Thanks! Here you can find more talks on various aspects of gaming industry. Hope it helps :)

The same happened to me these days

I bought a new Lenovo monitor a couple of days ago. So, I've have founded at least 3 bugs in its built-in menu. Insane...

dreadful

My Galaxy A31 doesn't even have a functional touchscreen when in landscape/horizontal mode. This happens randomly, for no reason, and sometimes it fixes itself, sometimes not even a reboot fixes it

now america and china in crisis mode over taiwan, maybe we ain't seen nothing yet

Lol, that comment aged well. The situation got even worse, now it's energy too.

@@switzerland yeah brace for winter

@@5cover Time to buy calls for blanket companies.

lemme guess - communism? Because that worked out so great. In world of scarce resources, you need to allocate them effectively, and to do that, you need to be able to calculate, and to do that, you need prices and to do that, you need - you guess it - money and free market. Competition is part of human freedom.

Exactly this, it's publishers and the entire economic system that drive this downward spiral. It's actually an established (and pervasive!) concept, called the "Race to the Bottom", where everyone keeps cutting corners to make the product cheaper and cheaper, so the end product is as mediocre as possible while still making the sale. It isn't unique to software and programming, it applies to virtually every industry right now. If only more of the commenters here would see this bigger picture.

True. Thanks for sharing this with people. More need to know.

Nice comment. I can't help wondering if people aren't *always* talking about decline, and when it actually happens then you can afterwards say "see, they knew" (forgetting all the times people were wrong). Just a thought though. Cheers!

I think I see what you mean: with all the people talking, of course someone was right even if they got it by just luck. Well, the one writer I'm thinking of the most is Sallust: he used to be a politician during the rough times of civil war before August took power. He in particular wrote two monographs, "De Catilinae coniuratione" and "Bellum Iugurthinum". With these two pieces he tells the history behind two fairly recent and important historic events, while doing so he tries to understand and explain why such terrible events (one involving a Roman that after losing elections wanted to take power by force, the other the history of a rulers that became increasingly corrupted along with some important Romans) were happening increasingly more often. In the end Sallust comes to the conclusion Rome's morality is doomed and there's no solution to this constant decay. Similarly think some extremely important writers under the peaceful time of August: they recognize August finally brought peace, some have the illusion this might last forever (we're not sure if this is propaganda or they actually thought that) but some still think the base problem that lead to the effective death of the Republic hadn't been solved. It has to be added that bond of them ever thought Rome was going to end though. It was common to think history was cyclic, but as far as I know they all believed Rome was going to last forever despite all these moral decay (which they thought was something affecting the whole world probably).

@@asgerms I think that's called "survivorship bias" :)

Great questions you've raised. How can we know if the doomsayers are right or if they're just grumpy old men? Same way we can know most things: good measurement. If we follow good scientific method we should be able to measure the decay. Unfortunately science is declining in productivity in a similar way to programming, but that's another topic... One such measure I think is very strong is demographics. If a population has less than ~2.1 children per woman, it is in decline (in the developed world - it's a higher factor in the rest of the world where mortality is higher). This is a problem because it creates an inverted demographic where there are more elderly than youth, so there's a steadily increasing burden on supporting older, less physically capable and healthy people. Japan is seriously struggling with this and the rest of the world is expected to follow - China isn't that far behind Japan (efforts to reverse the cultural norm created by the one-child policy have failed), most of the richest countries are only over 2.1 thanks to immigration, India is levelling off and expected to stop growing by 2050 (even Africa, the one remaining young continent, is expected to stop growing in the 2070s according to the most recent UN report). COVID-19 lockdowns are a good example of this burden - many governments are spending in a way that will take an entire generation to pay off - for the sake of an attempt to lower mortality mostly among elderly. It's hard to argue with these trends, and in itself it's enough to cause collapse because it's a accelerating spiral. Many people and especially this generation of the Green movement seem to believe that there's a problem with overpopulation instead - by their logic 100 million people in the world was overpopulated according to the technology of the time, which relates to the next point - the limit of effective technology at this stage is more due to mismanagement than physics or actual resources. Combine an aging population with the problems of increasing generation gap (highly related to infertility), increasing inequality (which also relates to the generation gap as older people have more wealth, you can see the problem in many countries with housing prices), increasing systematic corruption and government power, increasing mental fragility (suicide rates and mental disorders) - all these things are measurable. Now add the increasing complexity and consequent fragility this talk is about, you can see how all these things are interrelated and self-reinforcing, and it's hard *not* to see a problem looming that can't be fixed without either collapse or some other massive and disruptive paradigm shift.

