It's not just during peak. I think Amazon scales down during non-peak times and performance is almost just as bad.
I been a Prime customer since the year after it launched. I probably returned an item a year; if that. Today, our ordering has only doubled but we return atleast one package a month; primarily due to delivery issues.
I think we are on some "bad customer" list because every once in a while the rep says "You return a lot..."; implying that we shouldn't do that. We always respond with, look the notes and figure out who is at fault.
If we order blue, we shouldn't get red. If we order something gift wrapped, we probably want it before TK... not two weeks after the dinner. No, we aren't going to go to the other neighborhood or even down the street and pick up the bad delivery. No, we are not going to accept a box with a forklift hole in it. No we don't want another brand automatically substituted and billed to us. You credited us because your tracking number became active. Don't recharge me because you didn't get the return in time!
And all this we find out if we inquire or after delivery. I think Amazon just doesn't care anymore. They are more focused on AWS, warehouse automation, and another HQ.
That's almost the exact opposite experience I have with Amazon. Sure, every once and a while things are late. But often times I order something and it comes before they originally indicated. I can get same day delivery on most common items. The times that I have had a problem Amazon refunds me while allowing me to keep the package and has even offered me a gift card in addition to the free item.
Ohhh the "what about lag" argument ruins online play for so many games. Where you can basically cripple your own network connection by streaming a bunch of netflix videos or youtube videos on another machine so that you teleport around on everyone else's screen and they pause on your screen while you headshot them. I am sorry but if you're more than about 100ms behind everyone else in this day and age then there should be no compensation for your lag. In PUBG you can watch kill cams where people aren't even aiming remotely close to where you are, where they are shooting behind you and kill you because of lag compensation. It's the best when someone kills you while you're driving perpendicular to them at 120km/h and you watch the kill cam and they're shooting an easy 50m behind you.
He didn't pay random people, he paid friends. I don't know about your friends, but I generally trust mine to not pull a fast one on me over some change.
He said he paid friends of friends. And that he "put the feelers out" on who would be willing to do this for money. So... no. He got exactly what he should have, or most likely, did expect.
There are a lot of rage filled comments here about him being a liar, when it seems to me his main crime was gullibility?
I've occasionally watched his videos for years, and he seems to be a fairly decent, very smart guy. He did formerly work for NASA as an actual engineer - his interactions with current NASA personnel are proof of that.
His story is that obviously he couldn't get that many thief interactions on his own doorstep, so he loaned the build out to friends, and friends of friends with a cash reward for interactions with recoveries...I think that was his mistake. Once you bring money or fame into it, people are assholes; with both involved it's almost guaranteed. So some of them staged the reactions, ostensibly for the cash, but imo more for the internet fame...and now that's screwed him over.
His "crime" if anything isn't lying, it's being gullible and trusting.
Is anyone so naive to expect that if they pay someone to have a package stolen AND returned that the person is going to actually let any random person steal the package? Come on. I'd expect 9 out of 10 elementary school kids to know better than that. And what internet fame did the people who put the packages out receive? I don't recall seeing any of their names or faces in the video, though I admittedly watched it days ago and do not remember. I think the people being gullible are the ones that are giving this guy the benefit of the doubt - saying that he couldn't have possibly known that some of this would be fake. You are either saying that he's incredibly stupid or that he intentionally pretended to be incredibly stupid.
What we seem to have lost in this age is the ability of differentiated, graded responses. Everyone is either "omg, my hero!" or "that fraud, he's so evil!". There is no inbetween. Not in politics, not in economics and not in YouTubers.
Mod parent up?? He paid people to put this on their porch, but only if he got the package back. What did he think was going to happen? The ex-NASA guy is either a real idiot when it comes to people, or he knew that some or all of the reactions would be fake. Only the most naive person in the world would not expect this result.
ARM's always been an exception unfortunately, and the whole embedded space is a mess generally.
True. It's a lot easier to use non UEFI firmware with ARM, though. In fact, it is rarely used on ARM. AFAIK the only phone manufacturer to use UEFI is Apple, but I could be wrong. Though in the ARM server world there is a lot more UEFI firmware. If you have a very limited set of hardware requirements, CoreBoot is probably a better choice. But if you need to support expansion slots and a very configurable boot process then UEFI is way more mature. I know the Siemens uses CoreBoot to load up Intel based hardware on their industrial computers and UEFI firmware for their laptops and other home user products.
I take it as meaning they want to use the 'firmware as a service' aspects of UEFI (all UEFIs), but their current UEFI (TianoCore) is a bloated mess that is too difficult to maintain.
