@cmccullough Maybe people/orgs want to no longer be judged for their theme and we need something standard to be arranged for govtech and small businesses? We used to have HTML framesets and they made sites need less CSS… Now they are forbidden, in part because responsive design

What I also mean to say is that this is always about the prestige, the first reactions of upper management and of commonest of people. Things need to say they are modern while retaining the simplicity

@cmccullough Also can we have the internet not be so hostile to us through reducing our vulnerability / attack surface and have a readymade operating system that has the firewall and hardening more integrated with putting up services? Maybe we can’t have that because of how stalled we are in operating systems research for almost half a century?

@[email protected] @cmccullough It's not that the research is stalled.

It's that the development of old and broken by design systems that are currently popular gets enormous amount of funding, while modern designs are either one-off project for one high reliability application or developed as part of someone's thesis.

As a result you do have modern OS designs that have implementations that run on up to 1 piece of real hardware (sometimes only virtual), and you can run up to 1 application on them.

Support of every piece of hardware needs someone writing or porting a driver. Running an application requires that someone adapts it. Multiply that by thousands of pieces of hardware and thousands of applications and tools and the task is nearly impossible. If you are an enthusiast and willing to spend a lot of time on it you can demonstrate running an application on a piece of hardware that you specifically choose to be easy to support. For general 'just give us more secure OS to run our applications on' it's useless.

One recent attempt at popularizing more modern OS design is Harmony OS by Huawei. Similar to AOSP there is an opensource and fairly useless release that you can look at. Unfortunately, most of the detailed documentation I could find was Chinese which I don't understand.
From the bits I do understand it sounds like it's similar to Windows NT and OS X in that the OS core is relatively modern (for its respective time) and system services might make use of more modern API. On top is an environment that emulates something that is close to the currently popular OS environment. As a result most applications will not or possibly even cannot use the core OS modern API, and as a result cannot provide better security, either. ​​

@bunny @cmccullough Thanks for cool facts and observations. Although I think is what such stalling can realistically be expected to be, at its worst even; point by point you chime along with the point I stated. Also not just me claiming that, I’m actually repeating what I’ve heard and read from others

@[email protected] @cmccullough There is a difference between research being stalled, not having any idea how to design a system to be more resilient, and the current situation in which existing research results are not widely used.

Sure, it could take a while - see how long people were talking about e-ink before you could go to a store and buy an e-ink display. But modern OS designs were available even before people started talking about e-ink. And here we are. You can get an e-ink screen but not a secure OS.

My impression is that software development is somewhat like science in this regard.

There are some corporations, and sometimes government agencies that decide by their wallet what research is funded. Sure, you are not forbidden from doing research that is not funded but without the funding you only get so far.

And even if software is 'opensource' or 'free' similarly you can only get so far without funding. Not only will your development be much slower but also the software that has the most funding becomes the yardstick by which your results are measured ​​

@bunny @cmccullough I got reminded where I got the notion from, and it was either a talk with the utah2000 as slides or just maybe the following read mirrored somewhere:
doc.cat-v.org/bell_labs/utah2000/
doc.cat-v.org/bell_labs/utah2000/utah2000
("Systems Software Research is Irrelevant" by Rob Pike)