andrewf

They recreated some. https://www.oregonlive.com/portland/2024/04/pdxs-beloved-tea...

There's looottts of the 2015 carpet as well.

Sounds like you might enjoy switching jobs with https://news.ycombinator.com/item?id=42289955 (not because of the "seniors" bit, but the more general thing where they're rewarded for biz outcomes and don't have any tech colleagues)

You kinda want some machines to reboot less frequently than others, so issues don't creep up on you.

You also want some machines to reboot much more frequently than others, so you catch boot issues before they affect your entire fleet.

It seems most likely that the corruption happened in RAM or the local storage device, after the TLS integrity check had already happened.

> So how do we get to the place where we choose to build systems that behave correctly? I don't think we get there without severe penalties for failure.

What failure? The planes work. This is puritanism.

Random thing I noticed last week: there's a version of 2048 buried in the United Airlines iPhone app. Hamburger -> Games -> 2048.

I haven't looked but I bet a lot of the WAL complexity comes down to supporting consistency and durability guarantees, neither of which you necessarily need for your in-browser use case.

Reminds me of the Royal Society's contributions to science - we benefited from a bunch of landed gentry with time to spare, keen to self-actualize. (Stallman had that MacArthur Grant, too!)

I Perceive NSFW content in the middle of this thread :P

Adding to this: I like looking at graphs like https://calendar.perfplanet.com/images/2021/leon/image1.png . In this particular example, the "lzma" (ie xz) line crosses the zstd line, meaning that xz will be compress faster for some target ratios, zstd for others. Meanwhile zlib is completely dominated by zstd.

Different machines and different content will change the results, as will the optimization work that's gone into these libraries since someone made that chart in 2021.

There is a distinct Postgres entity which doesn't upsell, that's true. But looking at https://www.postgresql.org/community/contributors/ it seems like most of the actual people are employed to work on Postgres by companies who sell, among other things, proprietary Postgres derivatives and tooling. Maybe if one company came to dominate the list we'd be in trouble..

Back in the 2000s the Australian government provided software that ran on Windows to prepare and submit your personal tax return. I used to archive my tax return by preparing it in a Windows VM, then storing the whole VM image.

I wonder if 200 KB is small enough that TCP slow start will be the constraint on a download over a new connection, rather than the bandwidth.

BSD was the Unix distribution; BSD and Postgres/Ingres development did overlap at UC Berkeley.

> Autoconf is a perl program

Oof, did not realize.

Looking forward to the point where this can build autoconf. It's great that the generated ./configure script is portable but if I want to make substantial changes to the project I need to find a binary for my machine (and version differences can be quite substantial)

Something I didn't realize before trying one is the different assist settings available. The lowest one just made it feel like I'd slept better and was having an easy ride. The highest assist setting had a really obvious "kick".

The bike I used didn't have a "cruise" setting that moved the bike along while your legs sat still, like I've seen on some bikes.

Apple bolted Carbon APIs onto NeXTSTEP/Mach, and Microsoft bolted the OS/2 Presentation Manager onto NT (https://virtuallyfun.com/2021/05/19/presentation-manager-for...), so I'm not sure why bolting Carbon onto NT would be unreasonable.

Interesting you should mention inputs. One of the things I’ve often found useful to log are the data that are inputs into a decision the code is going to make. This can be difficult to reconstruct after the fact, especially if there is a cache between my code and the source of truth.

Compared to: how much are they spending on dev salaries? On cloud or infra overall?

Agreed you need to engineer the logging system and not just pray. "The log service slowed down and our writes to it are synchronous" is one I've seen a few times.

On "do not consider what to log and when" .. I'm not saying don't think about it at all, but if I could anticipate bugs well enough to know exactly what I'll need to debug them, I'd just not write the bug.

As an engineer I generally want logs so I can dive into problems that weren't anticipated. Debugging.

I get a lot of pushback from ops folks. They often don't have the same use case. The logs are for the things that'll be escalated beyond the ops folks to the people that wrote the bug.

Yes, most (> 99.99%) of them will never be looked at. But storage is supposed to be cheap, right? If we can waste bytes on loading a copy of Chromium for each desktop application, surely we can waste bytes on this.

My argument is completely orthogonal to "do we want to generate metrics from structured logs".

Ask yourself whether you want to optimize the P10, P50, or P90 outcome.

There's a warning when you step off the airplane onto the jetbridge at SFO.

Good luck refusing to get off the plane because you see the cancer warning and decide not to risk it.

> keeping speakers and microphones working outdoors across many different climates

Pretty common for two humans to have trouble understanding each other through the microphone setups, too.

From the file linked in https://www.vogons.org/viewtopic.php?p=534845#p534845, it looks like 'mov fs, ax' may have been 5 cycles compared to 3 for 'mov dx, ax', with no penalty in the other direction? So probably net faster than using memory instead.

But +15 cycles if you're in protected mode. In this case I think you'd also cause a fault by storing values into a segment register which weren't valid indexes into the LDT/GDT.

Quake didn't even present the bitmaps in a consistent aspect ratio or size, it did a 1-to-1 pixel copy regardless of the target pixel shape/size. Quake shipped with a 320x200 default resolution in 1996, with configurable support for a variety of resolutions including 320x240. Over time, most people ended up on a 4:3 resolution on a 4:3 display.

>some committee would only announce changes a few months in advance then people would get very upset.

This does happen, with daylight savings.

Sounds like the COMMON section in BASIC. Although there you couldn't even supply indexes. "Remember that variable order and type are significant, not variable names. The order and type of variables must be the same for all COMMON statements communicating between chaining programs." -- http://qbasicnews.com/qboho/#topic-qb45advr-T0085

It'd be interesting to benchmark this on texts from different languages. English text will be very friendly to the branch predictor because it's all 1-byte codepoints. I think a language dominated by either 2-byte or 3-byte codepoints would be as well, mixing codepoint lengths is what would trip things up.