline, on Wednesday directed sharp criticism at Sundar Pichai, CEO of Google’s parent company Alphabet, pointing to The Markup’s findings that Google Search places the company’s products and results above competitors’. “An investigative report published just yesterday found that 63 percent of searches that start on Google also end on Google’s own websites,” Cicilline said, referring to our findings about results on the coveted first screen. “And to me that is evidence that Google is increasingly a walled garden which keeps users on Google’s sites, even if Google doesn’t have the most relevant information. And it’s economically catastrophic for other companies online.” Cicilline mentioned the report after asking, “Increasingly what Google shows is what is most profitable for Google. Isn’t there a fundamental conflict of interest?” Pichai responded, “In a vast majority of queries on Google, we don’t show ads at all.” The hearing was the sixth in a year-long series titled Online Platforms and Market Power, which has included testimony from experts, competitors, and lower-level executives from the four tech giants, Google, Apple, Facebook, and Amazon. It was the first time all four had been questioned together, and Amazon CEO Jeff Bezos’s first appearance before Congress. They were asked about their respective market dominance and treatment of competitors. And some members of the committee asked the CEOs about their treatment of conservative media and causes. The hearing came one day after a months-long, painstaking investigation into Google Search that for the first time quantified how much the company highlights its own products and serves “answers” on the results page, many of which take information from other websites and are meant to satisfy queries without the need to leave the Google universe of products. The Markup published Our of more than 15,000 U.S. trending queries found that Google gave 41 percent of the first search results page to itself and dominated the coveted first screen. sophisticated analysis Google spokesperson Lara Levin called The Markup’s “flawed and misleading.” She criticized it for being “based on a non-representative sample of searches” and said using trending searches makes it more likely that results would include Google “knowledge panels”—information summaries curated from various sources—than a random sample would. As we wrote earlier, Google has not publicly released a random sample of searches. methodology Among the competitors who told The Markup that Google’s decision to highlight its own products in search is hurting other businesses is Chris Cummings, CEO of Curiosity Media, which owns SpanishDict.com. Cummings said that when Google serves SpanishDict as the top result, the site gets 80 percent of clicks—but when Google Translate appears above it, that figure drops to 2 percent. Google’s decisions have limited the company’s growth, he said: “We only exist because there are still some queries where they don’t put their stuff at the top.” Travel websites large and small said boosting Google Flights and Google Hotels on Google’s search results page has taken away business, even though some pay Google billions for ads. To determine the amount of space Google dedicates to itself, we built a custom scraper to gather U.S. Google Trends for two months, starting in November 2019. We built another scraper to run the searches as they would appear on mobile devices, where the majority of searches now occur. We wrote more than 1,000 lines of code to parse and analyze the resulting dataset, which contained more than one million rows. Members of Congress questioned Pichai, Bezos, Apple CEO Tim Cook, and Facebook CEO Mark Zuckerberg in a hearing that started more than an hour late and was ongoing as this story was published. During Rep. Greg Steube’s time for questions, he accused Google company Gmail of censoring Republican email. The Markup that Gmail was putting advocacy and political email from across the spectrum in its marketing tab rather than delivering it in the main inbox. (We found no political bias.) reported in February Pichai replied to Steube that the company is doing what users want: segregating bulk emails from messages from friends and family. Credits Originally published as " Congressman Says The Markup Investigation Proves Google Has Created a “Walled Garden” " with the Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0) license.

The Markup Investigation: Congressman Says that Google engineered a “Walled Garden”

The year was 2015 and my feature team was forced to learn Salesforce. I know that sounds extreme, but that was the reality for the six core members of our team who were required to adopt the force.com platform as a part of a corporate objective. Other than an introductory session from Salesforce during the 2008 Gartner Application Architecture and Design conference, I really had no depth of knowledge for a CRM product which promised a "No Software" implementation. However, I did receive a pretty nice T-shirt for attending the session. A Fast Journey The goal of the 2015 initiative was two-fold: get the team up to speed with the Salesforce ecosystem convert an existing CRM solution to Salesforce The six-person team consisted of four full-stack developers and two quality assurance team members. Prior to this assignment, our project sprints had been focused on using leading-edge, open-source frameworks and technologies. As one might expect, getting to know Salesforce in the pre-lightning era seemed like it might be a step back because we all felt like we were working on an assignment that lived within a closed-source proprietary system, or what is often referred to as a "walled garden". : a software system wherein the carrier or service provider has control over applications, content, and media, and restricts convenient access to non-approved applicants or content. This is in contrast to an open platform, wherein consumers generally have unrestricted access to applications and content. Walled Garden When talking about the project, at the time, my biggest hesitations with the Salesforce solution were the following: non-standard options for client customization inability to work in a git-based workflow inefficient request/response pattern for client/server communication inability to end-to-end (E2E) test Salesforce customer solutions Our implementation went live in November 2015. In fact, the project was considered the most-successful software implementation for my employer – which was delivered without any delays or compromise in functionality. Since it has been five years, I thought it would be appropriate to see how Salesforce has changed. After all, five years is considered an eternity in software design and architecture. A New and Open Course When I speak with team members in the Salesforce/MuleSoft practice at , they only nod in agreement when I recall the project which consumed about 50% of my employment in 2015. In every case, these technologists are eager to point out just how far Salesforce has evolved from being a walled garden. One key indicator to determine this is to check for the presence of open-source projects. Without much effort, I was able to locate the official Salesforce page on : CleanSlate Technology Group GitHub At the time this article was drafted, there were 211 repositories which I had access to view using my GitHub account. The next step is to understand how many of these repositories make the Salesforce experience more open sourced and break down the walls of a closed-source ecosystem. Sorting the repositories by the highest star rating to the lowest star rating, I found some very exciting and active projects—some of which publicly expose functionality internal to the Salesforce CRM solution. Client Improvements One of the biggest challenges of the Classic version of Salesforce was the lack of a rich-text client which could be extended. During my 2015 project, I remember our team included a compressed archive of JavaScript libraries to provide a user experience which was more in line with customer expectations. The Lightning upgrade with Salesforce introduced the (LCF) and a model which is called "Aura." The LCF upgrade solved a number of challenges, but still maintained some degree of proprietary design – until the following open-source projects were initiated: Lightning Component Framework : contains the source code for the Lightning Web Components Engine and Compiler lwc : -based implementation of the Salesforce Lightning Design System design-system-react React : provide the source code for the base components of the Salesforce Component library base-components-recipes Open sourcing the code related to the client-side of Salesforce communicates a zero-transparency message and allows those working in the Salesforce ecosystem to gain a better understanding of the underlying code driving the client experience. Additionally, feature teams interested in leveraging React no longer have to dive deep into the Aura model. But Wait ... There's More By open sourcing the repositories noted above, the following challenges from my 2015 implementation are no longer an issue: non-standard options for client customization inefficient request/response pattern for client/server communication The two remaining issues have also been addressed: inability to work in a git-based workflow is met with Salesforce DX inability to end-to-end (E2E) test Salesforce customer solutions is now possible with TestimSalesforce DX provides an integrated, end-to-end lifecycle designed for high-performance agile development. The lifecycle is open and flexible, facilitating integration with common and preferred tooling for building applications. By contrast, I have published the following articles on DZone in 2017 to demonstrate the steps required to automate the build process with Salesforce: Salesforce DX Continuous Integration With Salesforce Configuring Stash/Git for Salesforce With the Force IDE For Eclipse Deploying Salesforce Using Bamboo The Salesforce DX product eliminates a majority of the manually driven tasks, while introducing automation to create Scratch orgs which can be used to validate git-based updates. Once validated the Salesforce updates can be migrated using standardized deployment solutions. E2E Testing Using Testim While use of is possible with Salesforce for end-to-end tests, the product has gained approval by the Salesforce QA engineering team from an E2E perspective: Selenium Testim Automating Complex End-to-End Tests Use of the Testim product enables the confident execution of complex E2E tests, replacing the requirement of manually-executed tasks. With a stable E2E solution, QA engineers can finally embrace exploratory testing to potentially uncover bugs from alternate flows which had not been fully validated due to time and resources. Feature teams working on Salesforce-related functionality can add Testim to their development lifecycle toolset to perform the same degree of E2E testing prior to releasing new features and functionality. Conclusion At the end of 2015, the Salesforce ecosystem was a walled garden. While my feature team was successful at deploying a solution that end users loved and appreciated, the feeling of deploying a "closed" solution was bittersweet from a technology point-of-view. The current state of Salesforce reflects a great advancement from an open source perspective. In fact, a great deal of time and dedication has been allocated to support and maintain the client-related repositories noted above. Furthermore, the Salesforce DX product allows feature teams to adopt git-based workflows as a part of their development lifecycles. From an end-to-end perspective, the use of Testim handles the complexities related to complex and long-running scenarios which should be validated before the functionality is made available for customers. Salesforce provides a summary of additional open-source projects that can be tracked at the site: Salesforce Open Source This site provides another view into the mindset of the Salesforce engineering team, offering open-source contributions to concepts which do not map to CRM-related opportunities to solve. Some interesting items to consider are noted below: : a flexible, modular, and composable framework to model socio-economic behaviors and dynamics with both agents and governments AI Economist : state Management Tailored-Made for JS Applications Akita : AutoML library written in Scala that runs on top of Apache Spark TransmogrifAI : IAM Least Privilege Policy Generator. Policy Sentry Clearly, Salesforce is no longer a walled garden. Have a really great day!

