catblast01

Why does utf-32 not require a bom?

> Infinite self-joins are not elegant in relational algebra.

Somewhat ironic comment given the name of this site.

SQL stinks as a language in many ways, won’t argue there. However for many cases even PostgreSQL does just fine with recursive CTEs (though granted when it fails it does so hard). Oracle does better (and CONNECT BY has been around for over 40 years - someone’s been getting by with it). Graphs in RDBMSes are not uncommon.

Anyway, none of this has anything to do with the original claim. Are graph db implementations useful? Definitely (though often times sticking with the rdbms is the better choice). But they get their utility by being less flexible.

For one, Neo4j is an implementation and SQL is not. As for clarity, SQL, although by far the most widespread query language is a shitty language in many respects... again a separate issue from the relational model.

In any case you’re conflating a storage engine and a query language. There’s no reason an RDBMS implementation cannot be tuned for graphs, and there is no inherent lack of generality in the relational model which seems to be the prior posters claim. Whether Neo4j is a useful piece of software is a different topic.

> Honestly, i think graph db being more powerful is one of their more reasonable claims. Graph db trade off a more flexible data model (which some people would call power)

You can represent graphs in the relational model so that’s simply bullshit.

Matter of perspective I admit but it also could just be a weird pastiche of elements of C, COBOL, M, Ada, and JavaScript. Nothing here seems particularly ambitious really.. idiosyncratic however. A good counterpoint might be Go.

The author used to at least actively solicit feedback, I never once noticed acknowledgement of any of it though.

Nope. The guy posts with his real name on YouTube so this is like TempleOS and the creator is non anonymous, quite active on Reddit, and not very young (as if the style of graphics wasn’t a dead giveaway, like a time machine to the early 90s).

In any event I engaged with the creator before and what I can say is the conversations are entirely one-sided, in fact some of the most I have personally experienced. Will take one or two words from your post or reply, briefly mention it and then go completely off on a tangent only concretely related to those words. I’ve never known them to become outright offended or insulted and yet are completely impervious to any feedback. The upshot is, this will greatly limit a widespread following, because not all their tastes are exactly mainstream.

> not very RISC but extremely useful.. load and store multiple

Who was it that said that R and RISC as a nomenclature was always hogwash. A more apt name would be “load-store architecture”

Onne random thing I remember about LDM/STM was the earliest rev silicon of the Motorola (now Freescale) ARM based DragonBall (then better named the MX line) was the ARM9 core had a bug where LDM/STM would not work with the cache enabled - which of course was horrible, so we hacked gcc to not emit these instructions as a temporary workaround.

This is really low effort to look up from reputable sources if you don’t believe it. Electronics have far higher sales prices than clothing, I can’t even begin to fathom why that is a relevant benchmark.

GNU is heavily skewed to developer tools and infrastructure, and gcc is no counterexample. There are so many classes of software where this does not work. Pretty much anything for heavily regulated industries is not well served by FLOSS. There are few open source insurance software or medical records systems (the few that exist are highly niche and/or limited), EDA/CAD is not well served by FLOSS (I’ve toyed with FreeCAD, but even hobbyists gravitate to Fusion). Outside of developer tooling and infrastructure: commercial, generally closed source, closed development software is king.

* besides the hard part of standing up an EMR is not installing a prepackaged software.

cf Smart Pipe https://youtu.be/DJklHwoYgBQ

Is everyone here too old to remember MapQuest or CD based car navigation systems? Apparently. We had decent turn by turn in the 90s. In the US mapping data was heavily derived mostly from public Census TIGER data, which had flaws but was good enough.

Look up “protein coffee” and maybe try to lay off the snark a bit if you don’t know what you’re talking about.

Seriously though, the article was very well written. I knew it was a cheap shot. Sadly for some of us 1994 doesn’t seem like that long ago.

> This is the oldest and most well-known query execution model, and was first proposed in 1994.

Is the author younger than 25. Not picturing a world of databases before 1994? Did the world even exist?

I kid but regardless of when a paper describing some process was published the “oldest” query execution model is decidedly a lot older than 26 years.

Unfortunately this is always an unrealistic request for anyone with a public profile. Those that have no desire to speculate won’t, and it’s not going to stop the curious. Frankly I tend to agree it is moreover an unreasonable request as long as no one is actively invading the privacy of the family.

Pedantry in this case to be sure but Rs-485 lives quite nicely and cleanly at the physical layer of the osi model.

All the stuff you mention like collisions and addresses and shit starts to be layer 2 and above.

Maybe that is what the issue is? Some, well many, DNS providers provide value add transparent additional features like redirect and email. If their web infrastructure is down then that might not work even though the core DNS shit does.

This is really reaching. Channelization as being referred to upthread in reference to SSH or RDP is a protocol level abstraction, rs485 bussing is at the electrical level for multiple physical devices and also doesn’t specify any addressing scheme, that would be specified as a higher level protocol (that could be done just as well over rs-232). Device addressing and channelization are not even logically the same thing.

Ethernet as it is typically deployed is not a bus either but it is used in channelized communication (packet switched no less) just fine.

Edit: responding to your edit. The point still stands this is orthogonal to the physical or electrical layer. The vast vast majority of the internet including Ethernet and fiber links are point to point and clearly they carry a tremendous number of separate streams, and a small amount of broadcast. You can run layer 2 over an RS-232C serial port, whether its Ethernet, PPP, SLIP or ATM. In Linux PPP is in fact implemented as a tty line discipline.

