In other news today...
-
-
You thought you bought software – all you bought was a lie
Isn't code something worth paying for? Well, frankly, no
If you're a bit of a non-conformist, perhaps you chose a Mac
I stopped reading here.
-
Senator Wyden tells The Reg this latest infosec lapse is 'inexcusable'
Do I smell the whiff of a Okta like catastrophe in the wind?
-
@DogsB said in In other news today...:
Isn't code something worth paying for? Well, frankly, no
He keeps rambling and rambling and rambling and I when after skimming through I couldn't see where and if he gets to a point, I gave up.
@DogsB said in In other news today...:
Senator Wyden tells The Reg this latest infosec lapse is 'inexcusable'
Do I smell the whiff of a Okta like catastrophe in the wind?
Microsoft has been a catastrophe all the while, but it's big enough that users can't easily just flee.
Microsoft comes under blistering criticism for “grossly irresponsible” security
Azure looks like a house of cards collapsing under the weight of exploits and vulnerabilities.
linked around here last year is probably the most damning so far.
-
Jeff Bezos boosts fortune by selling $4bn in Amazon shares
Sale follows his recent move to Florida - where he will reportedly save hundreds of millions in tax
Amazon is about to hit the skids.
-
@DogsB said in In other news today...:
Isn't code something worth paying for? Well, frankly, no
If you're a bit of a non-conformist, perhaps you chose a Mac
I stopped reading here.
His point, such as it is, is that the lies the software industry tells itself are entirely just that, lies.
Basically it’s the usual drumbeating of why paid software is bad (because you don’t own it; because vendors have an intrinsic motivation in lock-in), and that you should use open source as much as you realistically can even if it isn’t the same comfortable UI you’re used to… pull the band-aid off and enjoy a destiny you at least have some control over. At least in theory.
His example is Outlook - you can’t have a free experience that fully and perfectly replicates Outlook but you can have one that is good enough without being entirely beholden to a specific vendor.
-
@Arantor said in In other news today...:
His example is Outlook - you can’t have a free experience that fully and perfectly replicates Outlook
No but you can have a free experience that is Outlook. It's free on Windows and macOS.
-
@Arantor said in In other news today...:
you can’t have a free experience that fully and perfectly replicates Outlook
Well duh. Neither a Mr Penetrator brand jumbo dragon dildo nor a 72 gallon drum of battery acid are available for nothing.
-
@Arantor said in In other news today...:
The speedier computing cake is a lie... so we got software bloat instead
Server chips have lots more cores, of course, because servers can use that better, but that has a ceiling too. Look it up: it's called Amdahl's law, after the late great Gene Amdahl, and it's quite scary. Even if a program can be made 95 per cent parallel, the maximum speedup you can get is about 20 times, no matter how many processor cores you throw at it.
Math is hard. Let's go coding!And, because computers aren't getting much quicker, as it gets bigger, software is getting slower. That's why we're not seeing big exciting new features, new capabilities and tools.
What world does this guy live in? The reason new features are less exciting is more that we're jaded and aren't impressed by that sort of thing any more. We expect it. Not that every new feature is amazing or whatever.
-
@Arantor said in In other news today...:
He also indirectly points at the UNIX philosophy: make more smaller tools that do a few things well rather than making large monstrosities that do everything.
Not really, because he whines about Debian (as
) pointed out. This was basically an unfunny Andy Rooney segment by someone who is out of his league (like how he generalized from the Wiki example of Amdahl's law).
-
@boomzilla The amount of speedup from going parallel depends on the algorithm, the computer architecture, the communications fabric, and the communications pattern. That's because the main thing that slows down parallel algorithms is the need for bits to communicate with each other and to synchronize their idea of the world, usually over a shared comms fabric. Other types of comms architectures can do much better and scale up a lot more... but aren't suitable for all applications.
The important thing to get very high speeds is to have very low consistency of model; while a CPU can know what its local state is with high fidelity, knowing what others know with that same fidelity is very costly.
-
@dkf said in In other news today...:
The amount of speedup from going parallel depends on the algorithm
Yes, which the wiki page explained pretty clearly, I thought, as opposed to the writer (or the original talker?) who glommed onto that "20 times" factor from the specific example.
Your low level explanation is completely irrelevant to his brain fart.
-
@boomzilla said in In other news today...:
@Arantor said in In other news today...:
The speedier computing cake is a lie... so we got software bloat instead
Server chips have lots more cores, of course, because servers can use that better, but that has a ceiling too. Look it up: it's called Amdahl's law, after the late great Gene Amdahl, and it's quite scary. Even if a program can be made 95 per cent parallel, the maximum speedup you can get is about 20 times, no matter how many processor cores you throw at it.