Microsoft basically wants the ability for the firmware image on the flash part to have multiple signatures. The ME or PSP portion of the flash part would be signed by the silicon vendor. Same with microcode, etc etc. UEFI allows modules. So if a generic module is being deployed then it can be signed by the developer of that module. This would allow security fixes to be pushed down to the flash without worrying about whether the OEM or ODM has decided to roll a new firmware image. This is actually the real issue in firmware security right now. A fix gets made available and many manufacturers choose not to pick it up for cost reasons. And it has nothing to do with TianoCore being too difficult to maintain and everything to do with the fact that changing the firmware has risks involved. This is a far more difficult problem than MIcrosoft cares to admit. Since they have a limited set of hardware they deal with it is not a problem for them. But if you look at a company such as Quanta they may be supporting dozens of boards for different customers.
UEFI: the only thing imaginable that could be worse than BIOS. Now, every computer has a shitty DOS-class OS burned into the firmware, which is permanently resident, and the perfect platform for back doors and spyware. That is on top of the IME, which reduces trust even further. Intel x86 considered harmful paints a vivid and damning picture of the modern x86 platform.
Any bootstrap mechanism should be simple and transparent, and leave the owner with full control of the machine. CoreBoot is a good starting point. Replacing x86 is a good next step, since Intel refuses to document the platform, and requires binary blobs to boot the platform.
There are worrying efforts which may also infect RISC-V platforms with UEFI and "secure" enclaves. It is ironic that open hardware efforts could help accelerate locked-down computing, if vendors widely adopt these user-hostile technologies.
CoreBoot has its place but it does not natively provide the security of the UEFI specification and also does not handle all of the potential use cases that the UEFI specification does as well. And how many exploits have you seen recently that revolve around issues in the firmware and not issues in the secure processor? I can tell you right now that most computers that have shipped in the last year or two, while potentially containing exploitable flaws, have generally not had any issue that allows an attacker to compromise the flash part unless they exploited a security issue in a BMC or through a secure processor such as ME or PSP.
If you manage desktops for a business, UEFI is both a pain in the but and a God-send.
A pain in the but because all these new-fangled UEFI machines broke your old imaging process. You've had to adapt to either learn to switch machines over to legacy boot more or (worst case) completely re-design your imaging process from the ground up with UEFI in mind.
A god-send because it enables some really cool new deployment scenarios, mainly through network booting. So you can set a machine to boot to a network location that will automatically deploy the OS. You can even configure the machine to reboot every night and re-image back to pristine, without any 3rd-party software.
A god-send because it enables some new security features: namely, validating the boot image, which makes it very difficult for root kits to take hold.
I remember the Slashdot reaction when MS first announced support for SecureBoot. So many comments about how MS would use it to lock out linux or other open source products. That was nearly 10 years ago, and nothing even close to that has happened, but we have seen a meaningful security benefit.
You should not be using legacy boot in an enterprise situation. The only exception to that would be if you have some old operating system that does not work except in legacy mode. If you’re booting in legacy mode then your system’s security is about as solid as a slice of swiss cheese.
I have built Tianocore images and booted them in order to understand UEFI so I think it's fair to say I'm in that other 1% and I assure you UEFI is a clusterfuck.
That depends entirely on what you are trying to accomplish. TianoCore provides drivers for just about everything. Way more than is needed to boot most systems. THat’s because UEFI is designed to work for everybody in just about every use case imaginable. You could use something like CoreBoot but the feature set is much smaller. Sometimes that is a good thing and sometimes that is a bad thing. It depends entirely on what you’re trying to accomplish.
Not true, quite a few computers, especially server and business grade allow a physically present user enter platform setup mode and upload the public portion of the PEK and then you can sign any KEK that you like.
Any OEM machine that ships with Windows is required to allow a physically present user to install their own key. The only OEM that can get around this? Microsoft. They do not follow that rule with the Surface, as an example. At least not the ARM ones. I don’t know what they do with the x86 ones.
Because they all need to petition redmond for a signature with the redmond key to be somewhat viable as a competitor to redmond. And, of course, it means supporting lots of redmond tech (fatshame32) just so you can boot. BIOS is crappy but so much simpler to deal with, that UEFI is not an improvement.
This is FUD. The UEFI Forum solicited proposals for companies to provide the root of trust for Secure Boot and all of the big security companies wanted a lot of money to host this. Microsoft offered to host it for free. Any other company could make the same offer the UEFI Forum would gladly accept it and add another root of trust. Not to mention that a requirement to get a system certified for OEM sale of Windows requires that the end user be able to install their own Secure Boot key. This means that anyone can sign their own boot loaders. You only need to have a signing key from the Microsoft authority if you want your boot loader, driver, etc, to run with secure boot enabled without having to load a key into the keystore on the machine.
