March 2025 - Ploum.net En - listes.ploum.net

Goodbye Offpunk, Welcome XKCDpunk!
by Ploum.net (English posts only) 31 Mar '25

31 Mar '25

GOODBYE OFFPUNK, WELCOME XKCDPUNK! by Ploum on 2025-04-01 https://ploum.net/2025-04-01-xkcdpunk.html For the last three years, I’ve been working on Offpunk, a command-line gemini and web browser. Offpunk.net https://offpunk.net While my initial goal was to browse the Geminisphere offline, the mission has slowly morphed into cleaning and unenshitiffying the modern web, offering users a minimalistic way of browsing any website with interesting content. Rendering the Web with Pictures in Your Terminal (ploum.net) https://ploum.net/2022-03-24-ansi_html.html Focusing on essentials ====================== From the start, it was clear that Offpunk would focus on essentials. If a website needs JavaScript to be read, it is considered as non- essential. It worked surprisingly well. In fact, in multiple occurrence, I’ve discovered that some websites work better in Offpunk than in Firefox. I can comfortably read their content in the former, not in the latter. By default, Offpunk blocks domains deemed as nonessentials or too enshitified like twitter, X, facebook, linkedin, tiktok. (those are configurable, of course. Defaults are in offblocklist.py). Cleaning websites, blocking worst offenders. That’s good. But it is only a start. It’s time to go further, to really cut out all the crap from the web. And, honestly, besides XKCD comics, everything is crap on the modern web. > As an online technical discussion grows longer, the probability of a comparison with an existing XKCD comic approaches 1. > – XKCD’s law XKCD’s law (ploum.net) https://ploum.net/xkcds-law/index.html If we know that we will end our discussion with an XKCD’s comic, why not cut all the fluff? Why don’t we go straight to the conclusion in a true minimalistic fashion? Introducing XKCDpunk ==================== That’s why I’m proud to announce that, starting with today’s release, Offpunk 2.7 will now be known as XKCDpunk 1.0. Xkcdpunk.net https://xkcdpunk.net XKCDpunk includes a new essential command "xkcd" which, as you guessed, takes an integer as a parameter and display the relevant XKCD comic in your terminal, while caching it to be able to browse it offline. Screenshot of XKCDpunk showing comic 626 https://ploum.net/files/xkcdpunk1.png Of course, this is only an early release. I need to clean a lot of code to remove everything not related to accessing xkcd.com. Every non-xkcd related domain will be added to offblocklist.py. I also need to clean every occurrence of "Offpunk" to change the name. All offpunk.net needs to be migrated to xkcd.net. Roma was not built in one day. Don’t hesitate to install an "offpunk" package, as it will still be called in most distributions. offpunk package versions - Repology (repology.org) https://repology.org/project/offpunk/versions And report bugs on the xkcdpunk’s mailinglist. xkcdpunk-users on lists.sr.ht https://lists.sr.ht/~lioploum/offpunk-users Goodbye Offpunk, welcome XKCDpunk!