Counterpostulate. In the early days of computers we were solving the low hanging fruit, the simple problems with simple use cases. Then as this those get solved, we move on to more complex problems and complex use cases. The stuff that was done first is obviously the most important, that’s why it was done first, and if supplies a foundation. But it facilitates the use of more complex systems. The same exact thing happened in physics, math, academic computer science, and basically every other academic discipline.

would you care to share the link to it? love to watch it!

@@dhfromkorra Can't give you a direct link right now but look up Eric Cline The Collapse of Cities and Civilizations during the Bronze Age. Any of the videos around or above an hour are what you're looking for.

e.g. this one: kzfaq.info/get/bejne/mLiTqJCq2NXHiGg.html

Look for Resilience Conspiracy. Resilience - obviously, Conspiracy - because we likely need to do this *despite* what management or monopolists want

@@alexnoman1498 Where do we look for that, google is useless

It's been a year. How has it gone?

@@dandymcgee He died. I mean, he didn't respond to your comment - it's the only logical conclusion.

@@DanielDavies-StellularNebulla plus his name literally starts with "Corpse". RIP in peace

Sometimes I will reboot my car so the radio works

You got me interested. How did your database system project go?

@@etmix7722 It was great to develop, but unfortunately the app didn't gain much traction. For some reason, to me, it's incredibly satisfying to optimise a search function. It has kinda changed the way I develop databases though, I still have to use SQL day to day, but I tend to use it differently now - more transactions and internal data structures, less 'connections' if you know what I mean.

@@stunthumb Sounds like it was a success after all! I see what you mean with less 'connections'. Decoupling the data, in addition to the logic, sounds like a great idea for keeping a project maintainable. Large database can be a nightmare do modify in the later stages of a project is everything is connected to everything else. That's one of the reasons why few people are using database triggers etc. anymore (and that's a very good thing). Thank you for satisfying my curiosity!

@@etmix7722 It's good to chat with someone who gets this stuff, pretty rare actually these days!

Too bad he's an idiot and his different path will lead to madness.

@@isodoublet This is probably not a good idea, but please explain in great detail why you think that.

I wrote another comment here where I explain in detail a _few_ of the wrong things with his line of thinking. Suffice it to say JBlow's ideas of what "simplicity" means are precisely backwards, and if he had his way we'd be quickly back in the win98 days where your computer blue-screened once a day for no apparent reason. But I guess that was "more reliable".

I read that post and really don’t agree. If duplicating shared libraries bound at compile time we’re taking up so much space, then why were Delphi’s executables a few 100k and now a Java or .Net program is megabytes on top of gigabytes or vm?

@@gedbyrne8482 Static linking is gonna be great, especially when the libraries were created to be modular and don't contain too many interconnected dependencies. It's gonna be a lot easier to make programs that use them faster, too.

I suspect it started when internet updates became a thing. "Ship now before the the other guy. Patch with update." Before internet updates, software shipped on a disk or disc and had to be properly debugged before mass distribution.

Maybe it happened as the masses became users. Half the time their ignorance led them to believe software was bad, and when they became less ignorant software could become more problematic without them noticing. Complaining about software became like complaining about traffic. It's practically a badge of honor.

I’m sure they have at least a dozen engineers, but even with that kind of ludicrously low-ball estimate, would you say the emerging features of the product reflect even that number of engineers?

I’m sure they have at least a dozen engineers, but even with that kind of ludicrously low-ball estimate, would you say the emerging features of the product reflect even that number of engineers?

Yep. And some people expect to hear only something they actually watching. It decided to serve neither of you! lol ))

I hate how when I turn on/ off a monitor connected via display port, my PC stops audio playback for a few moments.

I hope you figured it out

please purchase the "new document" DLC

Reminds me of why I largely stopped using Adobe software. Simply couldn't install it. Adobe support told me to reinstall Windows! I was like "Fuck that, I'm going to design my next resume using CSS grid instead of Indesign". Turned out - way nicer workflow, since making changes is much less of a hassle. Fortunately, Lightroom somehow still works. I'd be screwed otherwise when it comes to photography (if you think that there are good alternatives, I suspect that you're not doing photography on a professional level).

notthere83 : oh so you mean me using MSPaint for photography is just not good enough for you eh?

