My network and server setup eat about 200 W, split right down the middle. I have plans to cut the network consumption by replacing the PoE switch, so that I can use the savings to add another server.

So it’ll still be about 200 W, but more useful.

This appears to be an AI port of an AI project named after a conspiracy theory?

Is this for wood handling (ie avoiding splinters and sharp edges)? Or for use with rotating machinery? I don’t think the latter is advisable.

At a minimum, a lot of IoT devices that use IPv6 exclusively – eg the Matter standard – can finally stop compiling extra support for Legacy IP that they just don’t use. The kernel savings will likely be modest, but in an era where more and more things are being squeezed into less and less RAM – not necessarily due to DDR5 prices, since microcontroller SRAM is a different animal entirely – this will be appreciated as some breathing room for embedded developers.

There’s also the security argument, where disabling Legacy IP means that the (relatively few) CVEs against the IPv4 network stack simply won’t apply. Reducing one’s attack surface is always a good thing, if simply to reduce the number of warnings from software supply chain scanners.

I’m under no misapprehension that Legacy IP would be switched off as a default in the Linux kernel anytime soon, but offering a real choice means developers can actually customize their kernel build to be as optimal as possible. It removes a technical barrier, meaning all that remains is the political matter: the will to turn off Legacy IP entirely.

A few notes about the text of Prop 39:

• There is no exception from the requirement to present ID even for a provisional ballot. This means that someone who is robbed of their ID on the morning of Voting Day will have no time to get a replacement government-issued ID, nor are they allowed to cast a provisional ballot and have their ID checked later. They are SOL.
• It creates a private cause-of-action for anyone to sue the state or the counties related to their election laws, where under existing law, only those people actually affected by a law’s implementation have the right to sue. In the past, wacko advocacy groups had to go find people in each county as the vehicle to bring their lawsuit, but with this change, we can reasonably expect a barrage of suits, without ever having to ostensibly vindicate the voting rights of some Californian somewhere.
• The exact type of ID (state DL, USA Passport book or card, tribal govt ID) must be designated by each voter in advance of voting. So a voter who designates their USA Passport book is out-of-luck if their passport expires two months before the election and they didn’t remember to update their designation. Even if the voter has a state DL card on their person, only the pre-designated ID type is the one which will allow them to vote
• A minor point: counties would be prohibited from counting any ballots at all, including provisional ballots, until their ID or 4-digit PIN is verified. This means the verification process would become a bottleneck on the evening of Election Day. If it should happen that the state’s database has problems with uptime, then the counties will be constitutionally disallowed from even doing an unofficial count of the ballots. It is well documented from 2020 and 2024 that undue delays to ballot counting will undermine confidence in the election, and also open avenues for judicial malfeasance or gross misconduct by rogue election officials. A well-timed cyber attack on the verification server on Election Day would be devastating, since it would cause lines to queue around the block, until voters have to leave the line to use the bathroom, go to work, or anything else that a prolonged voting experience would cause.
• The state would have to provide a free ID to use for voting, but this is not the same thing as a general purpose ID card. Such a “voting ID” card is single purpose, making it as easy to misplace as a card to Top Golf for infrequent golfers. It would have been much simpler and easier if the proposition mandated full-and-proper state ID be free to everyone. The other issue is that “at no charge” does not mean “without any other requirements”. Nothing in the text would prevent a requirement to undergo a free-of-charge Equifax credit/identity check, for example.

These are just the obvious problems from the text of the proposition; future problems would almost certainly arise from the enabling law that implements this convoluted system. At the federal level, the SAVE Act was an unmitigated train wreck. In no way does that make Prop 39 a “moderate” or “common sense” solution that is worthy of passage. There isn’t really a Goldilocks zone when it comes to restricting the vote.

It’s also just not been a serious problem: of all the alleged voter fraud in the past decade, no serious suggestion has been made for the idea that fraudulent votes actually changed the result of a race. From a risk/reward perspective, intentionally casting a vote when not eligible means risking state or federal prison time, to change by 1 a vote count that usually is in the tens of thousands (city/local races) or in the millions (California-wide races).