1 0

The candid naivety of geeks
by Ploum.net (English posts only) 28 Mar '25

28 Mar '25

THE CANDID NAIVETY OF GEEKS by Ploum on 2025-03-28 https://ploum.net/2025-03-28-geeks-naivety.html I mean, come on! ================ Amazon recently announced that, from now on, everything you say to Alexa will be sent to their server. Pluralistic: Amazon annihilates Alexa privacy settings, turns on continuous, nonconsensual audio uploading (15 Mar 2025) (pluralistic.net) https://pluralistic.net/2025/03/15/altering-the-deal/ What surprised me the most with this announcement is how it was met with surprise and harsh reactions. People felt betrayed. I mean, come on! Did you really think that Amazon was not listening to you before that? Did you really buy an Alexa trusting Amazon to "protect your privacy"? Recently, I came across a comment on Hacker News where the poster defended Apple as protecting privacy of its users because "They market their product as protecting our privacy". I mean, once again, come on! Did you really think that "marketing" is telling the truth? Are you a freshly debarked Thermian? (In case you missed it, this is a Galaxy Quest reference.) The whole point of marketing is to lie, lie and lie again. What is the purpose of that gadget? =================================== The whole point of the whole Amazon Alexa tech stack is to send information to Amazon. That’s the main goal of the thing. The fact that it is sometimes useful to you is a direct consequence of the thing sending information to Amazon. Just like Facebook linking you with friends is a consequence of you giving your information to Meta. Usefulness is only a byproduct of privacy invasion. Having a fine-grained setting enabling "do not send all information to Amazon please" is, at best, wishful thinking. We had the same in the browser ("do-not-track"). It didn’t work. I’ve always been convinced that the tech geeks who bought an Amazon Alexa perfectly knew what they were doing. One of my friends has a Google Echo and justify it with "Google already knows everything about our family through our phones, so I’m trading only a bit more of our privacy for convenience". I don’t agree with him but, at the very least, it’s a logical opinion. We all know that what can be done with a tool will be done eventually. And you should prepare for it. On a side note, I also postulate that the reason Amazon removed that setting is because they were already gathering too much data to justify its existence in case there’s a complaint or an investigation in the future."How did you manage to get those data while your product says it will not send data?". But, once again, any tech person knows that pushing a button in an interface is not a proof of anything in the underlying software. Please stop being naive about Apple =================================== That’s also the point with Apple: Apple is such a big company that the right hand has no idea about what the left hand is doing. Some privacy people are working at Apple and doing good job. But their work is continuously diluted through the interests of quick and cheap production, marketing, release, new features, gathering data for advertising purpose. Apple is not a privacy company and has never been: it is an opportunistic company which advertise privacy when it feels it could help sell more iPhones. But deeply inside, they absolutely don’t care and they will absolutely trade the (very little) privacy they have if it means selling more. Sometimes, geek naivety is embarrassingly stupid. Like "brand loyalty". Marketing lies to you. As a rule of thumb, the bigger the company, the bigger the lie. In tech, there’s no way for a big company to not lie because marketers have no real understanding of they are selling. Do you really think that people who chose to advertise "privacy" at Apple have any strong knowledge about "privacy"? That they could simply give you a definition of "privacy"? I know that intelligent people go to great intellectual contortions to justify buying the latest overpriced spying shiny coloured screen with an apple logo. It looks like most humans actively look to see their freedom restricted. Seirdy calls it "the domestication of users". WhatsApp and the domestication of users (seirdy.one) https://seirdy.one/posts/2021/01/27/whatsapp-and-the-domestication-of-users/ And that’s why I see Apple as a cult: most tech people cannot be reasoned about it. The Cost of Being Convinced (ploum.net) https://ploum.net/the-cost-of-being-convinced/index.html You can’t find a technical solution to a lie ============================================ Bill Cole, contributor to Spamassassin, recently posted on Mastodon that the whole DNS stack to protect spammers was not working. > spammers are more consistent at making SPF, DKIM, and DMARC correct than are legitimate senders. 🆘Bill Cole 🇺🇦: "@jwz@mastodon.social The stats we collect for the…" (toad.social) https://toad.social/@grumpybozo/114213600922816869 It is, once again, a naive approach to spam. The whole stack was designed with the mindset "bad spammers will try to hide themselves". But was is happening in your inbox, really? Most spam is not "black hat spam". It is what I call "white-collar spam": perfectly legitimate company, sending you emails from legitimate address. You slept in a hotel during a business trip? Now you will receive weekly emails about our hotel for the rest of your life. And it is the same for any shop, any outlet, anything you have done. Your inbox is filled with "white-collar" junk. And they know this perfectly well. In Europe, we have a rule, the RGPD, which forbid businesses to keep your data without your express consent. I did the experiment for several months to send a legal threat to every single white-collar spam I received. Guess what: they always replied that it was a mistake, that I was now removed, that it should not have