​@@LKRaider I'm not sure that using MSAnything is ever good enough.

Dijkstra was my first software hero - way back in '78.

Microsoft is spyware. Were it to accept genuine articulate and expansive bug feedback not just passively capture metrics of everything you click for GOMS analysis, then it would be more ambiguous and not so easy to assert Microsoft is spyware, but Microsoft is spyware. They aren't gathering info for your benefit.

Not really quality has an upfront expense and serves as an insurance product to recover from major software and infrastructural issues, gives a programmer or teams the ability to recover and is more forgiving of mistakes during the exploration process. Modern programming emphasizes hyper efficiency at the cost of fragility hence it's more prone to large scale losses in output, software resources and is not welcoming of changes(incessant compulsion of having to predict the future with little margin of error) which reflects itself in the technologies. The cost down the road is extremely expensive for low quality. It's just someone in the industry convinces companies that they can guarantee the future and forces their warped paradigm of tech development, deployment and product quality to everyone else.

Meanwhile CEO at big tech are laughing their asses off.

The problem isn't to be solved from the bottom up, because the problem is capitalism. If a company has demand for developers but they don't want to pay engineers, they'll happily give the job to a developer who doesn't care about engineering principles.

100% this hits me like a truck, it feels like un-progress

>the problem is capitalism yanw

@@luiznsjsbsuhsbahahz4238 yawn? Otherwise I'm unfamiliar

I wouldn't worry about that)) If John is right, they will eventually go to nowhere and will be forced to start from the beginning. If John is not right, so... we'll successfully get the better world. Unfortunately, we cannot do anything about this situation. The world he was talking about - is the world, which were ruled by ideas, now we have world is ruled by money at first, so we just cannot return to the "good old days"))

Their argument is: "shut up idiot, my way is the best!"

And asking if you have 5+ years of experience in an environment that is barely 3 years old.

Well, those things might be an adaptation to another problem.

Sometimes I think recruiters make developers go through a shitty recruitment process with bullshit questions and tasks just so they will be discouraged from applying to another job later on and going through the same process all over again.

except for when it isn't, because that forum comment got deleted, or got shared in a DM, or, or, or...

very good point

People need to problem solve, people need self discipline. The thing you make should not inspire, it should be a vehicle toward inspiration. Illiciting a good feeling can be harmful, so can solving someone elses problems.

I think software HAS become less reliable/more buggy in the last ~15 years because there is always a route to fix it via the internet. Pre-internet if you shipped a show stopping bug in an application or device it was a costly process to solve. Sending out update disks or recalling hardware. Since the internet became ubiquitous there is a cheap route to solve almost any program or product failure. We see this now with the "50gb" release day patches for games, and if you ship a device with buggy firmware then you just push an update. This is a luxury unique to software and contributes heavily to what Jon is speaking about. The ability fix bugs post release cheaply means there is no pressure to "do it right" the first time. There is always a second chance. This attitude compounds until you end up with bad designs and fixes on fixes that increase complexity.

@@phatfish001 Something like this has happened in the film industry, too, though a bit earlier. There didn't use to be the possibility to "fix it in post".

No, he's right. Other industries use the legal system to solve this problem, but it doesn't have the same weight in the software business.

Indeed - I think that's a key point: one has to think beyond oneself to see any decline, which require effort, willingness and openness. Most people just keep themselves busy in their own world in the present until they're gone because it's the most logical and natural thing to do for your own sake, but the problem is that it doesn't advance or even preserve civilization as a whole unless they pass on their experience in the form of knowledge to others AND that said others accept and understand the value of the knowledge enough to pass it on further (and this cycle must never stop). I think, to advance/preserve civilization we have to fight against our natural instinct to care about ourselves in the present in favor of caring about everyone else throughout all of the future.

@@ReFreezed hit the nail on the head. The cause of all this is generational myopia. As a culture we are no longer oriented around the next generation, except to exploit them with marketing, rent and interest. This is a result of the shift of the center of society from the household to the corporation that came with industrialization. I'm not convinced that shift is necessary or inevitable, any more than the shift to insanely complicated and slow code is inevitable.

This whole thread has been archived into my notes. Fantastic deep thoughts :)

You don't need to use the "complicated" features. Also, C has preprocessors and whole lot of other stuff that seems unnecesarily complicated to me.

@@dawkot6955 The preprocessor is probably the least complicated part of the whole stack.

