wrigby

I find it really helpful to actively take handwritten notes, for two big reasons:

Firstly, hand-writing notes without falling behind the content forces me to choose the most relevant information presented and reword that information into a concise-enough form that I can quickly write it down. This helps me actually process the content instead of just hearing it. Mentally, it feels like the information soaks into my brain, rather than bouncing off of it.

Secondly - and this is comically simple, but just as important as the first - it keeps my hands off of my keyboard/mouse, so I don't get distracted by other interesting things on the Internet (like HN) and tune out.

Love this! I had read another paper from Wescott about sampling[1] years ago and loved it, but I didn't realize there's a handful of other great articles[2] that they've published.

1: https://www.wescottdesign.com/articles/Sampling/sampling.pdf

2: https://www.wescottdesign.com/articles.html

Yeah, most geostationary communications birds use analog transponders, even to this day. There's a lot of fancy command and control, but the actual payload is just analog RF circuitry, even using tube-based power amplifiers.

I have no idea about these older satellites, but on modern stuff there's a separate Telemetry, Command, and Control (TT&C) link. Maybe this bird is only transmitting telemetry though?

I feel like "lost" is a bit of a misnomer here - this spacecraft was of course being tracked still, and it only took a couple seconds of googling to pull up the detailed info on the orbit from Celestrak[1]. There's even a nice visualization available at [2].

1: https://celestrak.com/satcat/tle.php?CATNR=02866

2: https://celestrak.com/cesium/orbit-viz.php?tle=/satcat/tle2....

The discoverability aspect of this reminds me quite a bit of working on a Cisco IOS command line. At any point while typing a command, you can mash the ? key, and the shell will print out a list of acceptable inputs given what you've already typed (this example stolen directly from Cisco's docs[1]):

  Router# configure ?
   memory             Configure from NV memory
   network            Configure from a TFTP network host
   overwrite-network  Overwrite NV memory from TFTP network host
   terminal           Configure from the terminal
   <cr>

It's really quite a great experience.

1: https://www.cisco.com/en/US/docs/general/Test/dwerblo/broken...

Oh, that's a really good point!

My approach is mostly shaped by my company's culture (which is definitely not one-size-fits-all): if it's a quick thing to code up, I prefer submitting it as a PR with an [RFC] prefix (I've re-titled my PR to include that), so the maintainer knows I'm not too tied to it.

I'll probably lean towards opening issues before PR's in the future though - thanks!

I usually agree, but two of the three open issues should really just be PR's :\

...

After typing that, I realized that this comment should actually just be a PR[1].

1: https://github.com/anvaka/atree/pull/14

Yeah, I could have been more clear - the BigCo I'm at is still pretty light on process, and teams are free to use whatever project management approach they want. As seattle_spring conjectured, I was thinking more about technical best practices. I've learned many patterns from the software I work with now, and I've been able to grow from working with really smart engineers that have enough experience to be pragmatic without compromising on quality.

An (admittedly controversial) example is repos-per-service vs a monorepo: now that I've worked at a bigco with a monorepo, I would have used a monorepo at my previous company.

To be fair, he actually did join Palantir a couple years later.

There's a LOT of survivorship bias in this article. I personally worked in small companies (30 - 200 employees total) and startups (4 employees, including myself) before ending up at a big tech company, and I would have preferred to do this journey in the opposite direction - that is, work a few years at a BigCorp, and then work for a startup.

Getting to see what works and what doesn't work in the context of a bigcorp is really valuable, and I think this post discounts that. A lot of technical mistakes I've made could have been remedied by simply seeing industry best practices, and there was no way I would have learned these things in school.

I'm not 100% sure, but I'm assuming because of Double Data Rate transfers: https://en.wikipedia.org/wiki/Double_data_rate

I'm not sure about the rust package, but in the hardware world SerDes ("serializer / deserializer") is pronounced "sir-deez."

From what I can tell, OpenIndiana is to Illumos as Fedora is to Linux (GNU/Linux to be pedantic), and Illumos is based on OpenSolaris. So what we have here is essentially a Solaris distribution?

Looks like it's a fork of OpenSolaris[1], which is pretty cool - I didn't realize there was a Solaris variant still under active development.

1: https://en.wikipedia.org/wiki/Illumos

NORAD tracks everything in space, and works directly with satellite operators to reduce collision risks on the rare occasion that it could happen.

On my first read-through I thought 'F5 Storm' referred to F5 load balancers, not hitting the F5 key to refresh the page.

ETA is owned by Swatch group, which owns a ton of watch brands[1], including Hamilton, which I would say competes head-to-head with the most Timex pieces (at least in the $500 price range).