Using a bazooka to hit a nail is guaranteed to end with collateral damage: eligible voters will have to navigate a bureaucracy to do something that their constitution said was guaranteed to them. Meanwhile, actual vote fraudster will exploit the ten other ways to cast votes improperly, because election officials will be distracted by spending their time checking IDs more carefully than even bartenders, lest they get sued by some RW nutter organization in the lead up to every election.

For general verification, it’s fine I suppose.

For anything that logs into federal government services, it’s utterly absurd that they use the privately-operated ID.me when the government already operates login.gov to do exactly the same thing. The agencies that use ID.me exclusively are the ones which truly confuse me.

For other folks reading, be mindful that not all PC fans can efficiently draw airflow through a filter. The key word is “static pressure”, which refers to how much force can be exerted on a fully-blocked opening, which would then produce airflow through the parts that are less blocked. The simplest PC fans basically have zero static pressure, and so they’ll just spin uselessly when covered by an HVAC filter.

As an aside, dust and lint in PCs can be reduced by limiting dust and lint in the environment: replacing your home HVAC filter, not wearing shoes indoors, electrostatic fans, and using mechanical ventilation rather than opening the window are effective ways to keep detritus away from your machine in the first place.

dust blanket

Me: hmm, that’s an odd way to describe a bit of dust

[looks closely at photo]

Me: oh dear gawd!

The closest I’m familiar with is !askelectronics@discuss.tchncs.de , but it’s more about questions and answers, rather than general discussion. But then again, questions are a good way to start a discussion.

I will also post some of my stuff to !imadethis@lemmy.zip , but that’s for a complete build and not so much about the electronics journey that preceded it.

Ok. so now you got to understand that algorithms were generally viewed as positive back then.

The thing to keep in mind is that an algorithm is the instrument which optimizes for some incentive structure. In the early days of Google Search, the incentive was predominantly Adsense ads displayed alongside the search results. So to optimize views of these ads, the search algorithm had to be good enough so that people would keep coming back to Google Search for their web search needs. More users means more views means more eyeballs for ads.

What changed everything started with the incentive, when it became clear that marketers’ demands to hyper-target consumers could be met if websites like Google or Facebook could learn more about users than just the search terms they entered. This is when user tracking took off, where the incentive is now to monetize user habits and identity, ideally cross referenced to their social media profiles or login details.

With that new incentive, the way to optimize that is to use tracking cookies and to expand Google Analytics and AdSense so they appear on even more websites, each of which would report data back to Google as users moved around the web. Note how this means the quality of web search is no longer important, because whether or not the user lingers on Google Search, their habits are still recorded and monetizable. In this sense, we would say that the algorithm’s optimizing incentive is no longer aligned with the user’s interests, such as privacy or performance.

Capitalism, as always, requires an examination into incentive structures, and this explains pretty much most of what Google has been doing vis-a-vis web behavior for the last decade, with the notable exception of their Gemini AI. I readily admit that I have no idea how AI fits into their web search business model, apart from being a loss leader that keeps up the impression that they’re a major player in the current bubble. And I often wonder if they have any idea too.

Excuse me, sir, this is dull_mens_club, not awesome_mens_club. I’ll have to ask you to leave.

JK! This is awesome! I love recovering e-waste, though usually I just collect the li-ion cells from wayside vapes. Regarding how this thing is driven, I’m continually saddened that display manufacturers don’t bother with the tiny amount of extra SRAM needed to support a persistent image.

For one project of mine in the distant past – using a device similar to an Alpha Smart – the display did have its own memory but the board designers couldn’t be bothered to run the MISO line from the display to the CPU. So even though I could write to the display memory, I still had to keep a full copy of the rendered graphics in CPU memory, which was significant. All this made worse because the display chip itself supported a reasonable set of manipulations, like scrolling and shifting the display. But alas, none of it could be used effectively.

I’ve been rediscovering “old” tech. Not necessarily because DDR prices are through the roof, but because some things were genuinely well built in the not-too-distant past.