Math is hard. Let's go coding!And, because computers aren't getting much quicker, as it gets bigger, software is getting slower. That's why we're not seeing big exciting new features, new capabilities and tools.
What world does this guy live in? The reason new features are less exciting is more that we're jaded and aren't impressed by that sort of thing any more. We expect it. Not that every new feature is amazing or whatever.
His argument isn’t really that. Partly he’s getting at the “features for features sake” thinking, partly at “fuck me we don’t write software efficiently any more” and partly at “what if we slowed down how much software we wrote to do it nicely”
But all of that explained poorly with weird metaphors and analogies that don’t stack up to us because we’re not his target demo. We already know how shit it all is.
-
@Arantor said in In other news today...:
@boomzilla said in In other news today...:
@Arantor said in In other news today...:
The speedier computing cake is a lie... so we got software bloat instead
Server chips have lots more cores, of course, because servers can use that better, but that has a ceiling too. Look it up: it's called Amdahl's law, after the late great Gene Amdahl, and it's quite scary. Even if a program can be made 95 per cent parallel, the maximum speedup you can get is about 20 times, no matter how many processor cores you throw at it.
Math is hard. Let's go coding!And, because computers aren't getting much quicker, as it gets bigger, software is getting slower. That's why we're not seeing big exciting new features, new capabilities and tools.
What world does this guy live in? The reason new features are less exciting is more that we're jaded and aren't impressed by that sort of thing any more. We expect it. Not that every new feature is amazing or whatever.
His argument isn’t really that. Partly he’s getting at the “features for features sake” thinking, partly at “fuck me we don’t write software efficiently any more” and partly at “what if we slowed down how much software we wrote to do it nicely”
Yeah, and I'm saying he's wrong. There's all sorts of improvements in software that we gloss over and ignore. A lot of it isn't mind bending change, but it makes things a bit easier in lots of ways. And yeah, this takes more code. And yeah, also people do dumb things that make stuff worse. I'm not saying they don't. I'm just saying this guy is yelling at clouds and is mostly full of shit.
But all of that explained poorly with weird metaphors and analogies that don’t stack up to us because we’re not his target demo. We already know how shit it all is.
Maybe, but I'm not giving him the benefit of the doubt, especially after the retardery with Amdahl's law.
-
@boomzilla see, I’m not so sure.
From a user perspective I honestly don’t see it. Yes, more features. Cool. Except that a significant percentage of the time those features aren’t relevant and have all sorts of downsides.
Word is my favourite example here - I’ve been using Microsoft Word since 1998, I can’t tell you any appreciably new features it got since then (except, possibly, versioning and that I’m not sure about). But despite the frankly enormous computing power increase since then, it still feels sluggish. Unless I’m writing under a thousand words, Word feels fucking slow. It always has, and despite the orders of magnitude power increase since then, I don’t see it getting better.
Even as a dev I don’t really agree. We have more options now, yes, and I will grant that you can write more software quickly but I’d never suggest any of it was good.
I’m in PHP land, I use Laravel to build apps. It’s convenient, it has a good developer experience. But you know what it also has? 52MB of PHP dependencies that have to be shipped to the server before I even start writing a line of anything specific to my app.
And then we have things like Electron. I get why it exists - taking a rich web app and smooshing it into a local app is desirable, but could we not have invented a better way to do it?
And earlier today we had a bump in a topic on Windows desktop dev. The linked article - https://visualstudiomagazine.com/articles/2024/02/13/desktop-dev.aspx - had a wonderful table of the different frameworks available for Windows. And all of them from Microsoft as far as I can tell… why do we need so many options?
Remember we now live in a time where the humble calculator app is deemed to need a splash screen during startup. Splash screens historically were for apps that took too long to just appear and to reassure the user something was happening until the app could finish loading… what the fuck could a basic calculator app possibly be doing that would make this a sensible dev investment of time to implement?
-
@Arantor said in In other news today...:
And then we have things like Electron. I get why it exists - taking a rich web app and smooshing it into a local app is desirable, but could we not have invented a better way to do it?
We have. Tauri. Unlike Electron, it uses the HTML rendering libraries already present on every system instead of lugging them around. And I remember there was at least one other such tool.
@Arantor said in In other news today...:
And earlier today we had a bump in a topic on Windows desktop dev. The linked article - https://visualstudiomagazine.com/articles/2024/02/13/desktop-dev.aspx - had a wonderful table of the different frameworks available for Windows. And all of them from Microsoft as far as I can tell… why do we need so many options?
