💾 Archived View for soviet.circumlunar.space › sweater › smog › 20210130-SmogIssue7.gmi captured on 2023-04-26 at 13:54:29. Gemini links have been rewritten to link to archived content
⬅️ Previous capture (2021-12-03)
-=-=-=-=-=-=-
________ ________ _____ / ___| \/ o| _ |@ __ \ \*`--.|o. . | | | | | \/ `--. \ |\/| | | | | | __ /\__/ / | | \ \_/o/ |_\ \ \___o/\_| |_/\___/ \____/ News from the Free Internet Issue 7, January 30, 2021
1. Opening Thoughts: The Right To Repair
2. Gemini and Gopherspace News
3. Tech News
4. Cyberspace Musings
5. Classifieds
Unless stated otherwise, the material here is shared under the Attribution-NonCommercial-ShareAlike 4.0 International license.
Mirroring, copying, distribution, and derivative works are not only permitted, they are encouraged. If you wish to distribute Smog to your friends, I specifically encourage you to mirror it instead of linking to it. This will help keep hosting cost down and help Smog reach a wider audience.
Yeah, better redirect support would be nice, but what Gopher *really* needs is cookies!
by littlejohn <littlejohn@sdf.org>
In the last couple of weeks, there's been a steady bubbling of the stories about the Right to Repair -- in the media, on forums all over the Internet and so on. The most recent example?
Right to Repair off to the races in 2021 with 14 active states
But that's not all. Back in November last year, the European Parliament voted on a resolution to grant consumers a "right to repair" by facilitating their access to repairing services. Specifically, by "making repairs more appealing, systematic, and cost-efficient, whether by extending guarantees, providing guarantees for replaced parts, or better access to information on repair and maintenance", and by "call[ing] for measures to tackle practices that shorten the lifetime of a product, and endorse[ing] sustainable production."
You can read more about it here.
If you're not familiar with how European institutions work, let me ruin it for you a bit. This is the equivalent of the Guild of Snoozy Snails issuing a declaration of intent to look into the possibility of lodging a strongly-worded letter of protest with the Klingon Empire -- it's not as imaginary, but it will likely take as much time for something to happen as a result.
That being said, it was enough to trigger some PR alarms, and some of the ensuing propaganda slowly bubbled up through the news feeds.
Now, some of it is in the form of business strategy, which I readily admit to know nothing about and to be completely uninterested in. My strategy for keeping shareholders happy is basically "build quality things that run quality code". But many arguments are technical or, rather, pseudo-technical (as in "pseudo-scientific") and I think they are naive. These are very much on my turf, and I think they are bullshit.
The most common argument is that things are difficult to repair simply because they've reached such a level of miniaturisation and precision that they *cannot* be meaningfully repaired anymore. Making things more repairable would require using more discrete, bulky parts, which would result in increased cost, not to mention inferior technical quality.
Can miniaturisation make things harder to repair? Absolutely, a wire-wrapped computer backplane is usually easier to probe and repair compared to a modern-day computer mainboard. But does it have to make *all* repair procedures harder, or outright impossible? Not at all!
The most obvious example is a cracked screen, probably the most widespread kind of damage that phones, tablets and laptops suffer these days: making the screen easily replaceable requires some design effort, yes, but it barely makes a dent in the profit margin -- if it makes one at all -- especially in high-end devices that are already sold at tremendous margins. And miniaturisation has very little to do with it: the screen is one of the bulkiest components.
Making things "infinitely repairable" is unattainable, of course, and it's probably a bad idea, too. But making *some* types of repairs easy -- preferably, the ones that are most likely to be needed -- is a different story. Common repairs are hard to do today not as a result of the inevitable march of progress, but as a result of the fact that making things easy to repair has, at best, not been a design goal. In many cases, making things hard to repair has either been an explicit design goal, or just silently encouraged through the product management chain. I, for one, have only worked in one place where easy servicing was a deliberate design goal. It's a meager record.
The other red herring is that making things easy to repair and making them out of bulky, discrete parts go hand in hand -- the idea being that making devices "more repairable" would make them bulkier and heavier. That's not the case at all.
First, nobody is seriously suggesting that Apple should replace their A14 SoC with the equivalent discrete ICs, RAM and all. But that's hardly the obstacle, either. In fact, most attempts at repairing things are hindered by very different obstacles, such as lack of service manuals and schematics, or lack of replacement parts.