Salesforce: Walled Garden No More

The walled gardens situation is on everyone's lips, as it pushes ad-tech peers on both buying and selling sides to fight over the scraps remaining after the industry titans saved the best ad budget slices for themselves. Especially with the increase of CTV prominence, the GAFA squad - Google, Amazon, Facebook, and Apple - seek a way to sink their claws even deeper into the market to control it. Using their titanic authority and status, GAFA continues to put together a full ad stack by and consolidating the ad tech landscape. Inevitably, closed ecosystems built around just four major ad platforms who control ad buying and selling operations and limit their data access, hurt the industry. acquiring point-solutions On the surface, these big guys seem to provide ad tech companies with beneficial conditions, offering them first- and third-party access, analytics solutions, demand, and other opportunities, but what was given by one hand was taken away by the other. Apart from GAFA, many more middlemen are falling back on ad dollars flowing from advertisers to publishers. So, while advertisers desperately try to stitch things together in terms of their ad spend breakdown, publishers struggle with revenue loss too as infinite mediators are eager to reap their piece. On top of all that, the walled gardens use their position to charge buyers the ad tech tax. Based on the recent and much-discussed ISBA report, – an "unknown delta", representing around one-third of supply chain costs – could not be attributed. 15% of advertiser spend What budgets and fees are circulating across the programmatic supply chain? While programmatic supply chain opaqueness has reached its apogee, numerous mediators standing in the way of buyers and sellers keep promising the earth and maximum outcomes, especially to publishers. But the harsh reality doesn't look any nicer as this is where mediators start nibbling at ad budgets here and there. For starters, by taking cover behind mandatory traffic validation procedures, intermediaries charge about $0.1/CPM value. Then, instead of the vowed maximum yield and earnings without hidden fees, mediators charge their own fee that may reach 17%, plus pocket up to 33% of publishers' revenue, divvying it up with other 'partners' and tech providers neither publishers nor advertisers are aware of. So what does that leave us? The situation is this: up to 50% of ad buys vanish before reaching publishers, as endless supply chain intermediaries literally plunder ad dollars along the way. Hence, the only beneficiaries here are multiple links on the supply chain instead of actual advertisers and publishers. With 2021 on the horizon, are we up to any changes? The evolving ad tech market has given rise to several aspects that may become a way to win more transparency and revenues beyond the walled gardens. To combat the GAFA machine, ad tech players need to make their own contribution, especially when it comes to adopting industry standards. Today, IAB offers a whole range of solutions, including ads.txt, app-ads.txt, sellers.json, and ads.cert to help advertisers avoid suspicious deals related to counterfeit inventory and get awareness of their ad spends and ad tech fees charged on the way. To make it work, publishers have to implement these standards on their side. Advertisers, in turn, should also sabotage interaction with all unauthorized counterparties. With such a synergistic approach, the ubiquitous application of the txt-series standards will be just a matter of time. Another step that publishers can take to simplify the state of things for themselves and their advertisers is considering a direct partnership model. The concept of partnership with the first-hand relationship lies in eliminating numerous intermediaries and establishing an intelligent and fair business model for buying and selling parties. By leveraging a direct partnership model and remedies for improving transparency and eliminating ad fraud, advertisers will eventually know where the money flows, so that publishers receive the highest revenue for their inventory, as they are getting paid at the full price that the advertiser pays without any opaque and unexpected charges. Cutting the dead weights Walled gardens built by industry titans challenge the whole ad tech domain globally, and it won't be easy to combat the big guys in order to change the status quo. Yet, by banding together in the fight for real transparency and leveraging an alternative partnership model, ad tech peers will manage to defeat this dragon.

Walled Gardens Transparency and the Cost of Obscurity: Will 2021 Be Different?

In the last few years, I've been dissecting of any type, and companies of any size. At the same time, I've been talking, interviewing, and discussing business models and business model innovation with dozens of entrepreneurs, and practitioners. business models While doing that I found myself diving into the business models of tech giants, which taught me a few interesting things that tell us how the business tech world might look like in the coming years. As old gatekeepers are made irrelevant, new kids on the block took over, but now they are becoming like the old gatekeepers The Internet had enabled the decentralization, disintermediation, and disentangling of products and services that before were in the hands of a few or many gatekeepers. This explains the process of disintermediation of many industries. For instance, Google and Facebook have pretty much disintermediated the advertising industry, which budget before went through expensive and non-transparent advertising agencies managing billions of dollars at lofty fees. It is important to notice that while decentralization initially was a strong force on the web. As of 2019, we see companies that took hold of the web becoming tech monopolies. Those monopolies are applying the strategy of closing their platforms to create a complete user experience. Thus, the web, rather than completing a process of decentralization, simply disintermediated several industries by creating new centers of power. And where before there were many gatekeepers. We pretty much ended up with a few platforms which work as gatekeepers on a global scale. For instance, portals like AOL, in the first web age, had used the so-called walled garden strategy. Where users were locked-in and offered a plethora of services to keep them engaged as much as possible. Eventually, companies like Google opened up that walled garden by simply