And sure, I did enjoy servicing a 1980s VT320 serial terminal that I got for free, but I’ve also been examining what an original Raspberry Pi 1 is capable of doing. Did you know the RPi has hardware H.264 encoding capabilities? And mpeg2 hardware decode? With these things being fairly available – just ask some tech friends if they have some to spare – it’s not inconceivably to make a TV frontend to access Jellyfin, in lieu of a Google TV stick or whatever.

Likewise, some pursuits avoid new hardware outright: a FreeBSD-based router runs really well on older hardware, because the drivers will have been tested and rock solid by now. And Asterisk as a PBX software is great if you want to get into telephony. If modern packet switching is more your thing, dn42 is one such environment that you can plug into and pass traffic with other networks.

And if you don’t have PC hardware laying around, maybe try smaller: microcontrollers and embedded/IoT hardware is still fairly affordable, and has a breadth of applications, from home automation to just building things for the lolz. MeshCore and Meshtastic are comms networks built from low cost radios.

Just because some hardware became expensive does not mean the entire hardware space is out of reach, and in any case, the software domain is wide open. In that regard, Free and Open Source Software (FOSS) is timeless. This is akin to trying your hand at Software Defined Radio (SDR) before taking the plunge by buying an expensive ham radio.

The tech field is not limited to just Apple and Samsung. There’s also Dell, and Tektronix, and Linux, and Texas Instruments, and Arduino, and soldering, and KiCAD, and FlipperZero, and HomeAutomation, and IPv6, and more. Find your thing in this great big field!

In the end, programming is commanding a computer system to do your bidding. If your goal is to make electronic metal music, or complex 3d designs, simulate fluid dynamics, or solve for wheel settings on the Enigma machine, those are all programming tasks. Some may involve long runtimes, but the programming is giving the instructions to execute.

Even formulas in Excel are programming. But make no mistake that programming and software engineering are very different tasks: the latter is about devising good designs that balance countervailing interests. Software engineers define the problem which programmers would command onto a computer to find a solution or produce a product.

The way the question is framed is a tad confusing to me. The historical development of writing systems by humans is unambiguous, in that the spoken language came first and then writing systems later. So the idea that learning a spoken language is easier/harder due to its writing system is a strange mismatch of not-really dependent factors.

Insofar as Japanese, Mandarin, and Cantonese are concerned, my cursory understanding is that Japanese uses a different ordering for nouns and verbs, while the latter two have more similar ordering to English, Korean, Vietnamese, and western European language. Someone should fact check that.

That said, English is not a pinnacle example of a alphabetic language, because it has so many loanwords that its pronunciation system strains to logically capture all its rules. The romance languages do a more decent job of matching the written language to the spoken language.

At bottom, learning to speak a foreign language can be done without learning its writing system. When speakers can’t write the language, it is called illiteracy. But to learn a writing system without its spoken language, there are few examples, because every writing system conceived by humans always encoded some aspect of the spoken language’s grammar. Pre-1900 Vietnamese using the Chinese script is not fully comprehensible to a Japanese reader, as an example.

It is time for all computing professionals billionaires and corporations to accept responsibility for computing’s humanity’s current state

FTFY

In all seriousness, Vardi correctly identifies the problems with “move fast and break things” and “efficiency at the expense of resilience”. Yet then uses victim-blaming language that “we allowed” mega corporations to exist. Even if the “we” is read as “society at large”, it cannot be reconciled with the call-to-action, which unambiguously targets the individuals of the industry. Without effective organization, no amount of individualized action can overcome a structural problem. It’s unclear to me what his point is.

In a nutshell, there will not exist any premade, out-of-the-box collaboration platform that works precisely for your use-case. And indeed, for anyone’s use-case, because there are so many variables at play. Do you have:

• to deal with scale of 10-100 people or 100k to 10M?
• trolls and malicious actors that want to derail your project by bikeshedding?
• a user base that might not already be skilled at recognizing toxic team members?
• a legal entity for this project which is required to keep minutes or any other reporting requirements?
• paid developers around the world, which only use this platform to talk to you, thus it must be secure to store PII?
• and so on

The reason people think it’s easy to choose a collaboration tool is because they see Teams and Slack and think that they can just buy a product (or a FOSS equivalent) and they can hit the ground running. But anyone who has set up a Slack community knows that there are a lot of rules and conventions that have to be established. For example, FTX (the failed cryptocurrency exchange) was using Slack emojis to approve business expenses. Setting aside that example of colossally bad financial controls, every org of any size must establish what their baseline processes are, and that extends to the collaboration platform.

but I don’t want to worry about all the moderator tools and restrictions and BS either.