Second, and more importantly, *this is not just about high-end consumer electronics*. It's also about kitchen appliances and TVs and Bluetooth speakers and whatnot.
Most of these devices don't really feature high-density BGA chips that need expensive rework stations to fix -- a couple of screwdrivers, a hot air station, and a basic oscilloscope are enough for the most common breakage scenarios. But good luck getting replacements for the parts -- many of which are unmarked, undocumented, and only available, in bulk, to the manufacturer.
Another frequently-encountered frame is that, if people were to change their digital devices less often, the steady pace of innovation and progress would be reduced.
There is ample historical evidence to the contrary -- the field of computing did, in fact, progress by leaps and bounds back when the total market potential for computers was, notoriously, around five or so (fun fact: for what it's worth, though, Watson never *quite* said that).
But even if that weren't the case, this sort of prediction is very shabby. Whenever their exact scientific predictions turn out not to be quite right -- that is to say, pretty much all the time -- economists are quick to point out that economy isn't an exact science. Who knows: perhaps, if the threshold for buying a new device would be "the new one can do a lot more things" as opposed to "my hamster sat on the old one and now it's broken", manufacturers would have to put their R&D departments to even better use, and come up with even better tools faster than the comfortable pace of planned obsolescence allows them to do today.
I think an *improvement* to the state of repairability is, in fact, perfectly attainable at practically no cost -- in many cases, just by publishing some information, making some parts available, and not issuing cease & desist letters. It's perfectly sustainable, unless your goal is, quite literally, to make devices that are as bad as possible while still being allowed to sell them. In a healthy free market system, companies that do that should *theoretically* go under in no time.
A *radical* improvement -- this is a belief, this time, not an engineer's analysis -- is possible and affordable, and would have a tremendous impact over our lives. Not just in a way that 'dem dirty hippies would appreciate, happy sunshine and save-the-Planet and all, but in a way that even your average Silicon Valley sociopath would appreciate. It would make a big heap of money available for financing actual innovation, as opposed to spec-churning cheap plastic. It would lead to consumers demanding better and more advanced products, and sharpen competition.
It should come as no surprise that "the industry" doesn't want that. "The industry" doesn't want a lot of things, that's why we have anti-monopoly laws and antitrust laws and all sorts of other anti- things that are there *precisely* in order to keep the machine of competition well-oiled.
This one, I think, needs an oil change!
@nihilazo@tiny.tilde.website is teasing us with a screenshot of what might *eventually* turn into a Gemini client for the Kobo e-reader!
---
UK Weather Services over Gemini
Caolan McMahon built a very neat service that exposes UK weather forecasts over Gemini. It's a great little... Gemapp, I guess is the term? Seeing it in action feels very liberating for some reason.
Check out the weather in my favouritest place ever!
(Spoiler alert: it's probably raining).
---
Public HTTP proxy for Gemini space
Mansfield just announced a public HTTP proxy for Gemini space. It's a fast, neat service with a small browser footprint that makes browsing Gemini very easy and enjoyable.
---
The AXPbox, an open-source Alpha emulator forked from es40, just reached its 1.0 release. Among others, it runs OpenVMS 8.3 and 8.4 -- with plenty of caveats, but still!
---
pip drops support for Python 2
The latest release of pip has finally removed support for Python 2. It's one of the last plugs that hasn't yet been pulled and, while there is no shortage of unhappy users, this hasn't caused too many ripples. The writing had been on the wall for so long it was practically obscured by the time this announcement hit.
---
Heap-based buffer overflow in sudo
Qualys, whose recent security, ahem, exploits, are nothing short of fantastic, released an advisory regarding a vulnerability found in sudo. It appears to have been there for a long time -- almost ten years! -- but Qualys only recently discovered it.
The fixes have been published *very* quickly, a testament both to Qualys' work and to that of Todd Miller & co.. I warmly recommend you to read Qualys' write-up because it's very interesting.
---
The latest version of Firefox, which is... 85? Does anyone still keep count anymore? Anyways! The latest version of Firefox was released this week, featuring better protection against supercookies, bookmark improvements, and a bunch of security fixes. It also -- *sad trumpet sound* -- dropped support for Adobe Flash.
---
i386 will be a Tier 2 architecture starting with FreeBSD 13.0
It's official: starting with the 13.0 release, i386 will be a Tier 2 architecture. That means it will still see major development, but it won't be supported by the security officer, release engineering, and port management teams.