happened, that I checked the box (which was false but how could I prove it?) or even, on one occasion, that they restored a backup containing my email before I unsubscribed (I unsubscribed from that one 10 years before, which makes it very unlikely). In short, they lied. All of them. All of them are spammers and they lie pretending that "they thought you were interested". In one notable case, they told me that they had erased all my data while, still having the cookie on my laptop, I could see and use my account. Thirty days later, I was still connected and I figured that they simply managed to change my user id from "ploum" to "deleted_ploum" in the database. While answering me straight in the face that they had no information about me in their database. Corporations are lying. You must treat every corporate word as a straight lie until proved otherwise. But Ploum, if all marketing is a lie, why trusting Signal? ========================================================== If you can’t trust marketing, why do I use Signal and Protonmail? First of all, Signal is open source. And, yes, I’ve read some of the source code for some feature I was interested in. I’ve also read through some very deep audit of Signal source code. Reviewing the Cryptography Used by Signal (soatok.blog) https://soatok.blog/2025/02/18/reviewing-the-cryptography-used-by-signal/ I’m also trusting the people behind Signal. I’m trusting people who recommend Signal. I’m trusting the way Signal is built. But most importantly, Signal sole existence is to protect privacy of its users. It’s not even a corporation and, yes, this is important. Yes, they could lie in their marketing. Like Telegram did (and still does AFAIK). But this would undermine their sole reason to exist. I don’t say that Signal is perfect: I say I trust them to believe themselves what they announce. For now. What about Protonmail? ====================== For the same reasons, Protonmail can, to some extent, be trusted. Technically, they can access most of the emails of their customers (because those emails arrive unencrypted to PM’s servers). But I trust Protonmail not to sell any data because if there’s any doubt that they do it, the whole business will crumble. They have a strong commercial incentive to do everything they can to protect my data. I pay them for that. It’s not a "checkbox" they could remove, it’s their whole raison d’être. This is also why I pay for Kagi as my search engine: their business incentive is to provide me the best search results with less slop, less advertising. As soon as they start doing some kind of advertising, I will stop paying them and they know it. Or if Kagi starts becoming to AI centric for my taste, like they did for Lori: Why I Lost Faith in Kagi (d-shoot.net) https://d-shoot.net/kagi.html I don’t blindly trust companies. Paying them is not a commitment to obey them, au contraire. Every relation with a commercial entity is, by essence, temporary. I pay for a service with strings attached. If the service degrade, if my conditions are not respected, I stop paying. If I’m not convinced they can be trusted, I stop paying them. I know I can pay and still be the product. If I have any doubt, I don’t pay. I try to find an alternative and migrate to it. Email being critical to me, I always have two accounts on two different trustable providers with an easy migrating path (which boils down to changing my DNS config). Fighting the Androidification ============================= Cory Doctorow speaks a lot about enshitification. Where users are more and more exploited. But one key component of a good enshitification is what I call "Androidification". Androidification is not about degrading the user experience. It’s about closing doors, removing special use cases, being less and less transparent. It’s about taking open source software and frog boiling it to a full closed proprietary state while killing all the competition in the process. Android was, at first, an Open Source project. With each release, it became more closed, more proprietary. As I explain in my "20 years of Linux on the Desktop" essay, I believe it has always been part of the plan. Besides the Linux kernel, Google was always wary not to include any GPL or LGPL licensed library in Android. 20 years of Linux on the Desktop (part 3) (ploum.net) https://ploum.net/2025-03-08-linux_desktop3.html It took them 15 years but they finally achieved killing the Android Open Source Project: Google will develop the Android OS fully in private, here's why (www.androidauthority.com) https://www.androidauthority.com/google-android-development-aosp-3538503/ This is why I’m deeply concerned by the motivation of Canonical to switch Ubuntu’s coreutils to an MIT licensed version. Ubuntu 25.10 plans to swap GNU coreutils for Rust (go.theregister.com) https://go.theregister.com/feed/www.theregister.com/2025/03/19/ubuntu_2510_… This is why I’m deeply concerned that Protonmail quietly removed the issue tracker from its Protonmail Bridge Github page (making the development completely opaque for what is an essential tool for technical Protonmail users). I mean, commons! ================ This whole naivety is also why I’m deeply concerned by very intelligent and smart tech people not understanding what "copyleft" is, why it is different from "open source" and why they should care. We need more of Richard Stallman, not less (ploum.net) https://ploum.net/2023-06-19-more-rms.html Corporations are not your friend. They never were. They lie. The only possible relationship with them is an opportunistic one. And if you one to build commons that they cannot steal, you need strong copyleft. On Open Source and the Sustainability of the Commons (ploum.net) https://ploum.net/2024-07-01-opensource_sustainability.html But firstly, my fellow geeks, you need to lose your candid naivety. I mean, come on, let’s build the commons!