pushing people on as many websites on the Internet. However, as Alphabet squeezes its core asset, the Google search engine; algorithms keep users within Google's pages with new advanced features (from answer boxes to weather reports and travel platforms). In short, the same companies that built their strategy on an open platform now go back to become a walled garden. And that is understandable for several reasons. The reasons for arising walled gardens Google's business model is now mature. And as such, it only needs to keep monetizing its users for as long as it can. Second, as a new way to interfacing and consuming content arise, Google is trying to have a firm control on the distribution of that content. One interesting thing though, as those new tech monopolies have formed and as they are propelling new industries (think of all the billions of dollars Amazon and Google are pouring into the voice search war), those new industries are starting as closed ecosystems right off that bat. In short, Google, which once worked by providing dozens of blue links to sites across the web. With voice search, it narrows the selection to a very few. We might argue the reason is that Google itself is now a mature company. Just like AOL, founded in 1985, when it dominated the first web era was an already mature company. Google, now Alphabet, is an over twenty years old company, which in tech years, make of it a senior company. From many gatekeepers to a few global platforms The primary change that has driven this age is that while before power was concentrated in the hands of a few people. That same power seems to be concentrated in the hands of a few platforms, driven by algorithms tweaked by a few people. In the end, we might conclude it was all a change of power. And as new technologies emerge, those able to understand the logic and implications of them can build a new business playground, thus make the rules of this new game and dominate. Digital platforms are just like governments As a result of their success, those platforms collect fees (a tax), thus making fat profit margins. Take how Google collects a fee for the simple fact of showing a company first when a user is searching for it. For instance, if you type "the new york times" on Google, you'll notice NY Times showing as first paid result. That means the company is paying a small fee to make sure Google shows on top of the search results. This mechanism in paid marketing and distribution is called brand protection, and it makes sense. It is so inexpensive to protect your brand from the risk of now showing on top of Google if users were searching for it, that not paying for that tax would not make much sense. We don't know for sure how much money a company like Google makes on branded searches, but we can assume it makes for sure a few billions each year. And that looks exactly like a tax. Just like governments collect taxes for passing through their borders, Google asks for a small fee to make your brand show up on top of its search results. Do consumers really have options? One of the common beliefs that have formed throughout this digital era is that consumers gained power. As they got more options and they got empowered by more free tools, free information, and free entertainment, consumers at the end picked what they liked the most. With , it's easy to assume those tech giants follow the interest of those same consumers that empowered them. Amazon obsessing on customer experience And in part that is true. On the other hand, there are three aspects to take into account: 1. On the one hand, they create value for consumers while automating most of the jobs consumers need to buy things in the first place. And that makes sense from a business standpoint. If the reasoning is to add as much value as possible to consumers, machines help do the job many times over more efficiently. 2. New consumer behaviors are induced thanks to distribution sheer force. Think of how Amazon and Google are pushing out billions of dollars in spending to promote new behaviors within the voice assistants. And one can argue that only behaviors that are more congenial to human nature will survive. But what does that mean? Aren't there different ways we can do one thing and who control distribution than isn't in charge of deciding what's right or wrong? As someone that learned to appreciate the power of distribution, I've also learned that in many cases that same distribution can also influence the habits of billions of people. Which leads to the third point. 3. As monopolies form they reduce competition, increase fragility in the system and make the overall system more fragile. So what's happening next? Will tech save the world? As new technologies emerge (think of the Blockchain), fewer people look at them with a utopist view, although libertarian dreams of decentralization arise again. Yet, as business history has taught us, the same technology can evolve in many different ways. And while some technologies do have some intrinsic aspects, which make them move in certain directions. There are multiple ways that direction might look like. And it rarely depends on the technology itself. It depends on who controls distribution. In short, for those advocating the Blockchain will drive a new wave of decentralization. Beware that also Blockchain has a strong cultural component that depends upon the consensus of who's building it. Therefore, while it might be open and decentralized initially to enable that change of power needed to break modern monopolies, it might close again over time, once new centers of powers have formed again. In this context, to understand where these centers of powers will lie, you need to follow the money, distribution power, and consumer behaviors. And after over two decades of commercial viability of the web, power got more and more distributed in a few hands. Can we hope for tech to help this to change? Or do we need to be in charge of this process instead?

Are Tech Giants Building Walled Gardens (Again)?

