Still, your ideals mirror mine quite a bit.
I have long been against discrete GPUs and specialized APIs/units, which have led to fast triangle rasterization at the cost of the beautiful and creative flexibility pure CPU coding once allowed. In fact, I'd be happy with a step backwards in graphics for this cause. I don't think gameplay has advanced significantly from the days of Doom and Duke3D (at least with regard to graphical improvements), and the more advanced graphics have led us down a rabbit hole of production costs and complexity. Quite a few people enjoy coding on the Pico8 for this reason. (http://www.lexaloffle.com/pico-8.php)
I'd really like to build a cohesive UI/UX experience with fairly rigid standards and an eye towards elegant programming implementation, data-binding, fluid layout, etc. Distributing an application should be a no-brainer, but at the same time programmers should not be handcuffed by over-the-top security constraints. Rather than go nuts trying to satisfy every hardware combination, take a more console-ish approach with a sane set of limited hardware.
You could not please everyone, but they are trusting you to make the hard decisions. Which is another Jobs thing (I swear I am not a fanboy, more of a Woz guy, but many of his philosophies are relevant here) https://www.youtube.com/watch?v=65_PmYipnpk
Annual turtleneck pitch mandatory :)
What do you think about Qubes on a Skylake Skull Canyon NUC (small desktop with Iris Pro and Thunderbolt) as a desktop research platform? Qubes/Xen would allow existing Linux/Windows to run in parallel with an experimental OS. There is a cross-OS compositing graphics stack (like VMware Workstation Unity) in Qubes that tries to balance security with usability - could be extended.
Qubes also has hardware-rooted measured launch ("anti evil maid") to verify the integrity of the TCB, with keys under user/owner control. As systems get more complex, we need boot-and-runtime integrity measurement to ensure components have not been unexpectedly modified.
An interesting area of research is "data sovereignty" as a counterpoint to "surveillance capitalism". We need a modern storage API that shares the data portability goals of the EA/Amiga Interchange File Format, http://www.martinreddy.net/gfx/2d/IFF.txt & http://www.fileformat.info/format/iff/egff.htm.
As new data rendering capabilities are added to the OS, apps get support for free, as IFF delivered on the Amiga. In exchange, apps cannot lock user data into proprietary formats. This data interoperability then enables multi-app data analytics and UX compositing, with OS enforcement of user ownership and policy, rather than vendor or cloud policy.
Since app/service vendors are impermanent, standardize data interchange APIs with OSS implementations, ensuring the survival of user data. Combine this with KSWorld/Rebol ideas about DSLs for OS-enforced user extension of apps (think Excel declarative programming), and "apps" become data pipeline components, instead of fixed-function GUIs where users have to plead for changes.
I like the form factor of many NUCs but would be willing to go even bigger (maybe up to the size of a tactical suitcase, something like this but with user-friendly design: http://www.governmentvideo.com/Portals/0/IMT%20Suitcase%20Re...).
I am the least qualified person to comment on security, but I'll give my viewpoint for what it is worth: For me, there are two types of security: passive and reactive. Passive being that secure measures are in place to prevent foul-doing. Reactive being that threats are identified as they are created and appropriately eliminated. In the real world, our law system is passive and reactive. We are allowed to buy guns (but only by going through a passive security process [1]), but (severe) punishment is issued for misusing them (reactive security). In a way, reactive security functions as passive security - i.e. people do not steal because they are afraid of going to jail.
I believe in a little bit of both. I do not like security that obstructs users from doing what they want to do, especially if there are ways to circumnavigate that security anyway. A large problem I see in the software world is that we are too afraid to enforce security reactively, or the laws are simply not in place to be able to effectively do it.
([1] unless via a blackmarket/etc)
In the experimental phase, you want security to create strong isolation between the open area and the rest of the world, to maximize freedom within the protected area. As experimental code matures and wants to interact with external systems, and (possibly) the entire world via public internet, we can whitelist expected behavior and impose contextual limits, to regulate transitions across the biological system/world membrane.
One question is, who determines policy for non-experimental systems that interact with the world? Centralized app-stores are a blunt policy instrument, not to mention their propensity for business conflicts of interest. On the one hand, we want to encourage reuse of known-good policy defaults, which have been proven by other users against known threats. On the other hand, the long tail has many corner cases and even non-expert users can search the web to tweak config/policy settings to manage their own policies.
One techno-political precedent is CSS style sheets which at least tried to model a separation between "publisher" and "user" style sheets, acknowledging (a) the conflict of interest, (b) that users can override publishers.
What if every policy layer in client systems involved late-binding runtime negotiation among different stakeholders? An unplanned variation is happening with ad-blockers and anti-adblockers and anti-anti ... what if we developed DSLs for the purpose of efficient runtime and contextual policy negotiation? That would move us a lot closer to behavioral and contextual policy rather than semantic central policy.
While I know of quite a few people (at least 4) who owned, used Amiga, the Amiga was always the rich-mans computing option. Most home computing at that time was Commodore, PC even Trash-80s.
I'm probably viewing the Amiga through too much of a contemporary lens and just being too critical. But, at the time the Amiga was out and selling, I recall thinking it was a wacky piece of kit. It was brilliant but weird hardware with rough edges on the software side, and that wasn't what I was looking for in a computer.
Like this, maybe?
https://en.wikipedia.org/wiki/AMD_Accelerated_Processing_Uni...
Part of piling all the features into the CPU is it gives old software a chance at longer legs. Modern PC's have the GPU, which seem to have morphed into special-purpose general processors through herculean API and driver efforts. I do wish the GPU could fade into the main CPU though, it feels like a sidecar concept until it does. CPU vendors and OS vendors need to do a better job at adding features to new software while supporting old hardware. The planned obsolescence of hardware in newer operating systems seems foolish to me.
A ruthless productivity UX is my "dream". I envision taking a non-mainstream OS like Haiku or the open source Solaris and building a new desktop environment for it. Standard hotkeys for actions across applications that go far beyond copy/paste, a very specific and detailed User Experience Guidelines that ONLY target desktop/large screen applications, and a fixed, standard desktop theme and look-and-feel that will only be updated once per decade. I would strongly encourage users to choose not to reskin the desktop or customize their keystroke mappings. Productivity demands that a standard interface is presented to all users who sit down at the machine. I would be accessibility-first, and all standards and conventions would work magnificently with screen readers and magnifiers.
No chasing fads, no reporting user telemetry back to the mothership, privacy is a given. My unparalleled stability and north star vision will be my only real selling points. I'd charge good money for my OS, and my slavish fanboys would pay up for every upgrade and t-shirt I put on offer after my yearly mock turtleneck pitch. That's my dream, at least, hehe.