thequux

I can't judge the veracity of the history of hash functions, but the moment it starts talking about cryptography it goes completely off the rails: it seems to indicate that finite field exponentiation o'r high degree polynomials are used in cryptographic hash functions; they are emphatically not. It presents password hashing as just applying a suggest function to the password; in practice a KDF is used, which is a completely different design space (for a start, KDFs have a tweak parameter, usually called a salt in this context). Finally, there's a haven't reference to quantum computers breaking hash functions and needing post-quantum algorithms as a result. This does brush with reality in that Grover's algorithm does theoretically eat half the first preimage resistance security level of your hash function, but even SHA256 will require 2^128 iterations on a quantum computer, which will likely never be feasible. Worse, it doesn't help at all in attacks against second perimeter resistance or collision resistance.

Considering that everything I have personal knowledge of here is obviously bunk, best ignore the rest of it too

It's been a long time since I've touched any of this, so the details have slipped my mind. However, the general idea was that there were two different exit calls in DOS: terminate and terminate and stay resident. The difference between the two is that the stay resident option wouldn't release the memory used by your application. Further, the interrupt table, which told the processor how to handle each interrupt, was in RAM and therefore writable.

So, what TSRs would do is overwrite one or more interrupts to point to a routine that would check if the system call in question was one it wanted to handle (eg, to add a hotkey it would grab the keyboard handler and check for a special set of keys before passing control back to the normal handler). Once that was fine, it would call the TSR system call and control would be passed back to the OS with the hook still in place

How do you feel about ADD 5 TO X GIVING X?

I kid, of course, but "X is x + 5" brings Prolog to mind with all the unification and bidirectionality that implies.

So I think the following is IMO by far the biggest problem, no matter one's personal opinion: > "Rakoff entered a permanent worldwide injunction covering ten Anna’s Archive domains: annas-archive.org, .li, .se, .in, .pm, .gl, .ch, .pk, .gd, and .vg."

Legally speaking, the Southern District of New York can say whatever it likes, and Libera, Sweden, India, St-Pierre-et-Miquelon, Greenland, Switzerland, Pakistan, Grenada, and the British Virgin Islands are free to ignore what the US says. They all have national sovereignty over their respective ccTLDs, and of them, most are not going to simply accept the US telling them what to do considering recent geopolitical missteps.

NixOS using https://github.com/thequux/nix-zone-firewall/ worked well for me for many years. I only stopped using it because my poor embedded Linux machine started having issues and it made more sense to go with a Mikrotik than to buy a new device to run as a soft router.

$227k just to apply, and another few hundred thousand in legal, compliance, and contracting to reach delegation.

Source: I'm on the board of dotMeow and wrote the financial plan

Hi! I'm one of the founders of dotMeow, and it's been a long road to get to the point that we're willing to stake our reputations on a crowdfunding campaign. From applying to the applicant support program (we believe we were one of the first, if not the first organization accepted) to working out financial plans and overall strategy to achieve NIS2 compliance on a shoestring budget, it's been a year and a half of effort to reach where we are now.

It's quite late here in Belgium, so I'm about to go to sleep, but in the morning, I'll happily answer any questions people have about the project.

Do catgirls count? We've got one or two on staff. We'd have a cat, too, if it weren't for the fact that I'm allergic.

The sad thing is, this is one of the rhetorical flourishes that gets drilled into you in debate and comms education. ChatGPT picked it up because it's extremely common.

A few people have called us on this, but it's simply not true. The copy was all written by hand; the only machine assistance here is XCompose (you, too, can type an em-dash!).

It turns out that when everybody involved is some level of techie and you're targeting a sort of "corporate bland" tone so that you look professional and trustworthy by modeling your communications on other successful kickstarters, it comes out looking similar to what other marketing copy looks like, which is precisely what all of the LLMs were trained on.

As much as I dislike mandatory certification, I can understand the need for it in wireless battery powered devices: a malfunctioning decide can talk the battery life if everything within range, and most consumers aren't equipped to realize that this is happening much less identify the device that's causing the problem

Perhaps the solution is to make the spec open but make using the trademark contingent on certification (much like USB, for example)

Ehh, not really. OSI layer 5 was responsible for managing multiple non-overlapping sessions within a single transport stream, and routing those sessions to specific applications. This is precisely what HTTP/1.1 did (though the accept, content-type, accept-encoding, and transfer-encoding headers are really an implementation of layer 6); QUIC, on the other hand, covers the same layers that TCP does (3-5) plus the aforementioned layer 6.

I recommend actually reading X.200 (the specification of the OSI model) at some point: it's quite approachable (especially for an ITU spec, which are notoriously dense reading), and will quickly make you realize how silly it is that we still use it as a reference for modern stacks.

In that case, you don't want cloud; you want an MSP, whose core competence is running those IT services. They, in turn, have the skills to colo a rack at a DC or to manage rented servers, amortized across a number of clients.

In practice, there are two situations where cloud makes sense:

1. You infrequently need to handle traffic that unpredictably bursts to a large multiple of your baseline. (Consider: you can over provision your baseline infrastructure by an order of magnitude before you reach cloud costs) 2. Your organization is dysfunctional in a way that makes provisioning resources extremely difficult but cloud can provide an end run around that dysfunction.

Note that both situations are quite rare. most industries that handle that sort of large burst are very predictable: event management know when a client will be large and provision ticket sales infra accordingly, e-commerce knows when the big sale days will be, and so on. In the second case, whatever organizational dysfunction caused the cloud to be appealing will likely wrap itself around the cloud initiative as well.

