Most applications are DNS-routed and DNS is the security layer for all things HTTP.

So this feature really is for old and stuck technologies and bad designs that are hard to change.

I'm glad they did. Bad designs deserve security too. But they deserve redesign more.

"it's rather difficult for an attacker to actually connect me to a different IP address"

Is it really?

Have you run a router or honeypot?

The problem is not that C, one of the most successful languages ever, is not "memory safe". The problem comes from governments trying to add beaureucratic nonsense into programming language selection and then failing because fixing coding problems with language selection just produces more problems.

Torvalds pushed back against C++ for decades, using pure C in building objects, event handling and memory management. Clear guidelines and many eyes on the code worked better than the numerous attempts to use C++ trickery and now Rust, to fix coding problems.

Hmm where have I seen this before? Perhaps in aviation where low piloting experience is 'fixed' by more and more features of flight controlling computers.

Leave C as a low-level systems language that lets you do anything with the memory. Create a new language with all 'memory safety' certifications and let the commercial world chase that tail for another decade while in the backend they continue to run Linux/BSD because it just works.

Golang and python exist happily as higher-level non-system langauges that work great for large groups.

"Linux is a cancer that attaches itself in an intellectual property sense to everything it touches."
https://www.theregister.com/2001/06/02/ballmer_linux_is_a_cancer/

https://www.zdnet.com/article/ballmer-i-may-have-called-linux-a-cancer-but-now-i-love-it/

Ballmer doesn't head MSFT but not all top managers, developers and shareholders have been swapped out.

I've been involved with Linux/GNU/OSS from before MSFT took notice. I did not expect to see them so much in love with Linux so soon.

I hope to see Windows fully die and its internals be replaced with OSS-sourced code, before I fully die.

I much agree to this. Drivers, BSP, documentation and ecosystem can be more important than the SOC itself.

RPi has the stability and ecosystem (openness questionable)

Allwinner has the price and availability (stability and documentation are horrid, unless you can read Chinese and are located there)

NXP/SAM/Ti have kickass documentation and drivers relatively speaking, don't expect it to be cheap and always available.

They may all be Cortex-Ax, but their market usecases are mostly different.

The ram size is a disappointment, but hopefully more will be released later.
The price too. I think a 16GB version plus an M2 hat will take us closer to $150 or more.
I'm glad it uses less power or otherwise covers the overheating issues of version 4.

I personally use it for lower level code and hardware integrations and would have loved an embedded FPGA, M4 core or even the PIO from the pico against its GPIO. Give us a bunch more connectors in front of PIO/FPGA modules that sit in front of the GPIO to open up a lot more possibilities. This way they can still use the Broadcom SOCs.

I'll buy this version 5 if a 16GB version and an M2/nvme interface together add up to $100 or less. Otherwise it is too incremental for most of us.

Even though there were a few other server chips, until the M1, ARM was seen as an embedded CPU architecture.

Suddenly it is beating out high end x64 in various metrics and definitely in terms of performance per watt. And this is just ARM's entry into the field.

Right now Intel, AMD, Samsung, Qualcomm are resetting their 3- year and 5- year plans. They've all been dragged into high end chip markets.

I'm looking forward to the next few rounds of Neoverse. They're looking real good in public cloud as well. x64 servers will soon be placed in legacy racks along with Ultrasparc and Power series servers.

These companies should thank their gods that Apple only intends to sell Mx chips in Apple hardware and not sell them en masse as OEM chips.

Well she is mostly right. C based ABI is the standard between binaries libraries and OS. If you need a return of multiple variables from a function you must return a pointer to a struct as per C. I do not necessarily agree that thatâ€(TM)s a bad thing. C interface thus far is the common denomintor.

I disagree about hardware no longer being designed for C. Look closely at ARM, MIPS and RISC-V. Register pointers and symmetrical registers makes C implementation much easier. Hardware is also being designed for C

I remember the comments on Slashdot about one day Microsoft will release Linux distros of their own. Back in Ballmer days.

I can't believe it but here we go.

Are the FreeBSD people still fuming?

.. between:
- Pc Jr
- thin membrane 'keyboards' on some older devices including some gas stations
- the second last iteration of macbooks

I think macbooks have been at par with the PCjr keyboard until the most recent version which got slightly better.

We are reading fewer books, and that's a good thing.

At the advent of printed books, they were written by knowledgeable people and this knowledge was transferred to people who knew less about the subject.

Today you have far too many people writing books to make money or because it is fashionable in some circles to be an 'author'. A book or movie is strictly a one way exchange of information. Most online mediums are two-way, at least because of comments. I knew one guy who wrote books and sold them through Amazon with catchy titles like how to raise a teenager, without having raised one himself. Try posting that in a parenting forum. Or try lying about a highly technical subject in a place like Slashdot (or a forum specific to that subject matter). Many subjects are more accurately presented in Wikipedia than in Britannica (I personally know of some of them).

Book to online is a medium change. The real problem is not the number of books one reads. It is the signal to noise ratio online, and the lack of skill to navigate just that.

This is yet another statistic that finds a certain pattern helps weight loss.

Might it not be possible that just the act of monitoring a person's diet makes them reduce weight?

I lost most weight a few years ago when I aimed only to monitor and document what I ate. I had a goal of absolutely not dieting. But just the aspect of monitoring and thinking about food consciously made a bigger difference than all previous diets I had tried.

In a weird way it reminds me of Heisenberg's Uncertainty Principle.

Is that something that came after Webcrawler?

I used Webcrawler for years, especially for searching for mp3s on ftp sites.

Aircraft are not designed to be as reliable as possible by throwing in unlimited amount of redundancy. Usually if an environmental factor affects a sensor, it has a very high chance of affecting other similar redundant items (Remember the birdstrike and landing on the Hudson?).

In fact this is a frequent item of discussion when one is deciding between a single engine airplane and a twin. A twin engine has double the chance of AN engine failing and dramatically less chance of both failing at the same time. Unless an environmental factor was the cause of engine failure. And too many deaths came from just one engine failing and pilot trying to fight the remaining engine instead of focusing on a forced landing.

Most aircraft systems (check out the technical differences between non-TSO and TSO GPS, steamgauges etc) are designed to INDICATE a failure and stop working when there is a failure, rather than try to be as reliable as possible. This way, the pilot has an option to switch to a completely different system. Of course you need some redundancy, especially if the other item works in a different way.

So 5 AoA indicators would not be as good a design as a single AoA indicator that can reliably flag a failure. Since it can't two indicators can show a failure by showing the disagreement between them. Just fly through icing if you have 5 indicators and you'll see why that's not as good an idea. Once a failure is flagged, a pilot can disable systems that depend on it (if not done automatically) and remain in control.

Ideally a transducer that does not depend on vanes (maybe some kind of movement/capacitance sensor that can sense the movement of air along its surface) would be a perfect backup.

It is silly to blame documentation for everything.
If you're a pilot too, what would you do if you feel a runaway trim is happening?
Id' disable any system suspected of causing problems and manually fly the thing (smaller airplanes). With runaway trim or what feels like runaway trim without autopilot, there would be only one thing for me to disable here. Do you want a manual checklist to specifically tell you that or you wont do it?