Moderation is not optional. Every non-trivially sized collection of people will end up with disagreements that distract from their joint reason for being there. To not moderate is a choice, as demonstrated most succinctly by this post: https://www.techdirt.com/2022/11/02/hey-elon-let-me-help-you-speed-run-the-content-moderation-learning-curve/

At the very least, the collab platform you choose to use should have some tooling for a future moderator to utilize, so that you don’t have to switch platforms the moment someone reveals themselves to be a neo-nazi, starts spamming swastikas into the chat, and you’re powerless to do anything about it. For small groups, it’s less probable but the risk grows over time.

Originally, I was going to compare these barriers to the ones that Evan Edinger in the UK recently referenced, when discussing an under-river pedestrian crossing in London that nominally disallows cycling but is also the only available crossing for miles. The local council there tried adding barriers but had to remove them due to blocking wheelchair users, whom must not ever be treated as second-class citizens.

But instead, I want to focus on what White River NF said about these “step over” bars. As a fencing strategy, the practice predates the USA national forest system by a lot, with examples being found in the English countryside whenever private land is crossed by a public footpath. Such easements have existed since antiquity, so landowners demarcated their boundaries by fencing, with an opening that requires one or two steps so that animals would not necessarily migrate to adjacent land.

In an era before bicycles or wheelchairs, this works fine for filtering humans through, on paths that are explicitly for non-wheeled people traffic. But in the modern day, even the British look at such inaccessible steps as a historical quirk, not as a template to be carried forward to designated multi-use trails.

To that end, I believe White River NF when they say that such “step over” bars have been in use by the Forest Service for a long while. But I question whether they have correctly updated their design standards to conform to the Architectural Barriers Act (ABA) governing federal lands. How I understand it, the terrain exemption can only be applied to the portion of a trail where accessibility accommodations are impractical. There is no blanket exemption for a whole trail if only one part is exempt by terrain. And seeing how these bars are constructed, it seems very possible to remove them. Even if insurmountable terrain lays 100 meters beyond the bars, to keep the bars is to tell wheelchair users that no, they cannot enjoy even those 100 meters, and they are not welcome.

Every NF unit is managed separately, but ultimately each has to choose if they will continue to ratify the now-probably-disallowed choices from yesteryear, or if they will break with precedence and work to solve modern challenges as good stewards of the land. To be clear, I don’t want e-motos on trail system, but I refuse to exclude adaptive riders and wheelchair users to achieve that end. White River NF has to try again.

Is the question in relation to immigration requirements, or for job prospects? For the former, any accredited university will likely not have a problem being recognized, not unless it’s a for-profit university or has a sizable “controversies” section on Wikipedia. Online degrees from an established public university generally bear the name of the university at-large, so there’d be no noticeable difference when looking at the diploma.

As for the job prospects, that’s going to be very market specific. Perhaps the larger question is whether an online degree or not, is there enough demand in your prospective countries for MEs and EEs that employers are willing to go with a new resident from abroad, rather than from one of the established local universities?

Do you think it’s a good idea to get both a sleeve and a laptop bag to protect the bag?

This might be excessive, but I think the other problem you’d have is that most laptop bags are sized to take a naked laptop. So if a sleeve takes up the remaining margin, you might it very difficult to slide the encased laptop into the bag.

I’d probably start with the option that is most adaptable first, so if you always carry a USB C charger along with your laptop, then a laptop bag would store both of those and thus you can eschew the case. Though if you have free returns, then there’s no harm in also trying the laptop case for a while.