Exactly what i thought about, too

Yeah through much of the talk I was thinking about what Joel used to call "Architecture Astronauts". It never caught on, alas. Joel's blog is what, 20 years old now? Dig that. We've been saying "OMG that is so true it resonates so much" for a full 20 years now. But have we moved a finger to do something about it? Have we bollocks. Much of what Jonathan is talking about has only accelerated, or indeed only came about, in those 20 years. Heck, Joel himself went on to create StackOverflow. That jab from Jonathan at 52:20 was well-earned indeed. Joel went from criticizing Architecture Astronauts to making sure that absolutely everyone can be an Architecture Astronaut themselves. You don't need to know anything about programming. You just go to StackOverflow and rely on what few smart people are still out there. They will do your job for you, for free. And then you won't even read what they say. You'll just copypaste their solution and ask the next question. There is no learning taking place. The only lesson we learn these days is that we never need to learn anything ever again. Because there's Google and Wikipedia and StackOverflow. Until there isn't. Until the last Jon Skeet has died, or until China has blocked all of the Internet, or simply until we're out somewhere where there's no signal. Then suddenly it dawns on us. That we don't know anything at all. And even the things that we did used to know, we've long forgotten. Because it was so easy to just look them up or ask someone else. Jonathan's talk starts off like some kind of overtly grim and unrealistic, albeit exciting, thought experiment. But the longer it goes on, the more you can't help but think, damn. He's not talking about some post-apocalyptic future. He is actually talking about myself, right now.

Office isn't great, and yet, somehow, OpenOffice manages to be worse.

I guarentee OpenOffice is piggy-backs on millions and millions of lines of code inorder to run on your computer. And that code is outside of their domain. Even the simplest software today is hanging on the end of a very skinny branch waiting to snap.

OpenOffice is a kinda dead project now. Most of its developers moved to LibreOffice.

A Thompson bonner

IIRC most Lisps have garbage collection or at least some kind of memory management abstraction built in, which seems to be the distinction between the Fortran/C/++ level and the C#/Haskell/JS level. So he (and I) would most likely put it at the highest level of abstraction along with C# etc. (which, perhaps pointedly, he puts at the bottom of his slide).

I feel like lisp, or at least the lisp philosophy, is actually on the right track to solving a lot of the problems he mentions. It’s a language that at its simplest maps to machine language about as directly as C does, but which makes it very easy to create more abstract sub-languages existing inside of it. And in some sense (a very unusual sense) lisp is more low level than even assembly, since its notation is meant to mirror the structure of processes rather than the structure a computer uses when representing processes. For most purposes this seems to make it easier to figure out how to do things and make things simpler. Of course, when you start thinking about stuff outside of lisp’s model of what a process is, like for example trying to prevent cache misses (something Jon Blow seems to care a lot about), it isn’t as clear. Anyways, even if lisp offers a lot of the features of higher level languages, most lispers seem to have a pretty deep grasp of how computers work and don’t use lisp in quite the same way high level languages are normally used.

Also, in theory lisp machines could be possible, so if history was different we might all be comparing lisp to assembler.

Lisp is the most abstract, since it was designed without any thought of hardware. It was purely a notation for computational processes. Then people began to make interpreters and compilers that were very slow at first. Common Lisp is different as it has incorporated lots of lower level features to improve performance. I don't think Jon's idea of 'simple' means the lisp variant of simplicity, but is closer to the forth variant i.e. if I understand the problem I have and how the hardware works I can make a simpler solution, rather than a more abstract one.

@@LKRaider Weren't there at least two companies selling such a thing in the early 80's? More than theory, though I guess it depends on the definition of "lisp machine".

Beautiful.

My mother saw a V1 fly low overhead with a flaming tail, cut out, and go silent. Then her friend didn't show up at school the next day. Her house had gone in the night.

if von braun's v2 were really advanced then civilians would've been killed discriminately, just like the allied did to civilians in hamburg, tokyo, etc...

@@____uncompetative the house went silent too?

@@mito88 The flying bomb had a noisy rocket engine and when it used all of its fuel it was as silent as a glider. Obviously, its collision with the house made noise, but she didn't mention that part.

YES! How many people know how to fix it? How many even want to, regardless of payment?

What happened to cobol based systems?

@@EconaelGaming Many government systems run on COBOL. The most recent example being New Jersey's unemployment system, which started cracking under loads of people applying all at the same time due to COVID

lmao Unity has hired people who are more hardcore about performance than Jon Blow is. Look up Mike Acton.