As someone who lives in the UK afaict UK prime is better than us prime. AIUI US prime only gives free 2-day delivery while UK prime gives free next day or if you prefer nominated day on most Amazon-stocked items (some items seem to have more restrictive options, probablly as a result of being in different warehouses). Also unlike most next-day and nominated day services Amazon offers weekend deliveries at no extra cost.
Prime is great when you are doing something like preparing for an event when you want to order a bunch of stuff over a short period (but not all at once because you inevitablly find you need more stuff as the preperations proceed). I haven't yet found motivation to stay subscribed all the time though, I just buy a month when I want it.
I know amazon UK has a video service, but I dunno if it's better or worse than the US one, I haven't tried it personally.
I live in the US and can get free same day delivery on many items. If its a common item that is ordered a lot in my metro area they will deliver it the same day for free, up until noon. If I order after noon then I will get it the next day. Rarer and more expensive things typically come next day - though again this depends on popularity.
I won't comment on your other claims but, I do know that Google moved location services (as well as many other services) from the OS to Google Play Services in order to make them updatable separate from the OS. We've all seen the complaints about Android fragmentation right? There's a reason that most apps will run on any version of Android newer than 4 or 5 now and that's because of the way Google Play Services works in more recent versions of Android.
You mean the library is upgradeable through play or that it is a part of the Google Play framework? I mean that all location data routes through Play in stock Android. As in, you cannot just turn off location services for all of the Google apps. To use location services at all you have to give your location to Google. This has nothing to do with being an upgradeable library and everything to do with Google removing the ability to restrict location services from Google's own applications. I'm not sure if that is the case with AOSP but if the phone vendor uses Play then you have no choice but to give location data to google or not use location services anywhere.
Not only do I think this is disingenuous but he's also failing to point out that Google arbitrarily requires you to give them the whole kit and caboodle of personal data if you want to use certain features. Want to use the Play store? You have to enable location services that Google tracks. Want to use Google maps and actually store maps so you don't download them every time (whether you are on android or not)? You have to sign in and give over all your tracking data. And are there actual legitimate requirements for this? No - Google intentionally moved location services inside of the Play framework, it used to be completely separate and standalone. Do you need to be signed in to download maps to the local device? Of course not. But Google knows people want to use these features on their devices and the force you to give up everything to use them.
Quake had true 3D levels that can pass over one another. Quake had 3D adversaries. Quake had network gameplay with Quakeworld. Quake had OpenGL support with GLQuake that launched the GPU world, really starting with the Voodoo 1. Quake had translucent water, which was amazing the first time I ever saw it. And lastly, Quake is still the bar that any small platform must aspire to by answering the question, "Does it run Quake."
Man I always though Quake was garbage and that was when I stopped playing FPS games.
Ahhh one of my older brothers was a security guard at HP when they closed down a bunch of facilities over in Sunnyvale and threw away a bunch of hardware. We raided that trash for coax network cards and used Novell and IPX to play Doom and eventually Duke Nukem. I think that Doom allowed a 4 person multiplayer? I forget. Anyway, not long after that and I was playing Warcraft 2 over dial-up modem with my friend. And oh how the world changed when one of my friend’s dad, who worked at Cisco, got an ISDN line and we were able to download boatloads of MP3s and trade them all over school using CDRs. Man. Now I have gigabit fiber into the home and I can download gigs if data in just a few minutes. My how the world has changed.
LMOL first define what you compensation is then tell people it's not about the money.
You have bills to pay. Employers bottom dollar pay. So money is everything. When you get paid beyond a certain amount, then and only then do you think, maybe I would like ping pong. But with student loans, rent / mortgage, family and what not, it's about getting paid.
When I finished school, I was definitely looking for professional growth, work life balance, and interesting problems to solve. In fact, AMD offered me a very nice salary to work for them when I finished. I turned it down, even though it was the highest paying offer I received. I ended up going to another company that was about the same size as AMD, but transitioning from analog to digital. At AMD I would have been one of thousands of software people. At this other company I was one in about fifty. I started out making less than 50% of what AMD had offered me. However, thanks to a bit of luck and hard work, I was leading projects by the end of my second year at the company. End result? I was making more than someone I knew who took the exact same offer from AMD I had received and I was gaining much better experience than they were, too.
I fully agree with you that the Slack client is horribly inefficient. But my point is that anything that does the same thing more efficiently in RAM and CPU across all required platforms is ultimately economically unmaintainable.