All of them from Microsoft except React Native … well, Microsoft still maintains its Windows (and, strangely enough, also macOS) port.
There are so many mainly for histerical raisins:
- MFC is the original one. It is obsolete and nobody sane would use it for new development, but some ancient software is tough.
- Then with .NET they created first Windows Forms, but it didn't lend itself well to separate layout definition, so they also created WPF.
- The WinUI 2, WinUI 3 and MAUI are mostly evolution of WPF except not backward compatible. Because the Windows team in Microsoft cares about backward compati(de)bility, but the .NET team never did.
- Blazor is actually a web framework that got a almost-but-not-full HTML rendering backend similar to React.
-
@Bulb sure, I get the historical angle for how they came to be. Why are they all still receiving updates? (Albeit UWP looks to be headed towards EOL, but it’s still alive for now.)
Do we need this many current supported options? And remember this is just from MS (+React which I didn’t immediately spot in the picture when I came back to it for this thread, mea culpa)… it’s not like other vendors don’t also ship their own ideas on top of this too.
-
@Arantor said in In other news today...:
@boomzilla see, I’m not so sure.
From a user perspective I honestly don’t see it. Yes, more features. Cool. Except that a significant percentage of the time those features aren’t relevant and have all sorts of downsides.
Word is my favourite example here - I’ve been using Microsoft Word since 1998, I can’t tell you any appreciably new features it got since then (except, possibly, versioning and that I’m not sure about). But despite the frankly enormous computing power increase since then, it still feels sluggish. Unless I’m writing under a thousand words, Word feels fucking slow. It always has, and despite the orders of magnitude power increase since then, I don’t see it getting better.
Yeah, Word has always sucked. But also remember that no one uses more than a handful of its functionality. So, sure, maybe those features aren't useful to you, but they're probably useful to someone (except for the real stinkers, of course).
Even as a dev I don’t really agree. We have more options now, yes, and I will grant that you can write more software quickly but I’d never suggest any of it was good.
I’m in PHP land, I use Laravel to build apps. It’s convenient, it has a good developer experience. But you know what it also has? 52MB of PHP dependencies that have to be shipped to the server before I even start writing a line of anything specific to my app.
Here's something I was thinking about when I was writing that. As part of our dependencies, we've kind of standardized on a set of table widgets that include sorting and searching. It's way better than the old stuff we had before, but yeah, it requires some dependencies. But it makes the user's life a lot better than it was. Sort of a trivial example, but not insignificant.
And then we have things like Electron. I get why it exists - taking a rich web app and smooshing it into a local app is desirable, but could we not have invented a better way to do it?
Not so far. Lots of people have chased after the killer cross platform development platform but none have been as successful as electron. I think some of what you're expressing is the @blakeyrat...something. Where he rants about how stuff should be simpler, but the fact is that it's actually pretty complicated and no, it's really not that easy.
And earlier today we had a bump in a topic on Windows desktop dev. The linked article - https://visualstudiomagazine.com/articles/2024/02/13/desktop-dev.aspx - had a wonderful table of the different frameworks available for Windows. And all of them from Microsoft as far as I can tell… why do we need so many options?
Remember we now live in a time where the humble calculator app is deemed to need a splash screen during startup. Splash screens historically were for apps that took too long to just appear and to reassure the user something was happening until the app could finish loading… what the fuck could a basic calculator app possibly be doing that would make this a sensible dev investment of time to implement?
Well, some people just make shit. My calculator app doesn't have a splash screen, so I couldn't possibly comment on this alleged example.
-
@Arantor said in In other news today...:
@Bulb sure, I get the historical angle for how they came to be. Why are they all still receiving updates? (Albeit UWP looks to be headed towards EOL, but it’s still alive for now.)
Do we need this many current supported options? And remember this is just from MS (+React which I didn’t immediately spot in the picture when I came back to it for this thread, mea culpa)… it’s not like other vendors don’t also ship their own ideas on top of this too.
Because they are all incompatible and there is too much old software written in them, so they need to be maintained. I don't think it means they would be receiving any new features. But there still needs to be someone who makes sure they keep working on new versions of Windows and fixes and occasionally fixes some bug.
-
@boomzilla yes, some people just make shit. The problem is that “some” is “most” and despite the fantastic increases in the power available, nothing is getting better with it. That toy company in Redmond needed to ship a calculator with a splash screen at least for a while (haven’t used it in a while)
That’s the takeaway. People wouldn’t even mind so much about all the extra features, I suspect if it didn’t feel like we were going backwards half the time.