1 0

20 years of Linux on the Desktop (part 3)
by Ploum.net (English posts only) 08 Mar '25

08 Mar '25

20 YEARS OF LINUX ON THE DESKTOP (PART 3) by Ploum on 2025-03-08 https://ploum.net/2025-03-08-linux_desktop3.html > Previously in "20 years of Linux on the Deskop": After contributing to the launch of Ubuntu as the "perfect Linux desktop", Ploum realises that Ubuntu is drifting away from both Debian and GNOME. But something else is about to shake the world… 20 years of Linux on the Desktop (part 1) https://ploum.net/2024-10-20-20years-linux-desktop-part1.html 20 years of Linux on the Desktop (part 2) https://ploum.net/2024-12-16-linux_desktop2.html The new mobile paradigm ======================= While I was focused on Ubuntu as a desktop solution, another GNOME+Debian product had appeared and was shaking the small free software world: Maemo. It will come as a shock for the youngest but this was a time without smartphones (yes, we had electricity and, no, dinosaurs were already extinct, please keep playing Pokémon instead of interrupting me). Mobile phones were still quite new and doing exactly two things: calls and SMSes. In fact, they were sold as calling machines and the SMS frenzy, which was just a technical hack around the GSM protocol, took everybody by surprise, including operators. Were people really using awkward cramped keyboard to send themselves flood of small messages? Small pocket computers with tiny keyboard started to appear. There were using proprietary operating systems like WinCE or Symbian and browsing a mobile version of the web, called "WAP", that required specific WAP sites and that nobody used. The Blackberry was so proprietary that it had its own proprietary network. It was particularly popular amongst business people that wanted to look serious. Obama was famously addicted to his Blackberry to the point that the firm had to create a secure proprietary network only for him once he took office in the White House. But like others, Blackberries were very limited, with very limited software. Nothing like a laptop computer. N770, the precursor =================== In 2005, Nokia very quietly launched the N770 as an experiment. Unlike its competitors, it has no keyboard but a wide screen that could be used with a stylus. Inside was running a Debian system with an interface based on GNOME: Maemo. The N770, browsing Wikipedia https://ploum.net/files/old/Nokia770-fi-wiki-600x450.jpg Instead of doing all the development in-house, Nokia was toying with free software. Most of the software work was done by small European companies created by free software hackers between 2004 and 2005. Those companies, often created specifically to work with Nokia, were only a handful of people each and had very narrow expertise. Fluendo was working on the media framework GStreamer. Immendio was working on the GTK user interface layer. Collabora was focusing on messaging software. Etc. Far from the hegemony of American giant monopolists, the N770 was a mostly European attempt at innovating through a collaborative network of smaller and creative actors, everything led by the giant Nokia. During FOSDEM 2005, GNOME developer Vincent Untz lent me a N770 prototype for two days. The first night was a dream come true: I was laying in bed, chatting on IRC and reading forums. Once the N770 was publicly released, I immediately bought my own. While standing in line in the bakery one Sunday morning, I discovered that there was an unprotected wifi. I used it to post a message on the Linuxfr website telling my fellow geeks that I was waiting for my croissants and could still chat with them thanks to free software. Those days, chatting while waiting in a queue has been normalised to the point you remark someone not doing it. But, in 2005, this was brand new. So new that it started a running meme about "Ploum’s baker" on Linuxfr. Twenty years later, some people that I meet for the first time still greet me with "say hello to your baker" when they learn who I am. For the record, the baker, an already-old woman at the time of the original post, retired a couple years later and the whole building was demolished to give place to a motorbike shop. This anecdote highlights a huge flaw of the N770: without wifi, it was a dead weight. When I showed it to people, they didn’t understand what it was, they asked why I would carry it if I could not make calls with it. Not being able to use the Internet without a wifi was a huge miss but, to be fair, 3G didn’t exist yet. Another flaw was that installing new software was far from being user-friendly. Being