Electron does not have a monopoly on cross-platform development. In fact, it’s pretty late to the game. So this is only true if you define “economically maintainable” to mean that you can hire a bunch of high school kids in some third world country pennies to do the development. There are far better frameworks for this kind of work, but they require a more knowledgeable development team to use. Of course if you go the cheap route you’ll end up with an unmaintainable mess that is a pain to keep running. There‘s a risk of that happening even with an experienced development team, of course. But javascript is probably easier to shoot yourself in the foot with than C.
Hipchat supports all of those capabilities and you can serve hundreds of users from a single server instance that uses less RAM than the client side of Slack.
As of February 15, 2019, the client side of HipChat will use exactly the same amount of RAM as the client side of Slack because of Atlassian's deal with Slack to discontinue HipChat, including licenses for on-premises instances, and migrate all HipChat users to Slack. The server side of HipChat will no longer be available to the public at all, as Slack will host all instances on servers operated by Slack. On-premises users who cannot migrate to Slack will have to export their data and import it to some unspecified competing on-premises chat platform. (Source: "Slack Migration Frequently Asked Questions")
What's that got to do with anything? It means something better (from a resource consumption standpoint) existed and Slack is extinguishing it. But your claim was that no one had ever written anything like slack before and that is utter nonsense. Are you somehow financially tied to Slack? Because you obviously can't accept that the Slack client was horribly implemented.
[The features that proprietary web chat has over IRC] are all SERVER features.
First, I'd be interested to see what IRC servers support these features. Second, these features still need some sort of client integration, such as to request a particular history range or thumbnail image or to send an attachment, and I'd be interested to see what IRC clients have been customized to integrate with a server that supports these features. I agree that one could create a client to do these that's lighter weight than the existing Electron-based clients, but as of 2018, to the best of my knowledge, no one has.
Hipchat supports all of those capabilities and you can serve hundreds of users from a single server instance that uses less RAM than the client side of Slack. Do you need any more examples?
Funny. It's the native Mac apps that always seem to take up several GiBs of memory each. Run XCode plus a web browser and you're already past the available RAM on a Linux machine where I'm running half a dozen electron apps at a time.
Try again.
I've got 3 Xcode Projects open right now and it is consuming 161MB of RAM. Safari is consuming about 1.2GB. When I used to have to run slack for work, it would easily consume over 2GB of RAM. Opening Atom consumed 500MB of RAM without a single file open, though it did eventually slim down to 200MB. But I can see that Atom has a problem because RAM consumption is increasing by over 10MB every 10-20 seconds and it is sitting in the background doing nothing without anything open. It is now up to 350MB of RAM in just a few minutes.
There's little these apps do that vim and emacs and IRC haven't done for decades for a tiny percentage of these requirements.
Vim and Emacs I'll grant. But IRC out of the box doesn't do chat history, file attachments, link preview, voice chat, or channel groups (Slack "workspaces" or Discord "servers").
I know you keep going on and on about Slack's chat history, file attachments, etc but you're missing one thing. Those are all SERVER features. You do not need gigabytes of RAM on the client in order for the server to push you chat history out of a DB. It's not required for sending files, or any of those other things you're talking about. It's just plain sloppy programming and the fact that people consider Slack to be a commercial tool they're willing to pay money on is insanity. I'd be embarrassed if I helped deliver the slack client. It would never end up on my resume and I wouldn't want to tell any of my friends about it.
Oh, huh. That is pretty cool. I am not sure what those are, except for Skype, but sounds very good. I can feel my ironic beard growing.
What Tepples doesn't point out is that Slack, a simple IRC style chat interface, can easily consume multiple gigabytes of RAM and the more instances you have your slack client connected to, the higher the GB consumption soars. Discord also causes performance problems on my gaming PC when certain things happen. I'm not sure what event is causing it, but it temporarily hangs games while playing some sort of chime over my headset. And that's with a very high end gaming PC.
In fact, the throughput difference is so huge that switching from SATA to PCIe drops a compile time on one of my projects by 70%, So what did I end up doing? I ordered a Lenovo laptop that supports NVMe PCIe [...] I will never buy another Apple computer again.
Not even when asked to compile and test a macOS or iOS port of the application that you are compiling?
Actually, Apple licenses the source to one of the projects I work on and they compile it on... Windows! Though Windows is not my personal choice for any sort of work project. It is the only platform that has the build tools required, though the company that provides the build tools is starting to transition their tools over to Linux.
