I can see why Feynman became sexually promiscuous afterwards, undoubtedly to numb the pain of losing his wife; seduction allowed him to have a form of that connection, albeit without the depth and love of what he had with his wife. While that path looked appealing to me, I do not regret avoiding that temptation.

He married again in 1952 and divorced in 1958 (and married again a couple of years later).

I don't know if sexual promiscuity was part of the divorce complaint but apparently this fragment was in it: "He begins working calculus problems in his head as soon as he awakens. He did calculus while driving in his car, while sitting in the living room, and while lying in bed at night."

why? i dont follow. is academia so special there?

Opposed to "industry" -- where you're under constant pressure to figure out why some random docker container pull fails or mutating web hook won't let your pod be deleted. It's just not fun or meaningful. It's a slog and the needless complexity is doing nothing but increasing.

I think these things can all be true in industry though. I say this as an industry researcher myself that has struggled with work/life balance my whole career. In the more soul destroying industry jobs hopefully you can at least clock out at 5pm and pay attention to other interests, i.e. you don't have to care about work outside of work hours.

I actually think it's more to do with the person than the job though. It's about how much you value work/career achievements and recognition vs family vs hobbies. Caveat: some people genuinely have to work unreasonable amounts of hours to put food on the table for their family and that's unfortunate.

    > "He begins working calculus problems in his head as soon as he awakens. He did calculus while driving in his car, while sitting in the living room, and while lying in bed at night."

building close friends as an adult is a known Hard Problem when everyone has their own lives. any big insights or breakthroughs you can share for the rest of us?

And lots of flexibility and forgiveness (for small things).

It won't always work. I've done pretty well at this tbh, but it has helped living in a city like London where soicalising is one of the main activites and easily facilitated - to an extent at least. Context is key, no doubt.

Marathon, not a sprint. Everyone's busy so gradually making yourself a mutual part of another person's life will take time.

I try to steel myself about the coming eventuality, but all I know is it's going to be a deep abyss and it will take everything to climb out of it.

When it happened it was the worst emotional pain I had experienced, it does get better in time. Your feelings are valid. You don't have to feel whatever anyone else thinks you should feel. It feels cliche that it gets better in time, in my experience tho it does. I'm sorry your dad and family is going thru that. Oddly I found Cyberpunk 2077 to be therapeutic on coming to terms with that situation. In the game you play a character who essentially has a terminal disease, there are a lot of moments in the game that rang very true to the emotions I experienced.

When your dad eventually departs, you are allowed to feel anything. I have been the type to laugh at a funeral remembering something funny the departed said or did. If you feel like crying that's ok too.

What's odd is that I still don't feel the need to look at photos of him. Like, no need whatsoever. I don't have a photo of my late dad on the wall, even though we had a truly excellent father-son relationship with a lot of warmth. The true memory of him resides in my head; and what's more, I notice myself replicating his exact body language at times. Explicitly realizing this always feels funny, almost like I want to express gratitude to him for "inheriting" those movements to me :). I think the bonds still hold strong, and they forever will, but, oddly, there's no need whatsoever to visualize it with a photo. It's simply much deeper than that.

On another side, I often miss the ability to ask my dad "factual", practical questions. E.g. -- because he was a construction engineer -- how to do some specific thing in grandpa's house renovations; how to position a beam, etc. During the first months after his death, this was a common recurring thought: I'm stuck while building something, but "hey, I'll just ask dad tomorrow how to do it right; yeah, I'll just call him" -- following by "oh, I forgot I can't". After this, there was always an interesting feeling of emptiness or standstill -- not sadness, just a really deep understanding that, no matter what, life moves on. And I still get to live mine. I actually came to enjoy those moments of emptiness.

Thanks for sharing your story, it is very moving. Grief will have different stages; to me, it has mostly been really deep introspection. Take it easy, mate.

If that were to happen to me, I definitely want my wife to love and be loved again.

I think this is amazing. He wants her to lead a happy and full life. I for sure would want that for my wife if I was to die first. I love her to bits and would want her to be happy.

