This crap just never ends. It's amazing how things can go wrong, and it's irritating how there's zero accountability. In fact, there are armies of geeks out there to defend their pet tech from reasonable criticism, or just defend programming/tech in general.
Semi-related issue of my own:
I haven't been able to log into the iCloud.com website on my Macs for months. It started with my iMac, then also my MacBook Pro. I don't think iOS will let you even try, but by now I am sick of trying. It's not just me, as I've seen people post about it elsewhere (and they apparently had to call Apple Support, which I'm not keen to do just out of how much anxiety these kinds of obstructionist phone systems generate for me, though it supposedly should be a free call because it's about services).
So far, no news sites want to make an issue about it to embarrass Apple into addressing it. Luckily, I don't currently NEED to log in to iCloud.com, but some day I might. I assume I will have to use someone else's computer. I don't even get the login box from my Macs.
Give me a break. Name one piece of software that has been validated - ever - to be 100% free from bugs. We'll wait.
I'm a software engineer and mistakes happen. You ever make a mistake in your job? How's about applying that accountability to you as well? Should we fire you from your job?
This is what testing if for, and even then it's impossible to test for every possible scenario. Hindsight aside, it's been found and dealt with. Move on.
Considering the complexity of coding such a crucial subsystem like iCloud, I'm surprised it hasn't happened more often.
I'm not excusing sloppy code by sketchy software coders and trying to pass it off as production-quality products. iCloud has been pretty solid and depended-on by many.
Off the top of my head, INTEGRITY (a POSIX-compliant RTOS), seL4 (a reimplementation of a lot of L4 for embedded operating systems; Qualcomm's cell modems run it), and PikeOS (a commercial fork of L4 as an RTOS with hypervisor functionality) are all formally verified and are therefore mathematically proven to be free of bugs. Parts of VxWorks (another RTOS) have been formally verified, but not the system as a whole.
I call bullshit. The mathematically proven methods only consider what the mathematicians conceived of during testing. If you are talking statistics, then yeah - you can prove REALLY close down to zero bugs & maybe call it zero, but it still isn't proven to be zero.
In as much as proving such a negative is tantamount to impossible. But in which case the original ask "Name one piece of software that has been validated - ever - to be 100% free from bugs" was unreasonable. There is certainly software out there that has fewer bugs and has been verified and proven to a much a higher degree than Apple's. "Mistakes happen" is a pretty lame response to pretty basic coding errors in type validation from a company with Apple's resources.
This crap just never ends. It's amazing how things can go wrong, and it's irritating how there's zero accountability. In fact, there are armies of geeks out there to defend their pet tech from reasonable criticism, or just defend programming/tech in general.
Semi-related issue of my own:
I haven't been able to log into the iCloud.com website on my Macs for months. It started with my iMac, then also my MacBook Pro. I don't think iOS will let you even try, but by now I am sick of trying. It's not just me, as I've seen people post about it elsewhere (and they apparently had to call Apple Support, which I'm not keen to do just out of how much anxiety these kinds of obstructionist phone systems generate for me, though it supposedly should be a free call because it's about services).
So far, no news sites want to make an issue about it to embarrass Apple into addressing it. Luckily, I don't currently NEED to log in to iCloud.com, but some day I might. I assume I will have to use someone else's computer. I don't even get the login box from my Macs.
Give me a break. Name one piece of software that has been validated - ever - to be 100% free from bugs. We'll wait.
I'm a software engineer and mistakes happen. You ever make a mistake in your job? How's about applying that accountability to you as well? Should we fire you from your job?
This is what testing if for, and even then it's impossible to test for every possible scenario. Hindsight aside, it's been found and dealt with. Move on.
Considering the complexity of coding such a crucial subsystem like iCloud, I'm surprised it hasn't happened more often.
I'm not excusing sloppy code by sketchy software coders and trying to pass it off as production-quality products. iCloud has been pretty solid and depended-on by many.
Off the top of my head, INTEGRITY (a POSIX-compliant RTOS), seL4 (a reimplementation of a lot of L4 for embedded operating systems; Qualcomm's cell modems run it), and PikeOS (a commercial fork of L4 as an RTOS with hypervisor functionality) are all formally verified and are therefore mathematically proven to be free of bugs. Parts of VxWorks (another RTOS) have been formally verified, but not the system as a whole.
I call bullshit. The mathematically proven methods only consider what the mathematicians conceived of during testing. If you are talking statistics, then yeah - you can prove REALLY close down to zero bugs & maybe call it zero, but it still isn't proven to be zero.
For the third time, mathematically proven to be free of bugs. They match their specifications exactly. In technical terms, this means they have no undefined behaviors. This is not statistical inference, it’s concrete, verifiable proof that a program does everything in the specification and nothing which isn’t in the specification. The proof of this is as solid as a proof that 2+2=4, 2+2≠3, and 2+2≠5.
I don’t know why this seems to be so hard for people to understand. Software is math, and you can use math and formal logic to prove things which are incontrovertibly true. To use an earlier example, we can prove mathematically that it is impossible to tell whether certain programs will ever stop running by examining them ahead of time. This is actually a special case of Gödel’s first incompleteness theorem, which is one of the foundational discoveries of formal logic.
Comments
I don’t know why this seems to be so hard for people to understand. Software is math, and you can use math and formal logic to prove things which are incontrovertibly true. To use an earlier example, we can prove mathematically that it is impossible to tell whether certain programs will ever stop running by examining them ahead of time. This is actually a special case of Gödel’s first incompleteness theorem, which is one of the foundational discoveries of formal logic.