Slashdot Mirror
It's not just during peak. I think Amazon scales down during non-peak times and performance is almost just as bad.
I been a Prime customer since the year after it launched. I probably returned an item a year; if that. Today, our ordering has only doubled but we return atleast one package a month; primarily due to delivery issues.
I think we are on some "bad customer" list because every once in a while the rep says "You return a lot..."; implying that we shouldn't do that. We always respond with, look the notes and figure out who is at fault.
If we order blue, we shouldn't get red. If we order something gift wrapped, we probably want it before TK... not two weeks after the dinner. No, we aren't going to go to the other neighborhood or even down the street and pick up the bad delivery. No, we are not going to accept a box with a forklift hole in it. No we don't want another brand automatically substituted and billed to us. You credited us because your tracking number became active. Don't recharge me because you didn't get the return in time!
And all this we find out if we inquire or after delivery. I think Amazon just doesn't care anymore. They are more focused on AWS, warehouse automation, and another HQ.
That's almost the exact opposite experience I have with Amazon. Sure, every once and a while things are late. But often times I order something and it comes before they originally indicated. I can get same day delivery on most common items. The times that I have had a problem Amazon refunds me while allowing me to keep the package and has even offered me a gift card in addition to the free item.
That's just the viewport. What about the lag?
Ohhh the "what about lag" argument ruins online play for so many games. Where you can basically cripple your own network connection by streaming a bunch of netflix videos or youtube videos on another machine so that you teleport around on everyone else's screen and they pause on your screen while you headshot them. I am sorry but if you're more than about 100ms behind everyone else in this day and age then there should be no compensation for your lag. In PUBG you can watch kill cams where people aren't even aiming remotely close to where you are, where they are shooting behind you and kill you because of lag compensation. It's the best when someone kills you while you're driving perpendicular to them at 120km/h and you watch the kill cam and they're shooting an easy 50m behind you.
He didn't pay random people, he paid friends. I don't know about your friends, but I generally trust mine to not pull a fast one on me over some change.
He said he paid friends of friends. And that he "put the feelers out" on who would be willing to do this for money. So... no. He got exactly what he should have, or most likely, did expect.
There are a lot of rage filled comments here about him being a liar, when it seems to me his main crime was gullibility?
I've occasionally watched his videos for years, and he seems to be a fairly decent, very smart guy. He did formerly work for NASA as an actual engineer - his interactions with current NASA personnel are proof of that.
His story is that obviously he couldn't get that many thief interactions on his own doorstep, so he loaned the build out to friends, and friends of friends with a cash reward for interactions with recoveries...I think that was his mistake. Once you bring money or fame into it, people are assholes; with both involved it's almost guaranteed. So some of them staged the reactions, ostensibly for the cash, but imo more for the internet fame...and now that's screwed him over.
His "crime" if anything isn't lying, it's being gullible and trusting.
Is anyone so naive to expect that if they pay someone to have a package stolen AND returned that the person is going to actually let any random person steal the package? Come on. I'd expect 9 out of 10 elementary school kids to know better than that. And what internet fame did the people who put the packages out receive? I don't recall seeing any of their names or faces in the video, though I admittedly watched it days ago and do not remember. I think the people being gullible are the ones that are giving this guy the benefit of the doubt - saying that he couldn't have possibly known that some of this would be fake. You are either saying that he's incredibly stupid or that he intentionally pretended to be incredibly stupid.
Don't have mod points today, but mod parent up.
What we seem to have lost in this age is the ability of differentiated, graded responses. Everyone is either "omg, my hero!" or "that fraud, he's so evil!". There is no inbetween. Not in politics, not in economics and not in YouTubers.
Mod parent up?? He paid people to put this on their porch, but only if he got the package back. What did he think was going to happen? The ex-NASA guy is either a real idiot when it comes to people, or he knew that some or all of the reactions would be fake. Only the most naive person in the world would not expect this result.
ARM's always been an exception unfortunately, and the whole embedded space is a mess generally.
True. It's a lot easier to use non UEFI firmware with ARM, though. In fact, it is rarely used on ARM. AFAIK the only phone manufacturer to use UEFI is Apple, but I could be wrong. Though in the ARM server world there is a lot more UEFI firmware. If you have a very limited set of hardware requirements, CoreBoot is probably a better choice. But if you need to support expansion slots and a very configurable boot process then UEFI is way more mature. I know the Siemens uses CoreBoot to load up Intel based hardware on their industrial computers and UEFI firmware for their laptops and other home user products.
I take it as meaning they want to use the 'firmware as a service' aspects of UEFI (all UEFIs), but their current UEFI (TianoCore) is a bloated mess that is too difficult to maintain.