based on Debian, Maemo was offering a Synaptic-like interface where you had to select your software in a very long list of .deb packages, including the technical libraries. Also, it was slow and prone to crash but that could be solved. Having played with the N770 in my bed and having seen the reactions of people around me when I used it, I knew that the N770 could become a worldwide hit. It was literally the future. There were only two things that Nokia needed to solve: make it a phone and make it easy to install new software. Also, if it could crash less, that would be perfect. The Nokia (un)management guide to failure ========================================= But development seemed to stall. It would take more than two years for Nokia to successively release two successors to the N770: the N800 and the N810. But, besides some better performance, none of the core issues were addressed. None of those were phones. None of those offered easy installation of software. None were widely released. In fact, it was so confidential that you could only buy them through the Nokia website of some specific countries. The items were not in traditional shops nor catalogues. When I asked my employer to get a N810, the purchasing department was unable to find a reference: it didn’t exist for them. Tired by multiple days of discussion with the purchasing administration, my boss gave me his own credit card, asked me to purchase it on the Nokia website and made a "diverse material expense" to be reimbursed. The thing was simply not available to businesses. It was like Nokia wanted Maemo to fail at all cost. While the N800 and N810 were released, a new device appeared on the market: the Apple iPhone. I said that the problem with the N770 is that you had to carry a phone with it. Steve Jobs had come to the same conclusion with the iPod. People had to carry an iPod and a phone. So he added the phone to the iPod. It should be highlighted that the success of the iPhone took everyone by surprise, including Steve Jobs himself. The original iPhone was envisioned as an iPod and nothing else. There was no app, no app store, no customisation (Steve Jobs was against it). It was nevertheless a hit because you could make calls, listen to music and Apple spent a fortune in marketing to advertise it worldwide. The marketing frenzy was crazy. Multiple people that knew I was "good with computers" asked me if I could unlock the iPhone they bought in the USA and which was not working in Europe (I could not). They spent a fortune on a device that was not working. Those having one were showing it to everyone. With the iPhone, you had music listening and a phone on one single device. In theory, you could also browse the web. Of course, there was no 3G so browsing the web was mostly done through wifi, like the N770. But, at the time, websites were done with wide screens in mind and Flash was all the rage. The iPhone was not supporting Flash and the screen was vertical, which made web browsing a lot worse than on the N770. And, unlike the N770, you could not install any application. The iPhone 1 was far from the revolution Apple want us to believe. It was just very good marketing. In retrospective, the N770 could have been a huge success had Nokia done some marketing at all. They did none. Another Linux on your mobile ============================ In 2008, Google launched its first phone which still had a physical keyboard. Instead of developing the software from scratch, Google used a Linux system initially developed as an embedded solution for cameras: Android. At the same time, Apple came to the realisation I had in 2005 that installing software was a key feature. The App Store was born. Phone, web browsing and custom applications, all on one device. Since 2005, people who had tried the N770 knew this was the answer. They simply did not expect it from Apple nor Google. When Android was first released, I thought it was what Maemo should have been. Because of the Linux kernel, I was thinking it would be a "free" operating system. I made a deep comparison with Maemo, diving into some part of the source code, and was surprised by some choices. Why Java? And why would Android avoid GStreamer in its multimedia stack? Technical explanations around that choice were not convincing. Years later, I would understand that this was not a technical choice: besides the Linux kernel itself, Google would explicitly avoid every GPL and LGPL licensed code. Android was only "free software" by accident. Gradually, the Android Open Source Project (AOSP) would be reduced to a mere skeleton while