---
("Back" as in "All the cool kids sit in the back of the class")
Dissecting the Apple M1 GPU, part II
A while back, we linked to Alyssa Rosenzweig's first jab at the Apple M1 GPU. In the meantime, she published part II, along with a bunch of code based on all the goodies she uncovered. Check it out!
---
25 Years Since The First Java Release
It feels like a lifetime ago: January 23rd marks the 25th anniversary of Java's 1.0 release. Loathed by many -- but used by many more -- Java is still of the most pervasive technologies of our days. It has evolved in ways we'd never envisioned at the time, and is powering devices that few, if any, could even think about -- a testament to its flexibility and resilience.
---
The Common Desktop Environment is still developed in 2021
Okay, well, for... certain values of "still developed", but it still is! I'm sure many of Smog's readers either already knew this, or will be pleasantly surprised to hear about it because nostalgia is a hell of a drug.
My own memories about CDE aren't... particularly fond, as CDE wasn't exactly the apex of Unix workstations desktops. They're also interspersed with memories about writing Motif code so I guess some of these memories are outright gruesome. But hey, at the very least, it's far better to think about the days when I had to worry about my parents' "be home by ten or else!" curfew, rather than the government's lockdown...
---
Almost one year ago, Google's Project Zero published a very cool zero-click exploit for the iOS. Project Zero takes another look at iMessage, in the light the new protection mechanisms that Apple introduced. It's not an easy read but it's worth skimming, at the very least. It discusses a few very relevant real-world mitigations in an application with a great deal of consumer exposure.
---
Achieving 11M IOPS & 66 GB/s IO on a Single ThreadRipper Workstation
Benchmarks notoriously tell only a small part of the story, and they tell it badly, but this article convincingly paints the picture of a very near future where the rule of thumb about how slow non-volatile, large scale storage is (on account of hard-disks being slow) will be far less relevant than it is today.
---
Did you know about Globish? After the last issue's Opening Thoughts on language diversity and Gemini, one of our readers writes to tell us about it!
From: <redacted> To: littlejohn@sdf.org Subject: SMOG 6 Re: language diversity Dear John, I have read /somewhere/, that there is even a term for this particular English variant, which is not really English: Globish Native English speakers may have a hard time understanding it, if they have never been exposed to it. Oh, I see, there is even a wikipedia page: => https://en.wikipedia.org/wiki/Global_English
Smog and your friendly editor does not endorse any of the products, services, organisations, individuals or technologies mentioned below. However, I do not *not* endorse them, either!
Garden Gnomes Unite! @
@ /
I know you've heard of Astrobotany, \/ ,
and of course you have a plant. ||/
^^^^^^
Take your engagement to the next level with
gemini://gardengnome.ml
> #CircusInPlace --------------------------------------------------------- > The only jitsi call featuring legit theater clowns, hackers, and lovely > weirdos from all over the world! Hang out with terrific people as they > work on their passion projects, terrible puns, and circus skills. The > chat starts at 8pm UTC-6 MWF, and runs late enough to wind down with > friends overseas. > Contact @russsharek@mastodon.art, or check #CircusInPlace on Mastodon. > ------------------------------------------------------------------------
---
Smog runs ads on any topic for free and will never charge or accept money in exchange for publishing an ad. However, your friendly editor reserves the right to say no to an ad, or to stop running it at any time. That being said, I don't plan to say "no" to anything that's legal and civil.
Ads are published in random order, for up to 4 consecutive issues, re-shuffled each time. Ads will have to be no longer than 8 lines, wrapped at 72 columns, for a grand total of 576 characters. I encourage creative expression with ads though -- use ASCII art, sed one-liners, whatever you want!
If you'd like to see your ad here, drop me a line on littlejohn@sdf.org.
If I get interesting letters from my readers, I will publish them in the following issue. I encourage you to send your thoughts on your friendly editor's email address: littlejohn@sdf.org. If you would rather *not* get your letter published, please make a mention of it.
Want to see your writing in Smog? I will gladly publish or re-publish original articles, as long as they're not illegal or offensive, and if you're willing to license them under a license that does not prohibit free, non-commercial distribution and derivative works.
Please note that Smog is a non-commercial project with built-in SEO deterrence, an audience of maybe 12 people, and a business model that is best summed up by the word "nope", so all payment is in hipster points. If we are ever in the same pub, I will also buy you a beer (or a non-alcoholic equivalent!)