The web is an even better example - people look at me funny when I say that I generally don’t write JS but I don’t, I stick to classical “press the button to go to the next thing” and add JS as nice to haves. I then don’t need huge hunks of Vue everywhere and it all just works, and snappily because it’s not heaving around fast blobs of anything.
For all our progress, has it really gotten better? You might think so, but I’m yet to be convinced, and apparently so is the article author.
-
@Bulb said in In other news today...:
@Arantor said in In other news today...:
@Bulb sure, I get the historical angle for how they came to be. Why are they all still receiving updates? (Albeit UWP looks to be headed towards EOL, but it’s still alive for now.)
Do we need this many current supported options? And remember this is just from MS (+React which I didn’t immediately spot in the picture when I came back to it for this thread, mea culpa)… it’s not like other vendors don’t also ship their own ideas on top of this too.
Because they are all incompatible and there is too much old software written in them, so they need to be maintained. I don't think it means they would be receiving any new features. But there still needs to be someone who makes sure they keep working on new versions of Windows and fixes and occasionally fixes some bug.
6 of those 8 columns had a roadmap link. That suggested to me they were looking to add things.
-
@Arantor said in In other news today...:
The problem is that “some” is “most” and despite the fantastic increases in the power available, nothing is getting better with it.
This is a you problem. We all love to rant around here but I don't buy that for a second.
That’s the takeaway. People wouldn’t even mind so much about all the extra features, I suspect if it didn’t feel like we were going backwards half the time.
People are ungrateful and unobservant assholes (certainly including me).
For all our progress, has it really gotten better? You might think so, but I’m yet to be convinced, and apparently so is the article author.
I do think so. The author is writing about stuff he doesn't understand.
-
@Arantor said in In other news today...:
had a wonderful table of the different frameworks available for Windows. And all of them from Microsoft as far as I can tell… why do we need so many options?
And that's not even all the frameworks available. Qt, wxWidgets, and probably lots more (those are the 2 I use) (including MFC - it's still there)
-
@DogsB said in In other news today...:
Amazon is about to hit the skids.
Good. Amazon is a terrible company that treats its workers terribly. I'd like terrible things to happen to it.
-
@Arantor said in In other news today...:
you can’t have a free experience that fully and perfectly replicates Outlook
You say this like its something you might want to replicate.
-
@HardwareGeek said in In other news today...:
@Arantor said in In other news today...:
you can’t have a free experience that fully and perfectly replicates Outlook
You say this like its something you might want to replicate.
Me? God no. But I remember being in a company a few years ago that was all about open source - if you wanted anything that wasn't open source, you needed written permission.
And I remember a new starter who wasn't a dev, but was very firmly into the office life and was so disappointed that Thunderbird + calendar was such a poor imitation of Outlook...
-
@Arantor said in In other news today...:
I’ve been using Microsoft Word since 1998, I can’t tell you any appreciably new features it got since then
But muh Ribbon!!!!1!1!
I’m in PHP land
Sounds like a you problem.
-
@Arantor said in In other news today...:
some people just make shit. The problem is that “some” is “most”
Ever hear of a guy named Theodore Sturgeon?
-
@boomzilla said in In other news today...:
Amdahl's law, ... Even if a program can be made 95 per cent parallel, the maximum speedup you can get is about 20 times, no matter how many processor cores you throw at it.
Math is hard. Let's go coding!Needs more
-
@boomzilla said in In other news today...:
Well, some people just make shit. My calculator app doesn't have a splash screen, so I couldn't possibly comment on this alleged example.
That's because kcalc is a classic native Qt app, which weighs in at
7 kB500 kB, not 100 MB. (Looking at it from an extreme
perspective, it probably should be no more than a dozen kB, but I was still surprised to see it did. Most of the meat is in a library though.)
Who cares about a few hundred megs here and there? Well, you can get away with that shit for a few stupidly large things, but if all your little tools all packed their inner-platform-OS along separately, it would add up fast.
-
@topspin said in In other news today...:
That's because kcalc is a classic native Qt app, which weighs in at
7 kB500 kB, not 100 MB.In memory it weighs 88MB though.
ldd /usr/bin/kcalc | sed 's/.*=> \(.*\) (.*/\1/p;d' | xargs realpath | xargs du -ch /usr/bin/kcalcsays so. The shared libraries do only get loaded once for all applications that use them though.Also, relatively simple go applications, like command-line tools for the various cloud stuff, often weigh around 100 MB with no GUI at all simply because muh static linking.
-
@Bulb For each command-line calculator I tried though, by far the largest contributor was libc at 2.3MB.