The idea of a couple that belongs together beyond death might sound romantic but is ultimately too tough on the one left behind, especially if the loving couple are separated with the "leftover" still being relatively young.

(On a side note, I'm positively excited to read a quote on HN that dates from about 480 before Christ. It surely is still relevant two and a half thousand years later.)

And why not? Who knows the remaining spouse better than the departing one? Who has better insight than that? It seems like a deeply loving thing to do.

Even though her death was sudden and there was no chance to prepare for it, it makes me smile to know my sister's widower married someone who knew and loved my sister too. And it's probably easier for the second spouse this way too, because she does have context for her husband's grief. She also grieves the anniversary of my sister's death, obviously not in the same way, but with an understanding that someone who didn't know my sister would struggle to have.

To me, that is the saddest thought in the world. My sincerest hope is that he would find someone else he loves as much as me, and that he would tell her stories about me and let me be a happy memory rather than a sad one.

My uncle's wife remarried years after my uncle died after a long illness. She married a man who, himself, had been similarly widowed. She wore a gray wedding dress to indicate that she had been married before, and they placed two beautiful photos of their late partners behind the altar where they made their vows.

It was really beautiful and inspiring to see them carry on with their lives and love again, while proudly bringing along the memories of the two people they had lost and still loved as well. When I visit them at their home, it is filled with pictures of them with each other and with their late spouses, all intermixed, and it's deeply comforting to me. Even though they never knew each other before they were widowed, they tell me that it feels like they know each other's late spouses as if they had all been good friends together.

If I were the one dying of a long illness, knowing that something like that was on my partner's future would give me a more peaceful death. I wouldn't feel like I was letting him down and ruining his life — just saying goodbye to him in this chapter of his life, while finishing up my own chapter.

Everyone is different, and you don’t know what or how you’ll feel unless it happens to you.

For example, at a point, I felt guilty for not being sad. My son and I went to the beach and had a great time. I felt an intense guilt afterwards. But I focused on fitness and other things and tried to make each day better. Time and positive work heals.

In terms of love, etc, the heart has room. Some people are ready before the body is cold, others take years to contemplate a relationship. All i can offer is that as someone who has walked a mile in these shoes, I’d never cast a negative judgement on widower/widow behavior that isn’t plainly reckless.

Why?

I think it's important to distinguish these situations because I think there are a lot of people going through a lot of emotional pain because they believe that any sex outside of a long-term monogamous relationship is morally questionable, and it's not. The challenge is that insuring that all parties have given informed consent is hard, but the emotional toll of throwing in the towel on this can be pretty high, so I think it can be an effort worth making. Certainly it's worthwhile telling people that it's OK to try.

    > I have about two dozen very close friends (both male and female)

Sorry for your loss and glad you’re doing better.

If you believe you have a moral duty to be the best person you can then passing up that opportunity in favour of the facsimiles is morally wrong.

Maybe a better statement would be that casual connections only offer facsimiles of the depth of a deep and long relationship, because the sheer accumulation of time and experiences with another person cannot be replaced quickly. However, you can also say that a deep and long relationship only offers a facsimile of the breadth of many shorter relationships, as one person can only offer so many things. If you change your passionate pursuit every decade, finding a partner with the same one each time might make that pursuit more fruitful. It's not how I personally feel, but "best person" is subjective like that.

May I suggest retiring to The Villages in FL? They seem to have the multiple partners thing worked out.

https://duckduckgo.com/?hps=1&q=the+villages+florida+sex&ia=...

That's the sense I get. Back in early 2011, there was still an issue where crappy DNS servers would sometimes respond to AAAA record requests with a "Server fail" error. Because of this, I had to tweak my recursive DNS server to treat a server fail to an AAAA request as if the AAAA record did not exist. This made IPv6 resolution more fragile, but it had to be done because of the state of the internet in early 2011.

I finally removed that "feature" here in 2022, when someone asked why I handled server fail responses to AAAA requests in this unusual manner.

But, over the medium term, we will probably hit a point where unencrypted DNS stops being mainstream, just as unencrypted HTTP by and large stopped being used in the 2010s.