Microsoft basically wants the ability for the firmware image on the flash part to have multiple signatures. The ME or PSP portion of the flash part would be signed by the silicon vendor. Same with microcode, etc etc. UEFI allows modules. So if a generic module is being deployed then it can be signed by the developer of that module. This would allow security fixes to be pushed down to the flash without worrying about whether the OEM or ODM has decided to roll a new firmware image. This is actually the real issue in firmware security right now. A fix gets made available and many manufacturers choose not to pick it up for cost reasons. And it has nothing to do with TianoCore being too difficult to maintain and everything to do with the fact that changing the firmware has risks involved. This is a far more difficult problem than MIcrosoft cares to admit. Since they have a limited set of hardware they deal with it is not a problem for them. But if you look at a company such as Quanta they may be supporting dozens of boards for different customers.
UEFI: the only thing imaginable that could be worse than BIOS. Now, every computer has a shitty DOS-class OS burned into the firmware, which is permanently resident, and the perfect platform for back doors and spyware. That is on top of the IME, which reduces trust even further. Intel x86 considered harmful paints a vivid and damning picture of the modern x86 platform.
Any bootstrap mechanism should be simple and transparent, and leave the owner with full control of the machine. CoreBoot is a good starting point. Replacing x86 is a good next step, since Intel refuses to document the platform, and requires binary blobs to boot the platform.
There are worrying efforts which may also infect RISC-V platforms with UEFI and "secure" enclaves. It is ironic that open hardware efforts could help accelerate locked-down computing, if vendors widely adopt these user-hostile technologies.
CoreBoot has its place but it does not natively provide the security of the UEFI specification and also does not handle all of the potential use cases that the UEFI specification does as well. And how many exploits have you seen recently that revolve around issues in the firmware and not issues in the secure processor? I can tell you right now that most computers that have shipped in the last year or two, while potentially containing exploitable flaws, have generally not had any issue that allows an attacker to compromise the flash part unless they exploited a security issue in a BMC or through a secure processor such as ME or PSP.
If you manage desktops for a business, UEFI is both a pain in the but and a God-send.
A pain in the but because all these new-fangled UEFI machines broke your old imaging process. You've had to adapt to either learn to switch machines over to legacy boot more or (worst case) completely re-design your imaging process from the ground up with UEFI in mind.
A god-send because it enables some really cool new deployment scenarios, mainly through network booting. So you can set a machine to boot to a network location that will automatically deploy the OS. You can even configure the machine to reboot every night and re-image back to pristine, without any 3rd-party software. A god-send because it enables some new security features: namely, validating the boot image, which makes it very difficult for root kits to take hold.
I remember the Slashdot reaction when MS first announced support for SecureBoot. So many comments about how MS would use it to lock out linux or other open source products. That was nearly 10 years ago, and nothing even close to that has happened, but we have seen a meaningful security benefit.
You should not be using legacy boot in an enterprise situation. The only exception to that would be if you have some old operating system that does not work except in legacy mode. If you’re booting in legacy mode then your system’s security is about as solid as a slice of swiss cheese.
I have built Tianocore images and booted them in order to understand UEFI so I think it's fair to say I'm in that other 1% and I assure you UEFI is a clusterfuck.
That depends entirely on what you are trying to accomplish. TianoCore provides drivers for just about everything. Way more than is needed to boot most systems. THat’s because UEFI is designed to work for everybody in just about every use case imaginable. You could use something like CoreBoot but the feature set is much smaller. Sometimes that is a good thing and sometimes that is a bad thing. It depends entirely on what you’re trying to accomplish.
Not true, quite a few computers, especially server and business grade allow a physically present user enter platform setup mode and upload the public portion of the PEK and then you can sign any KEK that you like.
Any OEM machine that ships with Windows is required to allow a physically present user to install their own key. The only OEM that can get around this? Microsoft. They do not follow that rule with the Surface, as an example. At least not the ARM ones. I don’t know what they do with the x86 ones.
Because they all need to petition redmond for a signature with the redmond key to be somewhat viable as a competitor to redmond. And, of course, it means supporting lots of redmond tech (fatshame32) just so you can boot. BIOS is crappy but so much simpler to deal with, that UEFI is not an improvement.
This is FUD. The UEFI Forum solicited proposals for companies to provide the root of trust for Secure Boot and all of the big security companies wanted a lot of money to host this. Microsoft offered to host it for free. Any other company could make the same offer the UEFI Forum would gladly accept it and add another root of trust. Not to mention that a requirement to get a system certified for OEM sale of Windows requires that the end user be able to install their own Secure Boot key. This means that anyone can sign their own boot loaders. You only need to have a signing key from the Microsoft authority if you want your boot loader, driver, etc, to run with secure boot enabled without having to load a key into the keystore on the machine.