@@ShredST yes, ecs is great, but 19.3 is the most unstable version so far.

@@iamarugin I managed to install a version of 19 that literally gave an error message in an empty project. Just opening the thing resulted in an error, and they released it any way.

ECS is just one thing. It's not the whole solution.

As with everything else in computer science, we believe we're having revolutionary ideas, when actually we're just repeating the past. Just to illustrate why simplification was always a concern in computer science: The "Dragon Book", a 1986 book on compiler design, depicts a knight slaying the dragon of complexity on its cover.

shalamaby - re: installers I think he means we need installers to put the different parts of the program into place just so it runs, which could make sense if they were there to figure out the hardware architecture for us and then select the correct binary, but not even for that they are useful, you have to select the correct installer by platform first, then the installer will call all kinds of OS apis to register the program into the system, the binary is usually a fat one containing all possible architectures bloating your installed disk usage with unnecessary data, and since you cant just move the app to another system (because the registration calls and such will not be made) it's not useful at all, just a required bureaucracy we came up with and without real value to the end user.

I have hi-end computer with 12-core CPU, 32 GB RAM and last gen GPU inside and I still see latencies and freezes of GUI everywhere. Fun fact - I installed Windows XP just for Fun for my old laptop and its GUI works much more smooth)) Yes, it solves tasks (e.g. compilation) slower, but its interface works much better (not ideal, but better). We're on the way to hell)

CPUs are more complicated because the main processor has a digital cop sitting in the management engine frequently checking to see if the main CPU has permission to do something, like copy a file, disc or media stream, etc.

Yeah

When it climbs to the top of Distrowatch.......

A debugger made by JB would probably be waaay more useful right now than a new OS.

@@viciouswaffle A gaming OS would actually be really useful. Latency in modern operating systems is off the charts.

The problem is that modern desktop hardware has become too complex for one man to do that. We have to take a very very big step backwards in terms of hardware in order to take a better path.

Fuchsia maybe

We may not need a new kernel. We need a simple mental model, for the relationships between parts in the system. Interfaces needs to be well designed, in a coherent way. Then each part could hide its own complexity. That could be built from existing systems. BTW. I would not go with Fuchsia's Zirkon microkernel. Subpar capability model: you need revocation!. I'd suggest seL4 instead, which has it. You could also throw away C and enforce a type-safe language runtime, like Oberon, but I'd keep some memory protection just in case.

TempleOS, the only OS with Divine Intellect.

Thanks!

Thank you! And here I was feeling like the only one thinking he was being ridiculous. On the Boeing problem - the software was put in place as a hack to deal with the "hardware" problem of putting in larger engines and changing the plane dynamics. That's not a software issue - that's a rush to market issue! We're also building larger systems with more interconnecting parts. As you mentioned, it's a systems engineering problem.

Strange - but it is a software issue in that if you remove it you are forced to fix the actual problem. The software issue is also a social one: people rely so much on it, they lose track of what it can actually do, at a certain level of abstraction. When building higher up, you mostly assume the lower parts are a solid base, because otherwise, how can you build right? So the software people build on these false assumptions, and the resulting software is faulty. Let me tell you, false assumptions are 90% of software problems - the other 10% are the fixable errors. You mostly cannot fix false assumptions, in software or in people.

@@LKRaider completely agree - software is certainly _part_ of the problem, but to frame it as the whole problem or the biggest problem is, I believe, not giving proper consideration to the bigger... Ecosystem? Some word like that :P

You people are genius. I only know how to make a sandwich.

@@JK-zl9cq Do you? Do you know how to make a sandwich? I don't. I don't know how to make the bread. I don't know how to make the flour to make the bread. I don't know how to grow the grains to make the flour. I don't know how to make the butter to smear on the sandwich. I don't know how to get the milk from the cow to make the butter. I don't know how to farm the cows to get the milk from to make the butter. I don't know how to make the cheese to put on the sandwich, nor the peanut butter jelly, not the jam, nor the coleslaw salad and pickles... Nor do I know how to make any of the things that go into the pickles, the mayonnaise, etc. But fortunately these are all well-understood problems practised by millions of people day-in day-out and so as a society we are all still able to sustain production of sandwiches for a few hundred years already, despite major wars and disasters. (edit: fix typos)

With only one exception. Too many levels of indirections.