As for RS-485, there are widely used standards such as DMX512. It is not a strongly supported idea to put this stuff in the kernel, when performance in user space is sufficient.

Finally, RS-232 can do the same thing (multidrop) with electrical repeaters. See MAX3322 eg.

> (A harder problem is Japanese — Japan really got screwed over with Han unification, so choosing Shift-JIS over any Unicode encoding is often best.)

This statement needs more support. I think “screwed over” is a bit harsh, since I’m not aware the impact on Japanese was anymore than the rest of CJK. Despite the Han unification controversy, Unicode has been heavily adopted in Japan. The space requirements are basically the same as all CJK. Half-width kana is heavier since they are one byte in shift jis but they’re relatively uncommon.

How does that make it an RS-232 with channels? You’re confused. RS-485 just like RS-232 is an electrical specification for serial signaling (with multiple access), nothing more nor less. There are USB to 485 converters that basically operate exactly the same as their serial port counterparts and will show up as a tty device in Linux/Unix flavor or a COM port in Windows.

It has some features that zfs lacks such as a dedup story that is actually usable, plus more flexibility in reshaping. Other than its native raid 5/6 story what major features is it lacking in comparison to zfs? I agree the tooling is trash compared to zfs (zfs command line is a delight by comparison). Also stability wise i haven’t noted problems in years. Not to take anything away from zfs which has proven itself in its particular use cases, but btrfs is alright.

That makes no sense of course. You’re innoculating the whole body, especially bone marrow and lymph nodes, not a small area of muscle.

I think I addressed that: “how they’re organized in the spec”. They’re distinct, but outside of internal references in the spec the distinction between language and library doesn’t really matter. To a user of C++ it’s all one spec with distinctions for freestanding and hosted implementations, etc. You can’t have a conformant C++ implementation without elements from both.

The original point of this thread was about where std::format exists and from a implementation point of view it doesn’t really matter.

This is really a false/meaningless distinction though. The C++ spec covers the “language” (it’s one of the first sentences), and that includes the standard library. The spec uses the terminology language and library but there is no formal distinction between them other than how they’re organized in the spec. A conformant compiler could implement the entire STL as a built in if it wanted to.

There is a formal distinction between hosted and freestanding, but that’s not the same as library/language. Cf https://en.cppreference.com/w/cpp/freestanding

I really don’t care for pedantry, but when people start correcting others then had better be right.

I don’t deny any of that. But where is most of the hardware for just about any network and computing equipment manufactured? Questioning if tplink is made in China is pretty low effort and pointless.

A thoughtful analysis would ask why an onshore supplier would fundamentally be any less vulnerable to political adversaries.

> An intel goldmont won’t use much more power and can easily do gigabit sqm and wireguard/IPSec without breaking a sweat. Can any of these nearly 2 decade old MIPS/ARM designs come close? I don’t understand the hype for the hardware either.

It’s just silly to suggest this when quite simply not all IPs behave equally and the differences are not subtle. It’s annoying when “bad” IPs get reassigned and change, but that is a relatively infrequent event - and certainly CloudFlare is at least adaptive. It’s just not a principal aligned with reality.

> I think you forget how many people live in poverty in the USA.

FWIW, I work in public healthcare in Appalachia. In any event, I have no idea what the relevance of this statement is.

I made a statement of fact.. market wages for engineer/software types were in the 50s in the low COL areas 20 years ago - this can be verified - if they're not in the high 80s now then real wages have not grown.

Honestly, the OP comp might not be too too awful. Of course even then it shows wages haven't really skyrocketed either outside of FAANG. 86k - if that's base? - if that is $55k + "equity" then no - especially, they're not public and an 8 year vesting period! The longer vesting period is especially troublesome when you're not paying amazing wages to begin with - that's just how it is. Also, the problem with the bimodal distribution of wages in the US, if you're in that lower mode, you are kidding yourself if you can really maintain any realistic retention - there are too many opportunities for people to jump ship and it doesn't take much to make it worth their while. Another thing not covered by this post - how is their promotion structure? Sure the pay is transparent - but the vast majority of cases if you want a raise at the employer you're at - the best way to maximize it usually to quit and come back later.

Edit: looking at that table in the post and doing the math, looks like ~68k. Where do you get 86k from?

> Tech wages in the US are hugely inflated compared to the rest of the world too.

Sure, everyone knows this - well inflated not so sure - the market wages are higher. What's your point?

Edit: by low COL i mean, like Rust Belt, midwest.. as far removed from the coasts and SV as possible.

Don’t work in tech for a living anymore so no real horse. But wow. Market entry level pay in Low low COL (like not SV or Boston) 20 years ago when I got a job - about 55k. And this was no Microsoft or Apple. I’m aware FAANG and some unicorn packages are better than ever - but the other “half”, wow wages have really stagnated. Half the blog posts on this of course are intellectualization of why this isn’t sad.

Why? Bluetooth isn’t perfect but it the fidelity is not so poor as to reverse the “warmth” of the analog distortion, and even a cheap ADC won’t either... it should reproduce it pretty well.

What people get wrapped up in cognitive dissonance is that they equate “pleasing” with hi fidelity... but anything likeable about vinyl is because it is low fidelity — but if you tell certain people this they take it as a pretty personal attack. Anyway, a decent digital signal chain should reproduce this low fidelity pretty well including high bitrate SBC.