As someone who lives in the UK afaict UK prime is better than us prime. AIUI US prime only gives free 2-day delivery while UK prime gives free next day or if you prefer nominated day on most Amazon-stocked items (some items seem to have more restrictive options, probablly as a result of being in different warehouses). Also unlike most next-day and nominated day services Amazon offers weekend deliveries at no extra cost.
Prime is great when you are doing something like preparing for an event when you want to order a bunch of stuff over a short period (but not all at once because you inevitablly find you need more stuff as the preperations proceed). I haven't yet found motivation to stay subscribed all the time though, I just buy a month when I want it.
I know amazon UK has a video service, but I dunno if it's better or worse than the US one, I haven't tried it personally.
I live in the US and can get free same day delivery on many items. If its a common item that is ordered a lot in my metro area they will deliver it the same day for free, up until noon. If I order after noon then I will get it the next day. Rarer and more expensive things typically come next day - though again this depends on popularity.
I won't comment on your other claims but, I do know that Google moved location services (as well as many other services) from the OS to Google Play Services in order to make them updatable separate from the OS. We've all seen the complaints about Android fragmentation right? There's a reason that most apps will run on any version of Android newer than 4 or 5 now and that's because of the way Google Play Services works in more recent versions of Android.
You mean the library is upgradeable through play or that it is a part of the Google Play framework? I mean that all location data routes through Play in stock Android. As in, you cannot just turn off location services for all of the Google apps. To use location services at all you have to give your location to Google. This has nothing to do with being an upgradeable library and everything to do with Google removing the ability to restrict location services from Google's own applications. I'm not sure if that is the case with AOSP but if the phone vendor uses Play then you have no choice but to give location data to google or not use location services anywhere.
Not only do I think this is disingenuous but he's also failing to point out that Google arbitrarily requires you to give them the whole kit and caboodle of personal data if you want to use certain features. Want to use the Play store? You have to enable location services that Google tracks. Want to use Google maps and actually store maps so you don't download them every time (whether you are on android or not)? You have to sign in and give over all your tracking data. And are there actual legitimate requirements for this? No - Google intentionally moved location services inside of the Play framework, it used to be completely separate and standalone. Do you need to be signed in to download maps to the local device? Of course not. But Google knows people want to use these features on their devices and the force you to give up everything to use them.
Quake had true 3D levels that can pass over one another. Quake had 3D adversaries. Quake had network gameplay with Quakeworld. Quake had OpenGL support with GLQuake that launched the GPU world, really starting with the Voodoo 1. Quake had translucent water, which was amazing the first time I ever saw it. And lastly, Quake is still the bar that any small platform must aspire to by answering the question, "Does it run Quake."
Man I always though Quake was garbage and that was when I stopped playing FPS games.
Ahhh one of my older brothers was a security guard at HP when they closed down a bunch of facilities over in Sunnyvale and threw away a bunch of hardware. We raided that trash for coax network cards and used Novell and IPX to play Doom and eventually Duke Nukem. I think that Doom allowed a 4 person multiplayer? I forget. Anyway, not long after that and I was playing Warcraft 2 over dial-up modem with my friend. And oh how the world changed when one of my friend’s dad, who worked at Cisco, got an ISDN line and we were able to download boatloads of MP3s and trade them all over school using CDRs. Man. Now I have gigabit fiber into the home and I can download gigs if data in just a few minutes. My how the world has changed.
LMOL first define what you compensation is then tell people it's not about the money. You have bills to pay. Employers bottom dollar pay. So money is everything. When you get paid beyond a certain amount, then and only then do you think, maybe I would like ping pong. But with student loans, rent / mortgage, family and what not, it's about getting paid.
When I finished school, I was definitely looking for professional growth, work life balance, and interesting problems to solve. In fact, AMD offered me a very nice salary to work for them when I finished. I turned it down, even though it was the highest paying offer I received. I ended up going to another company that was about the same size as AMD, but transitioning from analog to digital. At AMD I would have been one of thousands of software people. At this other company I was one in about fifty. I started out making less than 50% of what AMD had offered me. However, thanks to a bit of luck and hard work, I was leading projects by the end of my second year at the company. End result? I was making more than someone I knew who took the exact same offer from AMD I had received and I was gaining much better experience than they were, too.