Android itself became more and more restricted and proprietary. In reaction to the iPhone and to Android, Nokia launched the N900 at the end of 2009. Eventually, the N900 was a phone. It even included an app store called, for unknown marketing reasons, "OVI store". The phone was good. The software was good, with the exception of the infamous OVI store (which was bad, had a bad name, a non-existent software offering and, worse of all, was conflicting with deb packages). The N900 would probably have taken the world by storm if released 3 years earlier. It would have been a success and a huge competitor to the iPhone if released 18 months before. Is it too late? The world seems to settle with an Apple/Google duopoly. A duopoly that could have been slightly shacked by the N900 if Nokia had done at least some marketing. It should be noted that the N900 had a physical keyboard. But, at that point, nobody really cared. When failing is not enough, dig deeper ====================================== At least, there was the Maemo platform. Four years of work. Something could be done with that. That’s why, in 2010, Nokia decided to… launch Meego, a new Linux platform which replaced the Debian infrastructure by RPMs and the GNOME infrastructure by Qt. No, really. Even if it was theoretically, the continuation of Maemo (Maemo 6, codenamed Harmattan, was released as Meego 1), it felt like starting everything from scratch with a Fedora+KDE system. Instead of a strong leadership, Meego was a medley of Linux Foundation, Intel, AMD and Nokia. Design by committee with red tape everywhere. From the outside, it looked like Nokia outsourced its own management incompetence and administrative hubris. The N9 phone would be released in 2011 without keyboard but with Meego. History would repeat itself two years later when people working on Meego (without Nokia) would replace it with Tizen. Yet another committee. From being three years ahead of the competition in 2005 thanks to Free Software, Nokia managed to become two years too late in 2010 thanks to incredibly bad management and choosing to hide its products instead of advertising them. I’ve no inside knowledge of what Nokia was at this time but my experience in the industry allows me to perfectly imagine the hundreds of meetings that probably happened at that time. When business decisions look like very bad management from the outside, it is often because they are. In the whole Europe at the time, technical expertise was seen as the realm of those who were not gifted enough to become managers. As a young engineer, I thought that managers from higher levels were pretentious and incompetent idiots. After climbing the ladder and becoming a manager myself, years later, I got the confirmation that I was even underestimating the sheer stupidity of management. It is not that most managers were idiots, they were also proud of their incompetence and, as this story would demonstrate, they sometimes need to become deeply dishonest to succeed. It looks like Nokia never really trusted its own Maemo initiative because no manager really understood what it was. To add insult to injury the company bought Symbian OS in 2008, an operating system which was already historical and highly limited at that time. Nodoby could figure out why they spent cash on that and why Symbian was suddenly an internal competitor to Maemo (Symbian was running on way cheaper devices). The emotional roller coster =========================== In 2006, I was certain that free software would take over the world. It was just a matter of time. Debian and GNOME would soon be on most desktop thanks to Ubuntu and on most mobile devices thanks to Maemo. There was no way for Microsoft to compete against such power. My wildest dreams were coming true. Five years later, the outlooadministrative hubris. The N9 phone would be released in 2011 without keyboard but with Meego.k was way darker. Apple was taking the lead by being even more proprietary and closed than Microsoft. Google seemed like good guys but could we trust them? Even Ubuntu was drifting away from its own Debian and GNOME roots. The communities I loved so much were now fragmented. Where would I go next? (to be continued) > Subscribe by email or by rss to get the next episodes of "20 years of Linux on the Desktop". > > I’m currently turning this story into a book. I’m looking for an agent or a publisher interested to work with me on this book and on an English translation of "Bikepunk", my new post-apocalyptic-cyclist typewritten novel which sold out in three weeks in France and Belgium.

1 0