This means that Deadwood (the caching/recursive DNS part of MaraDNS) will grow from being a tiny, efficient, 71680-byte server to being something a good deal more huge (TLS, HTTPS, etc. are really bloated compared to good old DNS-over-UDP).

While I don’t mind YouTube ads -- I have disabled ad blockers on that ___domain because I’m an electronic musician myself so strongly feel it’s important to support artists and musicians -- these Amazon ads are in my book unethical, since they look like legitimate results with only a light grey “sponsored” in small text underneath them.

Don’t get me started with how Amazon pesters me to join Prime every time I order with them.

I know, time_t is 64-bit with pretty much any new Linux distro out there, so why are people seeing Y2038 issues? It’s because the Windows 32-bit POSIX compatibility layer handles Y2038 very poorly. Once Y2038 is reached, the POSIX time() call in a 32-bit app fails with a -1. It doesn’t use a rolling timestamp somewhere in 1901 the way 32-bit Linux applications with 32-bit time_t do. It fails hard, returning -1 for every call to time().

Now, it’s true that Microsoft does have proprietary calls for time and date which are Y2038 compliant, and, yes, native Windows32 apps should use those calls instead of the POSIX ones, but in the real world, it’s sometimes a lot easier to, say, just use stat() to get a file’s timestamp instead of having to use CreateFile() followed by GetFileTime().

This is why a lot of Windows apps are still seeing Y2038 issues.

In terms of Linux apps, the Y2038 stuff is mainly seen in old 32-bit binary only apps. Since that stuff is mainly games, where an inaccurate datestamp isn’t a serious issue, I think we will see emulation libraries which give old games a synthetic time and date so they aren’t outside of the Y2038 window. New apps will use a 64-bit time_t even if compiled as a 32-bit binary.

HB567, a larger child welfare reform bill, and clarifies that parents are allowed to let their kids engage in independent activities as long as they aren't putting them in serious, likely danger. [...] Unfortunately, HB567 amended only family law, not criminal law. This meant the cops were still free to punish Wallace.

The solution is to create a law like HB567, but that amends Texas’s criminal law so frivolous arrests like this one can not happen.

Another thing would be for more states to have well defined laws which give children a minimum age when they can be left unsupervised.

Related: https://reason.com/2017/03/30/mom-arrested-for-leaving-kid-a... — https://reason.com/2021/10/11/cop-family-lawsuit-midland-qua...

The biggest issue for long term cell phone support is, even if we get an OS with a 10-year security update timeline like Rocky Linux, will the phone itself be able to make calls on whatever cellular networks exist 10 years from now? I have a number of 3G phones I bought as recently as 2018 which became paperweights in 2021 when all of the cellular telcos in the United States stopped supporting 3G, forcing me to update to a 5G phone. Is 5G going to still work in 10 years? Or are the telcos going to continue to convert perfectly good phones in to landfill?

As someone who has a 15-year-old laptop which is still a perfectly good Linux server (its screen went out two years ago, but it was a perfectly good desktop computer until then), it’s annoying seeing phones I bought less than six years ago be useless on today’s cellular networks.

Now they have a phone whitelist restricted to (mostly) mainstream brands...

What's the reasoning behind doing this?

Anyway I got a Pixel 5 and now use T-Mobile’s instead of AT&T’s network (via Ting, since I use little mobile data and that gives me a phone with a low-bandwidth data plan for $15 a month + tax), since T-Mobile is better about supporting legacy phones and protocols.

In terms of the “waste of spectrum” argument, I think it’s better for the earth to “waste” some spectrum than waste millions of perfectly good phones. One takes up landfill, the other doesn’t. But, to each their own.

The problem with long term support for Android devices is absolutely short security update windows, i.e. short term support for Android devices. Pixel comprises a tiny fraction of Android devices sold in the US, and a miniscule fraction worldwide.

The planned obsolescence of today's mobile phones makes me sad.

However, in case of mobiles, things are particularly awful because it is often not trivial to install and maintain an OS other than the one supplied by the manfucturer.