I fully agree with you that the Slack client is horribly inefficient. But my point is that anything that does the same thing more efficiently in RAM and CPU across all required platforms is ultimately economically unmaintainable.
Electron does not have a monopoly on cross-platform development. In fact, it’s pretty late to the game. So this is only true if you define “economically maintainable” to mean that you can hire a bunch of high school kids in some third world country pennies to do the development. There are far better frameworks for this kind of work, but they require a more knowledgeable development team to use. Of course if you go the cheap route you’ll end up with an unmaintainable mess that is a pain to keep running. There‘s a risk of that happening even with an experienced development team, of course. But javascript is probably easier to shoot yourself in the foot with than C.
Hipchat supports all of those capabilities and you can serve hundreds of users from a single server instance that uses less RAM than the client side of Slack.
As of February 15, 2019, the client side of HipChat will use exactly the same amount of RAM as the client side of Slack because of Atlassian's deal with Slack to discontinue HipChat, including licenses for on-premises instances, and migrate all HipChat users to Slack. The server side of HipChat will no longer be available to the public at all, as Slack will host all instances on servers operated by Slack. On-premises users who cannot migrate to Slack will have to export their data and import it to some unspecified competing on-premises chat platform. (Source: "Slack Migration Frequently Asked Questions")
What's that got to do with anything? It means something better (from a resource consumption standpoint) existed and Slack is extinguishing it. But your claim was that no one had ever written anything like slack before and that is utter nonsense. Are you somehow financially tied to Slack? Because you obviously can't accept that the Slack client was horribly implemented.
[The features that proprietary web chat has over IRC] are all SERVER features.
First, I'd be interested to see what IRC servers support these features. Second, these features still need some sort of client integration, such as to request a particular history range or thumbnail image or to send an attachment, and I'd be interested to see what IRC clients have been customized to integrate with a server that supports these features. I agree that one could create a client to do these that's lighter weight than the existing Electron-based clients, but as of 2018, to the best of my knowledge, no one has.
Hipchat supports all of those capabilities and you can serve hundreds of users from a single server instance that uses less RAM than the client side of Slack. Do you need any more examples?
Funny. It's the native Mac apps that always seem to take up several GiBs of memory each. Run XCode plus a web browser and you're already past the available RAM on a Linux machine where I'm running half a dozen electron apps at a time.
Try again.
I've got 3 Xcode Projects open right now and it is consuming 161MB of RAM. Safari is consuming about 1.2GB. When I used to have to run slack for work, it would easily consume over 2GB of RAM. Opening Atom consumed 500MB of RAM without a single file open, though it did eventually slim down to 200MB. But I can see that Atom has a problem because RAM consumption is increasing by over 10MB every 10-20 seconds and it is sitting in the background doing nothing without anything open. It is now up to 350MB of RAM in just a few minutes.
There's little these apps do that vim and emacs and IRC haven't done for decades for a tiny percentage of these requirements.
Vim and Emacs I'll grant. But IRC out of the box doesn't do chat history, file attachments, link preview, voice chat, or channel groups (Slack "workspaces" or Discord "servers").
I know you keep going on and on about Slack's chat history, file attachments, etc but you're missing one thing. Those are all SERVER features. You do not need gigabytes of RAM on the client in order for the server to push you chat history out of a DB. It's not required for sending files, or any of those other things you're talking about. It's just plain sloppy programming and the fact that people consider Slack to be a commercial tool they're willing to pay money on is insanity. I'd be embarrassed if I helped deliver the slack client. It would never end up on my resume and I wouldn't want to tell any of my friends about it.
Oh, huh. That is pretty cool. I am not sure what those are, except for Skype, but sounds very good. I can feel my ironic beard growing.
What Tepples doesn't point out is that Slack, a simple IRC style chat interface, can easily consume multiple gigabytes of RAM and the more instances you have your slack client connected to, the higher the GB consumption soars. Discord also causes performance problems on my gaming PC when certain things happen. I'm not sure what event is causing it, but it temporarily hangs games while playing some sort of chime over my headset. And that's with a very high end gaming PC.
In fact, the throughput difference is so huge that switching from SATA to PCIe drops a compile time on one of my projects by 70%, So what did I end up doing? I ordered a Lenovo laptop that supports NVMe PCIe [...] I will never buy another Apple computer again.
Not even when asked to compile and test a macOS or iOS port of the application that you are compiling?
Actually, Apple licenses the source to one of the projects I work on and they compile it on... Windows! Though Windows is not my personal choice for any sort of work project. It is the only platform that has the build tools required, though the company that provides the build tools is starting to transition their tools over to Linux.