Right now I have a rain coat and rain pants but I don’t want to wear them when it’s not raining since it gets hot

In this case, I’d roll the coat and pants into a cylinder and then strap them down with a bungee cord over the top of your rear rack. Clothes are soft and squishy, so the tension of an adjustable bungee cord will keep them in place. BTW, an adjustable bungee cord is a valuable bike accessory in any scenario.

and I need to remove it anyways so it doesn’t get stolen so it was more convenient for me to put it in another bag.

I don’t suppose there’s any way to store your bike indoors at work, or in a bike locker if your workplace has those? With how opportunistic bike theft is in the USA, you’ll want to holistically assess which parts of your bike can be stolen, against how annoying/difficult it would be to do so, and by the price or inconvenient of that part. It’s a rough calculus, for sure, but considering how the rear wheel of an ebike (which usually has the hub motor) is the most valuable removable component, everyone makes their own choice whether and which smaller components also need to be guarded and to what degree.

It’s a fair question, about which parts to take indoors if you can’t bring the whole bike in. But your risk assessment will be a lot different than mine, so I can’t really give a more specific answer.

One last thought before I forget: with waterproof panniers, always check them every so often, since if they can prevent water from getting in, they can also prevent water from getting out. I once went riding in the rain where it was cold enough for water in the air to condense inside the pannier, which drained to the bottom but had no way to escape. So it just sat there in a small puddle, until I noticed.

Waterproof does not mean airtight, and humid air can carry a decent amount of water under the right conditions, especially if it can accumulate over time and no checking is being done.

Given the disparate concerns, I think it best to divide them logically.

For the laptop, I would suggest looking at a laptop case or laptop bag, since it sounds like you were otherwise transporting the laptop naked or wrapped in a sweatshirt. Being your primary work tool, it does warrant its own protection, irrespective of an ebike or automobile or airplane commute. Laptop cases are like phone cases, in that they stay attached to the machine whether opened or close, usually by straps around the bottom. Some laptop cases are hard sized, and thus offer a lot of physical protection.

But a laptop bag would work equally well, such as those which look like messenger bags. Such soft sided bags offer less protection than a hard sided case to impacts, but they often are water/rain resistant when the bag is zipped up or the flap is laid over the top. I personally use a messenger bag with a laptop compartment, so I can carry my other work tools, like pen and paper, phone charger, and meeting headphones, whether I’m commuting to the office by car, bike, or walking.

In all circumstances, this layer of protection is bidirectional, preventing the laptop from sustaining damage but also protecting your panniers from holes. So now we get to panniers. Ideally, your laptop case/bag should fit into the pannier, and the pannier would be waterproof. This augments the rain resistancd of the laptop bag, while being flexible for other non-laptop payloads like groceries when you’re not commuting.

Since you mentioned transporting other electronics as well as lunch, a dual pannier setup can make sense, where you laptop case/bag plus electronics are on one side, and lunch on the other. I do carry dual panniers on my acoustic bike, and it’s terrible convenient when I don’t want my payloads to unbalance the bike.

For rain clothes, presumably you would be wearing these while commuting? Do these need to be stored in the panniers? Or would this be in preparation for a rainy commute on the way back home? I personally prefer rain gear that I can shed anywhere, so I use breathable snow pants that still function in the dry, and a windbreaker than isn’t too cumbersome to wear even if it isn’t raining.

My bike tools are stored in an undersaddle bag which never leaves the bike. This means I need them to be fairly compact, but I do keep the essentials on me wherever I may run into trouble.

I think the key thing to remember with a bicycle commute – electric or otherwise – is that things will bounce around. It’s not like in a car where something can be placed in the passenger footwell and it’ll mostly get to the destination in one piece. No, stuff in panniers will bump into things, the side walls, and sometimes even try to escape out the top. Things like bike tools won’t care about that, but electronics can be marred or damaged. Fortunately, this can be really easy to manage, with the simplest solution being a reusable grocery bag that you can tie shut. This clumps all the things together, so while the whole clump can still bounce around inside the pannier, they won’t smash into each other. A step up would be some kind of padded bag, like what’s used to transport cameras, with movable walls.

I wish you good luck in optimizing your bike commute! Don’t be afraid to try stuff out and do test things to their limits so you have confidence that it’ll serve you under all inclement conditions.