It's incredible how much electronic waste and security issues are generated by lazy manufacturers who do not mainline drivers into the Linux kernel.

An os that will let an old device live on as a wifi only device is sorely needed.

But unless I'm mistaken, if a device only has 3G, and you want to travel with it away from WiFi, you can't even get 3G data on the device, as they shut off the 3G networks in the USA. Best you're gonna find in some rural areas may be 2G or 2.5G.

5 years of updates is still a short time. We only have one planet...

> The biggest issue for long term cell phone support is, even if we get an OS with a 10-year security update timeline like Rocky Linux, will the phone itself be able to make calls on whatever cellular networks exist 10 years from now?

I don't think this is the biggest issue : operators usually maintain a certain type of carrier for at least 20-30 years (in France, we are only talking about shutting down 2G - which still raises a lot of issue because of the many IoT devices using GSM...)

The biggest issues are IMO

  - lack of parts to repair old phones
  - no possibility to manage bootloader keys / relock the bootloader (not even mentioning devices with locked bootloaders)
  - "stable-api-nonsense" ideology and no BIOS/UEFI/ACPI for smartphone => no way to have "one firmware to rule them all"

With an Android bootloader UEFI app on top to mimic what Android wants on top...

UEFI alone isn’t exactly useful when you need per device kernels (and associated modules). It’s just one slice of the problem.

I'm fairly certain that some phone models out there are probably better in this regard, especially if you can swap out the battery, the components are of decent quality and they're built in a rugged way. Also, somehow the idea of repairing your phone has gone out of the window, since you can just buy a new one.

> Your laptop may be good, but it will probably only support some ancient insecure slow WiFi profiles.

For many, slow Wi-Fi is acceptable. Same with a dated and slow CPU/RAM, which many will still consider better than their devices becoming e-waste. If the OS wasn't so locked down, many would enjoy having a stand-in for a Raspberry Pi for all I care, since the small form factor of a phone would lend itself nicely to DIY hacking, especially because of included camera and networking. Even if the hardware is lacking, that doesn't mean that we shouldn't or couldn't support the software for longer amounts of time.

Your point about the hardware itself and what it supports becoming insecure is a good one, but there's no guarantee that the amount of time for something like that to happen would be much shorter than any improved OS EOL period. Of course, if it's some non-critical functionality that's insecure, it might as well be turned off in software, like older versions of TLS in web servers.

> Distributions start to drop old hardware. And there are always new CPU and other chip security bugs discovered, how do you get fixes for that into your system?

We could cross that bridge when we actually get to it, and try to figure out the things that are easier to do first and foremost: notably software support. If something like Ubuntu LTS has an EOL of 5 years and AlmaLinux has security updates for 10, I don't see why Android versions should be any different, unless governed by a profit oriented corporation.

Aside from that, it's surprising that 3G can just be tossed away like that on a national level, since the amounts of e-waste this would generate is kind of staggering, even more odd is the fact that in many places 2G is still in operation. I guess at least that is a bit of a silver lining, if the claims were to be true (citation is needed, but it sounds like a sane argument): https://en.wikipedia.org/wiki/3G#Decline_and_decommissions

> Technology that depends on 3G for usage will soon become inoperable in many places. For example, the European Union plans to ensure that member countries maintain 2G networks as a fallback[citation needed], so 3G devices that are backwards compatible with 2G frequencies can continue to be used.

It’s nothing like Bash; it’s a Lisp dialect.

The obvious comparison is Lua, which is generally the standard embedded language people use when a scripting language needs to be tiny. I compared its size; Zuo is 55% the size of Lua, so apples to apples, Zuo is only a little over half the size of Lua. On the other hand, Lua uses a Pascal like syntax which is a good deal more Bash like than Zuo.

Going back to the Bash comparison, Bash is huge but the subset ash is pretty small, somewhat smaller than Zuo. Busybox’s AWK is, as I recall, about half the size of Zuo. We can have a POSIX compliant scripting setup (ash + AWK) using around the same amount of space Zuo takes up.