Rob Walling has bootstrapped multiple startups to exit, most recently Drip. These days he’s working on TinySeed, the first startup accelerator designed for bootstrappers. He also founded and runs MicroConf and Startups for the Rest of Us. Davis Baer : What’s your background, and what are you working on? I’m a software developer by training and feel like I officially became an entrepreneur in 2005 when I acquired a software product that was in its alpha release. Since then I’ve started or acquired a few dozen products, most notably SaaS applications and , both of which were acquired. HitTail Drip In addition I’ve been spearheading a community for bootstrapped founders through my , , , and now , the first startup accelerator designed for folks who would traditionally bootstrap. essays podcast conference TinySeed We will focus on SaaS companies that won’t become unicorns, but could reach $1M-$50M in annual revenue. These are companies with real business models. Charging real money. To real customers. How is TinySeed different than other accelerators? Most accelerators run for three months and founders have to relocate to a specific location. But there are thousands (if not tens of thousands) of bootstrappers who aren’t able to relocate due to family or other personal obligations. TinySeed is unique in that it’s a remote accelerator, and companies will be mentored for a full year. In addition, accelerators focus on demo day — the day at the end of the program where the companies pitch their progress in an attempt to raise a seed or Series A. Since bootstrappers tend to be capital efficient, a founding team can live off that same low six-figure investment ($100k-$150k) for an entire year if they can participate remotely, without the necessity (but the option) of raising more capital at the end of the 12 months. TinySeed companies don’t need to exit for our accelerator to work. We’ve designed a structure where the goal is to align both founder and investor interest such that founders don’t need to sell in order for the investor to get a return. Obviously, if the founder does decide to sell, everyone wins — but we’re working on a flexible profit/dividend share model where the founder can focus on growth if they want to, but also optimize for taking cash out of the company (when that makes sense). We believe that designing a funding structure that fairly aligns founders and investors in this way will mean a lot more capital will flow into this ecosystem and hence allow many more businesses like this to be built. This also means they can grow at a healthier pace and don’t need to force growth to raise their next funding round. I like to think of this as building a . That is, a startup that values people over results, has reasonable working hours, provides ample days away from the office, and generally doesn’t burn out the people involved. sane startup What motivated you to start TinySeed? There is a gap in the market that makes it very difficult to find funding if you don’t aspire to build a $1B company and exit in 7–10 years. The venture capitalists, angels and accelerators most of us are familiar with focus on Unicorns (companies that have the potential for a $1B valuation). If you apply to YCombinator with a SaaS product that can “only” grow to $10M or $20M per year in recurring revenue, your likelihood of being accepted is close to zero. But my experience working with bootstrappers for the past decade+ has given me first-hand exposure to a large group of SaaS companies that may not grow to Unicorn status, but have a good shot at becoming highly profitable, multi-million dollar businesses. As an angel investor, I’ve realized there is a serious lack of capital and support in this corner of the startup ecosystem ( is really the only player I’m aware of in this space). I also believe these companies are worth a similar level of investment (both time and financial) as those seeking to become Unicorns. Indie.vc A SaaS application that’s past its fast growth stage can have net margins of 30–50%. When you’re talking about $5M or $10M in revenue you’re not going to IPO, but it can be a life-changing outcome. To that end, TinySeed aims to “fill this gap” by writing accelerator-sized checks and providing a year of mentorship to SaaS companies that have the potential to reach $1M-50M ARR. This allows a founding team to work full-time on their business for a year. How have you generated interest in this idea? When we announced the idea in October with a landing page, a blog post and a tweet, the response was shocking, both from the founder side and the mentor/investor side. , 617 likes, 342 upvotes on Hacker News, and close to 3,500 interested people who provided us with their name and email address. About half of those provided their product URL. 175 retweets I think the key to the success of this announcement hinged on a few things: There is pent-up demand for something like TinySeed. It’s a new approach to a problem many people experience. Working a day job and launching on the side is hard. very Given my involvement with bootstrappers I have an audience in the space and had easy reach into many email inboxes and Twitter feeds of people who would care about this. We did a good job of summarizing the pain and our solution in a well-written, founder-focused landing page ( ). here What are your goals for the future? The next few months are typical of a new startup — everything happening at once and not enough time to get it all done. In the next 90 days I expect we’ll: close our funding round handle legal and corporate stuff (incorporate, bank accounts, etc.) launch a marketing website open batch applications narrow down to our final 10 or 20 companies and about 100 other things we haven’t thought of My goal is to execute on the above, handle the unforeseen, and get the first batch of founders chosen in early 2019. What are the biggest challenges you’ve faced and obstacles you’ve overcome? If you had to start over, what would you do differently? The biggest challenge we’ve faced so far is trying to educate founders and investors on the merits of this approach. The Unicorn mindset is so ingrained in startup culture that questioning it is outside the realm of reality for many people. When I’ve spoken at meetups in the Bay Area, I inevitably have to convince companies that they do not have to become the next Facebook. I explain to founders that you can be a $10 million company and be really successful, insanely profitable company without institutional money. Have you found anything particularly helpful or advantageous? The ability to control your emotions is highly underrated. I used to be terrible at this but had to get better (or implode) while I was growing Drip. At times the stress is immense, failure is almost a certainty on any given day, and if you can’t figure out a way to function with that level of stress, bad things will happen to your body, your relationships, or your life. I’ve found mindfulness and careful attention to my state of mind has become one