I recall reading somewhere that the entire point of solitaire (at least the original implementation that came with windows 3) was to teach users how to click and drag, so I'm not surprised that it was good for teaching your colleague how to use a mouse

OSI's session layer did very little more than TCP/UDP port numbers; in the OSI model you would open a connection to a machine, then use that connection to open a session to a particular application.

X.400 was a nice idea, but the ideal of having a single global directory predates security. I can understand why it never happened

On X.509, the spec spends two chapters on attribute certificates, which I've never seen used in the wild. It's a shame; identity certificates do a terrible job at authentication

Alas, the market can remain irrational longer than I can remain solvent.

Years ago, I had a CD marked "Windows 2000 for Alpha RC1", which suggests that it was cancelled quite late in the release cycle.

I think that this can change the semantics though; with the preceding check you can miss the shared variable being decremented from another thread. In some cases, such as if the shared value is monotonic, this is done, but not in the general case.

If you actually read ITU T-REC X.200, which specifies the OSI model, you'll find that it doesn't match the modern internet at all. E.g., we don't have an OSI-style transport protocol at all (connections themselves aren't addressable independent of the SSAPs), TCP and UDP are actually layer 5, the presentation layer is protocol-specific, and pretty much the entire stack falls to bits if the network layer isn't packet switched.

There's a separate term for the bits of the OSI model that are actually relevant; it's called the IETF model.

There was an ISA I saw a while back that featured an "enhanced multiply and accumulate signed" instruction, which of course got the mnemonic "EMACS"

I second your recommendation for trying fountain pens. I suffer from some form of arthritis, and fountain pens let me write again.

There are a variety of cheap ones available; I'm fond of the Platinum Preppy. They're cheap as chips, write nicely, and have a fine version that actually lives up to its name. The Lamy Safari is also popular, but I found it too chunky to be comfortable.

Many barcode scanners these days can scan QR codes. I have a NetumScan NSL5 that I got for €30 or so that can handle QR, DataMatrix, and even Aztec codes.

What this misses is that, in a properly-functioning organization, a team that reliably delivers gets a bigger pie to divvy up. If your organization isn't like this, then perhaps you should consider finding a new one.

Set the AllowedIPs wireguard setting (and/or the route, if you can set that separately) to one larger than your home network (i.e., if your home network is 192.168.1.0/24, use 192.168.0.0/23). Then, block wireguard packets from the internal network on your router. Then the tunnel will always be running; it just won't be used when you're at home because there's a more specific route

I know that Len Sassaman had r@ai for quite some time, so your friend didn't have the absolute shortest address. Still a cool one though

The US does identity verification by asking for a driver's license (which has no chip or biometric data) and possibly a series of questions about your past drawn from public data. All of these credentials are laughably easy to spoof. Compare this to Europe, where every resident has an eID containing a keypair and X.509 certificate signed by the government containing their personal details. It is trivial to check the validity of these cards and nearly impossible to forge without subverting either the national PKI or printing apparatus

The USB HID protocol is designed to support basically any device that regularly reports a set of values; those values can represent which keys are pressed, how a mouse has moved, how a joystick is positioned, etc. Now, different devices have different things that they support: joysticks have varying numbers of axes, mice have different sets of buttons, some keyboards have dials on them, etc. So, there's no single format for a report that simultaneously efficiently uses bandwidth and supports all the things a human interface device might do. To solve this, the HID protocol specifies that the host can request a "report descriptor" that specifies the format and meaning of the status reports. This is great for complex devices running a full OS; there's plenty of memory and processing power to handle those varying formats. However, these HID devices needed to also work in very limited environments: a real mode BIOS, microcontroller, etc. So, for certain classes of device such as keyboards and mice, there is a standard but limited report format called the "boot protocol". IIRC, the keyboard version has space to list 6 keys that are pressed simultaneously (plus modifiers), all of which must be from the same table of keys in the spec, and the mouse has an dX and dY field plus a bitfield for up to 8 buttons (four of which are the various ways you can scroll). To implement a more complex device, you'd want to be able to specify your own report format, which the ESP driver doesn't seem to allow you to do.

Calendar apps expose all of the weird edge cases in dates and times. For a specific instance of a specific event that never needs to move, TAI will work. However, suppose that you schedule a meeting for 28 March at 13:00. You then move that meeting forward a week, crossing a DST switch. If you simply add 604800 seconds to the TAI, that meeting will be at 14:00, which is surprising. OK, easy enough to solve; convert the TAI back to the local timezone, add 7 to the day field (carrying into month and year, as needed), and you're good to go. Now suppose that the meeting was set up in the US, and you realized that you needed to reschedule the meeting while you were travelling in Germany. You get back to the US, and suddenly the meeting is an hour early, because DST rules aren't universal.

Worse, imagine a recurring meeting, every week at noon. You've got the TAI for the original instance, but as you cross a DST boundary, some places will shift according to DST and others won't, and thus you have half the attendees showing up at the wrong time.

Finally, TAI introduces the leap second bug for calendars. When you schedule a meeting for noon next year, you want that to happen at noon, not now plus however many seconds. If a leap second is introduced, you don't want that meeting to happen at one second before or after noon.

TBH, if you find yourself in a situation where you need to deal with calendar events, I recommend updating your CV.

Indeed, BFS, DFS, and A* are the same algorithm just with a different data structure to track unexplored nodes. BFS uses a FIFO queue, DFS uses a LIFO stack, and A* uses a priority queue (generally a heap)

One of the things that Data Matrix got right was being able to shift between encoding regimes mid-stream. Many character sets can be represented in radix-40 (so three characters per two bytes), and the occasional capital character can be handled by a shift byte. If you have a long string of digits, they can be encoded in 4 bits/char. You can even put raw binary in there if need be