Using Helvetica in a font stack with Linux has issues because many desktop distributions use the Ghostscript Helvetica equivalent font which has vertical alignment issues. Other fonts are often times substituted with different fonts which both change the look and metrics of a web page.

The reason why compatible metrics is important is that elements using lettering have the exact same size on a web page, so there aren’t issues with a font overflowing or otherwise looking wrong when it looks fine with Arial. Substituting Arial with Helvetica might break how things look in a web page which assumes Arial metrics.

Personally, I wish every computer out there (or at least every browser in widespread use) came with Noto Sans, Noto Serif, and Noto Sans Mono so that something like “font face: Noto Serif” would look and act the same way everywhere, but Arial is as close as we get to that.

Second, you are overstating the importance of metric compatibility on the Web. Web pages aren’t built with pixel-perfect character placement (unlike PDFs or some GUI frameworks) - text is flowed by the browser, and the web developer cannot specify nor depend on the precise appearance of it, as the appearance of text may differ depending on the browser engine, underlying OS, system scaling, user zoom, accessibility settings etc.

"Hello" in 10px Arial takes up 29x10 px on a Windows 11 PC (in Paint), but it might take up a little more or little less 32 px on Linux, or macOS, a different Windows program, or a 150% scale display, due to antialiasing/hinting/rendering technique differences. On a webpage, you wouldn’t make a button containing the word "Hello" precisely 29x10 px — the button would have more breathing space anyway, or it might just have a margin/padding instead of width/height set, and the layout will work even with tiny rendering differences.

Also, while a handful of very-badly-designed pages may break if you replace Arial with Helvetica for aesthetic reasons, the bigger crime they simultaneously commit is that they won’t work if you replace Arial with OpenDyslexic for accessibility reasons.

That’s may be true for Firefox on Android, which has, as per caniuse.com, a 0.28% global usage share. Chrome for Android, which has a 40.28% usage share (well over 100 times the users as Firefox/Android) shows things without serif.

There’s also a number of uniwidth typefaces which do not get wider in their bold form, for GUI elements which need to stay the same size (a use case which we should not completely dismiss as bad design), but those need to be downloaded as .woff/.woff2 fonts.

Large is relative; the 116,472 bytes my font stack needs has five fonts: Roman, bold, italic, sans serif bold, and sans serif bold small caps. Mind you, this only supports English and Spanish, but it does have smart quotes, an em dash, and other typographic characters: “”‘’— etc.

Then again, 116,472 bytes is a drop in the bucket in an era when people by and large consume video content online.

Arial is designed with the same metrics as Helvetica because Arial is a copy of Helvetica.

Replacing Arial with Helvetica is just restoring things to the original-not-encumbered-by-licensing-decisions standard.

Some people may prefer Arial due to familiarity etc (a distaste for quirky capital Rs?). Some people may not care. But Helvetica is the original.

Mostly true, as per https://wiki.archlinux.org/title/Metric-compatible_fonts but Arimo is designed to have the exact same metrics as Arial, which is important for situations where one is testing a web page/app and can only avoid testing a different font if it’s guaranteed to act in the very same manner. I don’t think Apple guarantees that their Helvetica is compatible with Arial to that degree. Using Helvetica in a font stack on a web page[1] isn’t practical unless one discards that 1-2% of Linux desktop users out there, since they will often get Nimbus Sans which has a different vertical alignment.

In terms of the power users with custom CSS which replaces all instances of Arial with Apple’s Helvetica, they are probably clued enough to understand it might screw up the alignment of an element tied to a font’s exact metrics on a web page. I agree it’s unlikely, but I also remember when clients insisted on “pixel perfect” HTML/CSS renderings of their Photoshop-generated layout template.

[1] If they have a Helvetica implementation as a .woff/.woff2 web font, this doesn’t apply. I’m assuming that this is a web page which doesn’t have any downloaded fonts.

Pixel perfect is tough, but a quick lookaround suggests that the two fonts should match glyph widths at least. There are other metrics to consider though, for sure!