of my superpowers. What’s your advice for entrepreneurs who are just starting out? It’s going to take longer than you want to find success, but if it were easy everyone would do it. If you base your timeline on TechCrunch articles it would seem like you can launch a startup today and be a millionaire next month. I would encourage you to focus on a more realistic timeline, and think in years, not months. Also, read books. Some of helpful books I’ve come across during my entrepreneurial journey include: by Gabriel Weinberg and Justin Mares Traction by Ryan Battles SaaS Marketing Essentials by Ben Horowitz The Hard Thing About Hard Things Where can we go to learn more? is the best place to stay up to date, and I also keep up to date with the latest happenings. TinySeed my website Feel free to ask your questions in the comments. This interview is brought to you by , a tool to schedule and automatically repeat your posts on Facebook, Instagram, Twitter, Pinterest, LinkedIn, and Google My Business OneUp

Founder Interviews: Rob Walling of MicroConf, Drip, and TinySeed

This is Part 3 on my series about cross-compilation. You can check out Part 1 and Part 2 first ! _Wherein we try to use the Microsoft Visual C++ Compiler on Linux_hackernoon.com A C++ Hello World And A Glass Of Wine, Oh My ! _Wherein we cross-compile Qt Aplications for Windows with a couple of alien tool-chains and the power of rainbow._hackernoon.com A C++ Hello World And the Cute Heartless Rainbow © https://linuxnewbieguide.org You cannot caters to the needs of Windows and Linux users while ignoring the third major, well, second actually, desktop operating system. The Operating System I’m talking about is of course developed and commercialized by a company best known as the one who gave Clang to the world, is mostly responsible for maintaining WebKit (after most of the industry moved to Chromium), and created some other amazing open sources softwares such as CUPS. And for that we should be grateful. You would think that a company who went to the trouble of starting a completely new compiler to offer a better user experience would make it easy to cross-compile to their platform. However. That company is Apple. Like on Linux and Windows, we will need to acquire and setup 3 pieces. A compiler, some system headers and libraries like and a Sdk for desktop integration purposes. libc++ If you have done OS X development before, you know that all off that can be found in XCode, a 5GB package of bundled tools most of which we won’t need. XCode is free software. As in beer. Which we will need a lot of. XCode is however proprietary, which is fine. The issue is that if you read the , you will find the following clause. Terms and condition attached to XCode The grants set forth in this Agreement do not permit You to, and You agree not to, install, use or run the Apple Software or Apple Services on any non-Apple-branded computer or device, or to enable others to do so. 2.7 Restrictions; No Other Permitted Uses I’m not a lawyer. However it seems to me that Apple is actively forbidding cross-compilation using their libraries, regardless of the technical do-ability. Therefore, some of the ideas discussed in the rest of the article could, if applied, void any agreement you have with Apple. You need an Apple ID, and so you need to create an account on apple.com. I don’t recall having a more terrible account creation experience in a while. The biggest offense is certainly their antiquated security policies. How not to do account creation and security, courtesy of Apple. They will then send you an email for verification, which is great. But instead of having a link in the email you will get a code that you cannot even paste and have to type in manually. You will then search for XCode. Fortunately, some good Samaritans maintain on Stackoverflow since 2012. valid download links This turned into a “All software is terrible” rant again. I’m sorry. On a more positive note, someone already set up a nice collection of scripts to get you started with building an OsX toolchain on Unix. Works with Cygwin too ! You will need to clone it. _osxcross - OS X cross toolchain for Linux, *BSD and Windows (Cygwin)_github.com cor3ntin/osxcross It’s a fork from ’s work which I had to patch up. Thomas Pöchtrager XCode 7.3 is shipped as a DMG and while this is an osx-specific file format, osxcross come with a script to extract it, making good use of . More on that later. Darling osxcross/tools/gen_sdk_package.sh Xcode_xxx.dmg Unfortunately, the open source community is still waiting on apple to release a ld64 linker with support for TBD files v2 that are used in later version of osx to not have to ship the .dylib in the sdk. TBD files are pretty cool, they are a YAML representation of the symbols included in a dynamic library, alleviating the need to ship the actual library. They are pretty similar in concept to the files that are generated by MSVC when building a DLL. I think TBD files could be used across platforms but for now and the open-source ld64 can’t handle the new version. .lib LLVM can’t handle them (yet ?) So we will have to stick to a 10.11 SDK. It’s reasonable ! I went to the trouble of supporting files that are used to package the later versions of XCode. A format that is inspired by babushka dolls, but with compressed archives instead. Unfortunately, we can’t use anything more recent than XCode 7.3. I hope it will change soon ! xip You can then run move the generated to then launch MacOSX10.11.sdk.tar.xz osxcross/tarballs SDK_VERSION=10.11 ./osxcross/build.sh You will also need to run osxcross/build_llvm_dsymutil.sh And in no time, you will have a complete toolchain for OSX, for both and ( even if you have absolutely 0 reason to build anything in 32 bits when targeting OSX ). i386 x86_64 It evens builds my personal favorites : and . If you ever built something on OSX, you know how terrible these tool ares. Or rather how terrible the OSX loader is. otool install_name_tool I’m properly impressed by the work that went into . osxcross Configuring QBS is rather straight forward, though there are some things to take care of. In , run: osxcross/target/bin ln -s x86_64-apple-darwin15-ld ldcp osxcross-llvm-dsymutil x86_64-apple-darwin15-dsymutil This will later help clang finding the proper tools. If you want to support multiple tool-chains, put in a subfolder ld Here is my profile configuration than you can adapt The options tells clang where to find the proper (ld64) since the system linker is not suitable to link Mach-O