-
Relatedly:
DSL 2024 is not as svelte as it used to be – but who is?
Warning: same author.
-
@Bulb said in In other news today...:
In memory it weighs 88MB though
Actually loading it doesn't use nearly as much. And you don't run all OS utilities at the same time. But if they were all electron, then each one would use a gigantic amount of disk space.
-
@topspin said in In other news today...:
@Bulb said in In other news today...:
In memory it weighs 88MB though
Actually loading it doesn't use nearly as much.
It does require that amount of memory. Much of it may be shared with other KDE applications, so if you are running full KDE desktop, it might only add a megabyte when you start it, but that 88MB gets loaded. And then of course some dynamic allocations are made for data.
And you don't run all OS utilities at the same time.
I didn't count any OS utilities. I only counted all shared libraries that
kcalc, as currently installed on my Debian Testing system, references (lddlists shared libraries used by a process, the sed extracts the resolved names, anddu -chto calculate total size over all arguments).But if they were all electron, then each one would use a gigantic amount of disk space.
And memory, because data can't be shared in memory if they come from different files even if the files have the same content, yes. However if they were Tauri instead, they would again share all the browser runtime.
-
Stingray’s mysterious pregnancy could be a shark’s tale
Bite marks found on the female’s body may indicate a paternal link to the unborn pups and the shark next door
Nope!
-
@Bulb said in In other news today...:
It does require that amount of memory. Much of it may be shared with other KDE applications, so if you are running full KDE desktop, it might only add a megabyte when you start it, but that 88MB gets loaded.
That 88MB gets mapped into memory, but might not be loaded from disk.
-
@dkf … which is true for electron and go and such apps too.
-
@dkf said in In other news today...:
@Bulb said in In other news today...:
It does require that amount of memory. Much of it may be shared with other KDE applications, so if you are running full KDE desktop, it might only add a megabyte when you start it, but that 88MB gets loaded.
That 88MB gets mapped into memory, but might not be loaded from disk.
Resident set size then:
$ ps -eO rss PID RSS S TTY TIME COMMAND 24223 69928 S ? 00:00:00 gnome-calculator 24267 70284 S ? 00:00:00 qalculate-gtk 24373 1036 S pts/1 00:00:00 dc 24409 2396 S pts/2 00:00:00 bcAll of them had been started only; none had been asked to perform any computation.
-
@PleegWat Does RSS include memory-mapped drawing surfaces?
-
@dkf It's all RAM associated with the process, including shared RAM, but not pages which are swapped out.
-
@PleegWat So some of that will be the drawing surfaces, as that's how Wayland works (and you're probably using that with Qt now). That might cover a few MB, but most of the space is still
To my eyes, it's at least 60MB too much...
-
@dkf I'm still on X. Zoom didn't like wayland when I first configured this system.
-
Lets go with some Math today:
Smoothed asymptotics: from number theory to QFT
Inspired by the method of smoothed asymptotics developed by Terence Tao, we introduce a new ultra-violet regularisation scheme for loop integrals in quantum field theory which we call $η$...
-
@Dragoon said in In other news today...:
Lets go with some Math today:
High Energy Physics - Theory
Well ok, fair enough. Most of theoretical physics is just math masturbation by now.
inspires a regularisation scheme for non-abelian gauge theories coupled to Dirac fermions that preserves the Ward identity for the vacuum polarisation tensor.
What did Clarke say again? Any sufficiently advanced abstract is indistinguishable from ChatGPT.
-
@PleegWat said in In other news today...:
@dkf I'm still on X. Zoom didn't like wayland when I first configured this system.
Then WTF is it doing? It doesn't take that much memory to do GUIs on X...
-
@dkf said in In other news today...:
@PleegWat said in In other news today...:
@dkf I'm still on X. Zoom didn't like wayland when I first configured this system.
Then WTF is it doing? It doesn't take that much memory to do GUIs on X...
Dunno. I took a quick look at
/proc/../mapsbut that only lists start and end address, not segment size, and I couldn't figure out a shell script to do the subtraction (in hex) offhand. I did notice that apart from libraries gnome-calculator also had quite a bunch of font files loaded.
-
This couldn't possibly go wrong.
-
Finally, a useful suggestion from one of this red state's politicians:
Why would Pluto become Arizona's official state planet?
Percival Lowell's theories of a 9th planet in our solar system came to fruition with the discovery of Pluto at the Lowell Observatory in Flagstaff.
-
@Arantor I do love how he clearly states what interference was done. Love being told to "google it!" as part of the references cited...
