Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Reelya]

There's a very old saying "If it ain't broke, don't fix it".

It's pretty common that the "new and improved" version of something is worse than the old version. Just in case you wonder why people buy the new version, it's because they stopped making the old version.

[Naturegirl1999]

Huh
I thought that updates would fix things or add security. I have heard the if it ain’t broke don’t fix it before, I thought updates wouldn’t be done unless needed, I was wrong

[wierd]

The UX/UI developers have to justify their existences.  Once something is "good", they are no longer needed as heavily.  This weighs on their minds, and they feel the need to redevelop the wheel every year.

This causes user confusion as things move around without warning or rational explanation. It causes new bugs to enter the software. It causes new dependencies that break things.

[Kagus]

There's also the matter of how the military adaptation cycle moves considerably slower than the software development cycle. The series of tests, checks and paperwork required to implement something new take years to complete, so if they were in the business of trying to outpace security threats with new defenses, they'd be hopelessly and utterly behind the curve.


When I was in the Norwegian military, there was a great big undertaking with converting all standardized military computers (a task enviable by absolutely no one) over to the newly-approved updated operating system, which had just finished going through all the rigorous testing required to qualify it for use.

The new OS was Windows XP. This was in 2011.

[Reelya]

Huh
I thought that updates would fix things or add security. I have heard the if it ain’t broke don’t fix it before, I thought updates wouldn’t be done unless needed, I was wrong

That's the reason they're still using the same gear. It already works. If they bought new gear they couldn't be sure it works, and it doesn't actually do anything that they actually need. For example, if you need to hammer in nails then you use a hammer whose design hasn't changed in probably 150 years. You don't need to buy the latest hammer each few years.

Also a 1970s computer is much simpler and has much less code in it, there's just less that can go wrong and it's easier to prove that it's secure.

Look up attacks such as Meltdown, Spectre and Rowhammer. These are specific attacks that only work on new processors or on recent memory chips, due to changes in processor design
https://meltdownattack.com/
Meltdown for example was first fully disclosed in 2018, and is a vulnerability in roughly the last 20 years worth of CPUs.
https://en.wikipedia.org/wiki/Meltdown_(security_vulnerability)#Overview

Rowhammer is a memory-hacking technique, which only works on newer memory modules, because there's less physical space between the components, so more signal bleeds over between them, and you can steal data this way
https://en.wikipedia.org/wiki/Row_hammer

So, bugs creep in and may open up security holes and it may be decades before they're fully documented. If the Air Force had upgraded their gear with new CPUs from any time between ~1998 to 2018 the nuclear computers would have had the Meltdown vulnerability.

[Naturegirl1999]

Huh
I thought that updates would fix things or add security. I have heard the if it ain’t broke don’t fix it before, I thought updates wouldn’t be done unless needed, I was wrong

That's the reason they're still using the same gear. It already works. If they bought new gear they couldn't be sure it works, and it doesn't actually do anything that they actually need. For example, if you need to hammer in nails then you use a hammer whose design hasn't changed in probably 150 years. You don't need to buy the latest hammer each few years.

Also a 1970s computer is much simpler and has much less code in it, there's just less that can go wrong and it's easier to prove that it's secure.

Look up attacks such as Meltdown, Spectre and Rowhammer. These are specific attacks that only work on new processors or on recent memory chips, due to changes in processor design
https://meltdownattack.com/
https://en.wikipedia.org/wiki/Row_hammer

Meltdown for example was first fully disclosed in 2018, and is a vulnerability in roughly the last 20 years worth of CPUs.
https://en.wikipedia.org/wiki/Meltdown_(security_vulnerability)#Overview

So, bugs creep in and may open up security holes and it may be decades before they're fully documented. If the Air Force had upgraded their gear with new CPUs from any time between ~1998 to 2018 the nuclear computers would have had the Meltdown vulnerability.

Thank you for the information.

[Iduno]

These are the nuclear launch computers. If they updated them all the time, they risk introducing bugs. i.e. accidentally launching some nukes, or introducing backdoors for hackers. Just be glad they're not updating them every time there's a new Windows operating system, or running Windows Update on them in general, heh.

"I see you're trying to end the world. Would you like help?"

[wierd]

"Because our presidents have been getting less and less presidential with each passing decade (with increasing levels of forgetfulness), we have finally decided that the nuclear launch codes can be reset to a system default value using an easy button press combination at power on. While the missiles will not launch with the default codes, one can then set the codes to anything they like, THEN launch. There is a 30 minute safety window after which the new codes become valid. We expect no problems with this arrangement."

[Reelya]

Uh. Reality is ahead of your joke. The actual launch codes were 00000000. For real.

https://www.gizmodo.com.au/2013/12/for-20-years-the-nuclear-launch-code-at-us-minuteman-silos-was-00000000/

"Our launch checklist in fact instructed us, the firing crew, to double-check the locking panel in our underground launch bunker to ensure that no digits other than zero had been inadvertently dialed into the panel."

This ensured that there was no need to wait for Presidential confirmation that would have just wasted valuable Russian nuking time.

Yeah, so the supposed presidential unlock codes were just a fiction in practice. They pre-dialled all zeros (the actual combinations) into all the missile unlock systems at all times. They probably gave the president a briefcase full of fake numbers to make him feel like he was needed.

[wierd]

LOL. Awesome.

To counter useless numbers, here's a recent paper that I am nowhere near qualified to even make sense of, let alone critique--- but it makes a very fascinating claim.

Numbers may be multiplied with time O(n log n).

https://hal.archives-ouvertes.fr/hal-02070778/document

[Reelya]

Yeah, I came across that on Slashdot. The important fact is that the new algorithm is only better than existing algorithms for very long numbers. How long?

https://en.wikipedia.org/wiki/Galactic_algorithm

An example of a galactic algorithm is the fastest known way to multiply two numbers, which is based on a 1729-dimensional Fourier transform. This means it will not reach its stated efficiency until the numbers have at least 2^1729 digits, which is vastly larger than the number of atoms in the known universe. So this algorithm is never used in practice.

The link is to the Harvey algorithm mentioned in your document. It hits O(n log n) efficiency once you exceed 2¹⁷²⁹ digits in length. There's another existing algorithm that's more efficient for "shorter" numbers, but nobody yet knows the exact point at which the new algorithm becomes more efficient than that:

"The question is, how deep does n have to be for this algorithm to actually be faster than the previous algorithms?" the assistant professor says in the video. "The answer is we don't know.

"It could be billions of digits. It could be trillions. It could be much bigger than that. We really have no idea at this point."

[Zangi]

So... Murrican 5G.  Is it just a half-arsed cash-in/'look at us, we can do 5G like the Chinese too!(but not really)' kinda thing?

[Telgin]

From what I understand nobody has even codified what 5G is yet, despite lots of people claiming to support it or be actively working on it.  Kind of a head scratcher for me.

[scourge728]

We just need bigger and more powerful emitters, shove more signal so some can force its way through or something 

[Madman198237]

We just need bigger and more powerful emitters, shove more signal so some can force its way through or something 

You're right. But we should also use the highest frequency possible to maximize benefits, so I'm thinking we should use a nuclear x-ray laser for the most powerful and highest frequency information transfer.

It is completely flawless and absolutely guaranteed to transmit one important piece of information in a split second. Of course, the only piece of information it transmits is "DIE!" but hey, it's extremely effective at transmitting that information at very high bandwidth and nearly unlimited fidelity to the original message!