application. -prefix ld The rest is just giving qbs the proper search paths. Unfortunately, support for .plist in qbs is not portable, so you will encountered an error ERROR: TypeError: Result of expression 'PropertyList' [[object Object]] is not a constructor.at JavaScriptCommand.sourceCodeat Rule.prepare in /opt/qtcreator/share/qbs/modules/cpp/DarwinGCC.qbs:262:18 Comment out the rule in DarwinGCC.qbs to fix the issue. Of course, not being able to create files will be very limitating and it would be great if QBS could handle those files in a platform-agnostic manner. info.plist For now, in all our .qbs project files, we will put the following to disable bundling and therefore Info.plist generation Depends {name: "bundle"}bundle.isBundle: false At that point, we are able to build the simple console Hello World seen in Part one # file helloworldMach-O 64-bit x86_64 executable, flags:<NOUNDEFS|DYLDLINK|TWOLEVEL|WEAK_DEFINES|BINDS_TO_WEAK|PIE> But can it run ? Oh, darling ! To run our windows application we used wine. There is a rather recent effort — It started in 2012 while WINE started in 1993; Windows 3.1 was just released !— to offer a translation layer, called The project is far from being as mature and it doesn’t seems to have any kind of financial support. I hope it will catch on. darling. _Darling - macOS translation layer for Linux_darlinghq.org Darling | macOS translation layer for Linux You can clone and build darling, F . On my machine, it took a bit under one hour for the whole thing. Once installed, it’s about 800MB. which is unsurprising as it is a complete system which comes with all the usual tools including g++, ruby, python, Perl, git, bash, swift, ssh… ollow the instructions on github But, the build complete with no error, and amazingly, it works, and seems very reactive. Being more modern that wine, it is containerized ! Oh, Hi Mac Adding magic with binfmt So now we can run a mac command, but what if we could just hide the magic away ? Using a and a , I was able to create a file so that the kernel can handle launching Mach-O files. kernel facility systemd service /etc/binfmt.d/darling.conf :Mach-O 64b:M::\xcf\xfa\xed\xfe::/usr/bin/darling_interpreter::Mach-O 32b:M::\xce\xfa\xed\xfe::/usr/bin/darling_interpreter::Mach-O FAT:M::\xca\xfe\xba\xbe::/usr/bin/darling_interpreter: /usr/bin/darling_interpreter is a script that launches — it should be executable. darling shell #!/bin/bash/usr/local/bin/darling shell "$1" The darling documentation suggests that should work, but it does not. darling <binary> After restarting the binfmt service ( ), we are able to launch OSX application transparently. systemctl restart systemd-binfmt Which mean we can do this Any sufficiently advanced technology is indistinguishable from magic Oh, and by the way, you can do the same thing for windows executables and WINE. Some distributions do that out of the box. In part 2, I attempted to install the win32 version of the Qt Framework on Linux without using windows. I failed. Can we get Qt for mac without a mac ? I did download the installer. It’s a .dmg. It would be an issue, but we can mount DMGs on Linux. No problem at all. It’s the kind of things we do here. with the power of darling But mounting the Qt installer dmg reveals that it contain a binary and a file rather than a simple folder or something manageable. .dat Presumably, the binary is the installer. Maybe it runs on darling ? No. Hard dependency on the OpenGL framework. Great software bundled in crappy unusable packages seems to be a recurring theme. Is all hope lost again ? Not this time. We can build Qt for mac, like we attempted for Windows. But it will work. Mac has . It knows about clang and gcc, it’s very much like Linux in a lot of aspects. There is an UNIX under there after all ( But I always had the feeling that OSX internals are terrible, hidden under a nice user interface. For starter, , 10 years ago ). make a large number of tools haven’t been maintained after their upstream version moved to GPLv3 Alas, it means dealing with Qt’s complex build system. It took some time to hack the build files. See, Qt makes the terrible assumption that all toolchain on osx involve xcode. We don’t have xcode. qmake But once you bypass all the automatic probes and assumptions about what’s installed on the system…. ….you can get it to work ! #configure -release -opensource -confirm-license -xplatform macx-cross-clang -skip qtwebengine -nomake examples -nomake tests    -prefix /home/cor3ntin/dev/cross-compilers/osx/qt5_10 It was not quite the end of the road. Qt Widgets failed to build because of missing dependencies. QtLocation failed to build because the libc++ headers were too old or broken ( I fixed that by copying the latest libc++ version within the OSX SDK. It worked). Then QtLocation complained because was not defined so I hacked a few headers. std::auto_ptr I tried to get (chromium ) to build, but it uses another build system still, and I had done enough build system hacking for one night. qwebengine But in the end most of Qt did build. It took me a few hours but I got Qt ! And what we have then, is something quite interesting. The binaries are natives Linux ELF, while the frameworks and libraries are Macho-O. That will be handy in a minute. Hello, I’m a Mac — And I’m a PC. And I am a Mac too. — What ? Qt is a big piece of software making full use of the underlying system capabilities, in term of OS integration. If we can build that, we can build almost anything. I initially called my mkspec file . I was a bad name. it also prevented qbs to understand that it was a mac build. Rather than renaming the mkspec and rebuilding Qt, I modified qbs. the code of is actually parsing the mkspec’s files with regular expressions. Somehow, it works. Just don’t breathe on it. darling-clang qbs-setup-qt .conf Eventually, once I gave qbs what it expected to understand that we were dealing with mac, I was able to call qbs-setup-qt osx/qt5_10/bin/qmake osx-x64-qt510 Which created the right profiles and modules. I cleaned up my profiles manually to merge and clang-osx osx-x64-qt510 And then we can compile or magnificent app ! Hello World Compiling an OSX application based on Qt on Linux, using qbs : check What now ? Well, we have a complete toolchain, maybe we can check some things ? Using , the osx equivalent of we can assert that we indeed link to some libraries otool -L ldd We can use other tools. such as or or the dreaded nm strings install_name_tool Unfortunately, Darling can’t handle anything remotely