There are other Swiss automatic movement makers though, like Sellita (the SW200 is basically an ETA 2824[2]).

1: https://www.swatchgroup.com/en/brands-companies

2: https://www.christopherward.com/etasellitastory.html

Kolaches have to be one of Texas' best kept secrets. I discovered them after I moved to NYC (from DC) because a friend from Texas is opening a kolache bakery in Brooklyn, and man... what a treat. I highly recommend everyone travel to Texas and try a kolache from a roadside stand.

I think the data locality benefits would hold for one to many relationships too (that is, one parent with many children).

I may have some insight here. As a smoker that's trying to quit (14 days clean so far), a Juul just doesn't appeal to me - I'd rather quit than switch. Apart from the nicotine addiction, smoking involves a lot of rituals - for example, I personally choose a lucky[1], and I love the sound and smells of striking the Zippo that I carried. A Juul doesn't offer those experiences, and cigars are much too heavy for me to smoke daily.

This scene from Frasier is spot on: https://www.youtube.com/watch?v=ptqRrwvYVTs&t=126

1: https://www.urbandictionary.com/define.php?term=lucky%20ciga...

Yeah, I take issue with that 150GB number. For one, it doesn't take into account browser caching, and if you're trying to save origin bandwidth, you're better off making sure you're using your CDN as well as you can than reducing your JS bundle by 30kb. I'd be very surprised if there was actually a 150GB drop in total data transferred to users over the month.

I only have a cursory knowledge of DSP, but it sounds like this was implemented by using the 5 second clip to build a matched filter kernel. This would require convolving the samples of the 5 second clip with the entire song to find out where it matches, which should work super well, but I think it would be really tough to scale this up to millions of songs. Is this how it was implemented?

Microsoft's UI toolkit in Windows 95/98/2000/ME/NT4/XP did a fine job of letting me know that I clicked on a button, and it added no latency to the interaction.

On mobile, this is more difficult - the user's finger is obscuring whatever button they hit from view, so I kind of understand the "ripple" after the button is released (and ostensibly the user's finger has moved away, thus allowing them to actually see the ripple).

Still, unnecessary delay is infuriating on desktop apps. I agree feedback is important, but I don't accept the premise that latency is required to provide feedback.

Well... maybe. You would just register the domain once (maybe `mycoolawesomehome.net`), and then you have the authority to arbitrarily decide that your printer's hostname should be `thebestprinterever.mycoolawesomehome.net`.

But now you can just use `mycoolawesome.home.arpa` without registering it.

I think it's the "Joint Comprehensive Plan of Action" [1], which seems to be the official name of the Iran nuclear deal.

1: https://en.wikipedia.org/wiki/Joint_Comprehensive_Plan_of_Ac...

We used to use RFC1918 (172.16/12 IIRC) addresses for the communication between internal nodes in a cluster-in-box system that I worked on, which worked great until we had a subnet collision on a customer's network. Leaves me wondering if link-local (169.254/16, fe80::/10) would have been a better option - while technically the customer could decide to make the external (customer-facing) network have a link-local interface, the chances of that configuration actually happening are pretty slim.

I'm still not entirely sure what the best option is there. Maybe some clever use of network namespaces, with a named pipe to bridge between the "internal" and "external" universes? Just typing up that idea makes me cringe though.

I could understand this if "America" was a singular entity. In reality, it's much more complicated than that. It's American filmmakers creating a movie about how a war, which American politicians chose to fight, effected American soldiers.

Attributing the collective actions of every single person from the US to a single entity called "America" and then pointing out that these actions are inconsistent is a silly reductionist argument.

Edit: I should also add that I strongly think the world would be better if the filmmakers were creating movies about the effects of war on the _other_ side, not just their country's side.

I'm not up to speed on ZBLAN at all (had to google what it is just now), but if it can really do 0.05dB/km at 1550nm [1], then it seems like it's actually ideally suited for long distance telecom. This is like taking an off-the-shelf "80km" SFP, and turning it into an "800km" SFP. Am I missing something?

1: https://en.wikipedia.org/wiki/ZBLAN#/media/File:Zblan_transm...

The satellite antennas used to uplink a video feed like this have a very small beamwidth (around 1° depending on antenna size and the frequency they're using), so you need to keep them pointed precisely at a specific point in the sky. They rely on a sensitive IMU (accelerometers) and fast motors to actively track the satellite while the platform moves. They're quite impressive to watch in action[1], but I'm guessing they're no match for the shock and vibe of the rocket landing.

1: https://www.youtube.com/watch?v=k79vfwUkeYk

Yeah, they've made it a point to emphasize that they're working on crypto / bitcoin related things[1].

1: https://www.investopedia.com/news/overstock-goes-all-bitcoin...