graphical yet, So we need a Mac to actually launch our app. A real mac; It is illegal to visualize Mac OSX, unless it’s on a mac. Some words comes to mind. Imagine my French. Let’s talk about the mac. You probably know the one. Most companies have it. The Mac. This Is A True Story. It used to belong to Bob. It’s Bob’s mac. But then, 5 years ago, Bob died, so T_he Mac_ was offered to Alice. Alice left the company soon after - Pure coincidence, probably. Yet was always It has no master. It has no puppet either. You can connect to it at . The password is simply . It’s not in the same LAN as the other servers running on virtual machines somewhere. It’s not administered in anyway, presenting a convenient back door in an otherwise tightly secured network. The Mac The Mac. ssh bob@mac.yourcompany.com pass When it’s running. Sometime it’s just down for days at a time. It’s not just like people don’t care, they also don’t know where it’s physically located. A mac mini is easy to loose. Under someone desk, wedging an old chair, used as a coffee table. Last time it crashed, you had to track it down for three days straight, like chasing after a big cat in the mountains. You even tried to call Bob’s widow. You finally found sandwiched between Carol’s screen and the Oxford Dictionary. “It was the prefect height!”, objected Carol when you took back her 800$ monitor stand. You traded for an old IKEA catalog that Carol found to be as practical, if not more, than a Mac-Mini. the mac the mac You plugged back in and diligently updated it it to the last version of OSX, “Cougar” (or something, it’s hard to keep track). the mac And when, a few days later, your coworker got a new car and you lost your home, I was left wondering : Is it appropriate to require a credit card to apply free security fixes ? But truth is, is doing an important job. Running all the tasks that can only be run on . Signing installers, doing mac packages, iOS stuff, running the New York stock exchange… you are not quite sure, it’s Bob’s after all. the mac the mac Maybe life would have been different if we could virtualize OSX. I do happen to have a 2011 mac-mini, gift of a former employer. Its life was a a bit different than It was never loved and spent the last two years in a box. It only gets to see the life of the day for the needs of this article. It is also the reason why I’m 4 days late in my publication schedule. I tried to install High Sierra — Grey screen; I had to reformat, install Lion, then install El Captain. So El captain is what we will be using. It features a whooping 2GB of memory, the system using 3/4 of that. the mac’s. You should enable VNC, remote share and SSH in the systems parameters of your mac. This article is starting to be a bit long. So here is a quick visual summary of the work achieved so far: Definitively improving our productivity. We should stop fooling around. Copy you OSX build of Qt on the OSX Machine. You can use , or a shared folder ( through samba ) scp -r rsync Copy your on the machine, again, using . helloworld-gui scp -r Our cross-compiled build of Qt does not contain . You could get it by installing the official version of Qt on the mac directly. To avoid doing that and to not have to deal with , we can set up to point to the folder containing all the . DYLD_FALLBACK_FRAMEWORK_PATH is somewhat sane but may not work and has some associated security risks. Please don’t use it in shipped applications. macdeployqt the hairy mess of [install_name_tool](http://doc.qt.io/qt-5/osx-deployment.html) DYLD_FALLBACK_FRAMEWORK_PATH Qt*.framework export DYLD_FALLBACK_FRAMEWORK_PATH=/Users/cor3ntin/dev/cross-compilation/qt5_10/lib Like on windows, we need to provide a file next to our , to tell Qt where to load its plugins from ( the application will simply not run otherwise ). Mine looks like that qt.conf helloworld-gui [Paths]Prefix=/Users/cor3ntin/dev/cross-compilation/qt5_10/Plugins=plugins Now, connected to the mac through you can execute your application and it will appear on the mac screen and on the remote display / VNC session ssh It feels like we climbed a mountain ! Can we make all that legal ? Clang, LLVM, ld64 and the associated tools are open source projects. That does not mean the open source versions match the version Apple is using. In fact, Apple’s Clang is a modified version of Clang proper and they lag a few versions behind upstream. Which is ironic given they started the project. LD64 And the “cctools” are released under the “Apple Open Source licence” And the versions of that tools used by XCode are 2 years ahead of what the open source community has access too. The frameworks themselves are not open source, and, as I mention at the beginning, not redistributable. And the open source alternative which is now only maintain by the people is insufficient. cocotron Darling There are a few issues with it They do not have a build script to actually build an SDK and only install the . this could probably be fixed rather easily. .dylib They have a limited set of frameworks and that set is insufficient to build Qt. Qt uses the following framework, those prefixed by are missing in Darling !!!! * AppKit* ApplicationServices!!!! AssetsLibrary* AudioToolbox!!!! AudioUnit!!!! AVFoundation!!!! Carbon* Cocoa!!!! CoreAudio!!!! CoreBluetooth* CoreFoundation* CoreGraphics!!!! CoreLocation!!!! CoreMedia!!!! CoreMotion* CoreServices* CoreText* CoreVideo!!!! fftreal* Foundation* ImageIO!!!! IOBluetooth* IOKit!!!! OpenCL* QuartzCore* Security!!!! SystemConfiguration You could possibly compile the Qt Framework on OSX and then copy them to your Linux machine, it would probably work in most cases. Utilizing an SDK that is not the official SDK kinda defeats the purpose of cross compiling to test your software works on mac. you are just testing it works on Darling. There is no guarantee that Darling SDK headers match the official one. Mingw suffers from that issue too. So it’s technically possible to cross compile complex applications ( including Qt and Qt based one) for Mac on Linux. It’s even possible to run non-graphical applications and unit tests directly and seamlessly. But it’s illegal to use the SDK making the whole exercise a bit pointless depending your legislation. Open sources alternative exist but may not be sufficient and sufficiently reliable. And while we can be hopeful, it’s doubtful Apple will ever have more developer friendly politics. And on that terrible disappointment, it’s time to end !

Stories, News & Mentions about WALLID

Wallid’s Stories on HackerNoon