@emsenn

Restarting from scratch the whole computer stack is not possible in a foreseeable future. The thing is even if at the processor level things are broken (heartbleed, et al) not all processors are broken (RISC-V?).

I can not imagine things to change overnight, even if **some** people size the means of production. Like you wrote the whole (software) system is built around domination. Sizing production facilities will, in my opinion, change the people in power, not the system. In particular, it will not change people's minds Earth-wide.

FOSS was a step in a good direction. Because the system is the way it is. FOSS has put another oligarchy in power and created new monsters, but those kind of monsters existed before FOSS. The good thing about FOSS is that it exhibits, once more, that **together we are stronger**.

One can swap a processor with another, given enough work.

One needs to give some existence to an alternative system, and that should be done at every level of the stack.

edit: that should be done at *every* level of the stack

@theruran

@stman I understand. I was boosting @zig post for visibility since I think it's important to hear different angles on the topic.

yes, we need a decentralized economic system to support the decentralized information system development and stewardship. to the extent possible, these should be designed together.

I sent you a DM earlier about definitions. Lemme know and we can iterate on some things.

Peace!

/cc @emsenn

@stman @theruran @emsenn

Hello,

I introduce to you @mouloud who he is more aware of philosophical implication and sometime technical details of the actual the (sort of) system we have been working on.

I forward a very **serious** question from him, since the convo was not federated on his instance yet:

What is the point of money or crypto-money?

@zig @stman @theruran @emsenn

My question is indeed what is the point of money whether it is crypto or not. I know a little about bitcoin (proof-of-work: evil for climate so far) and Ethereum (algorithmic contracts, good idea as far as I understand, but still PoW).

The idea of a single source of thruth is neat, and would be useful to avoid lies or fakes in a distributed system.

@zig @stman @theruran @emsenn

Money used to be used a means of exchange something of value for the rich that is gold. That by itself shows how dubious money seems, because gold is almost useless in practice. Nowadays, money has mostly only virtual value, because people trust the system, and the people in power somewhat trust each other and they agree through the market on exchange rates.

@zig @stman @theruran @emsenn

Anyway, take of instance "carbon budget" of countries, it can be exchanged for money. And they that "carbon budget" can be used to produce new products.

With the money, a low carbon footprint country can bargain to buy some products.

During this exchange the low carbon footprint might have lost value because conditions of the exchanges and the dubbed added value to the products.

@zig @stman @theruran @emsenn

It is far fetched, but to me their should be no money, hence probably no crypto-money.

A single-source of truth is helpful, but I am not convinced it is necessary, and is certainly not necessary in a fully cooperative system with no evil.

Thanks including me in the convo.

@stman @zig @theruran

I agree with what you write.

Where does the code happen?

@stman @mouloud @zig Thanks for the discussion. I enjoyed reading so far!

Yes, and about code - that is the kind of knee-jerk reaction that people have nowadays and it prevents everyone else from understanding what they are doing. Documentation of every kind is key, unlike the prevailing software engineering practices that lack rigorous conceptual design development phase. Visual documentation is also of course important, and to maximize the utility it must also be an executable architecture model.

Well - if we can theorize another way of achieving an equivalent security model and utility to Bitcoin without the energy consumption, that would be incredible. As far as I know, there is no alternative yet conceived and the energy consumption keeps the system honest. And unfortunately, no one I have met in the fediverse thus far is qualified to theorize such an alternative. There are real engineering constraints and trade-offs that are glossed over in these kinds of discussion, and I doubt that billions of transactions per second is achievable due to laws of physics. It is my expectation that such a decentralized and trustworthy cybersystem will be slower in many ways but is nevertheless fast enough for us to get real work done and not just mindlessly consume Big Media.

P.S. come to hackers.town - we got 10,000-character toots!

@stman Time Sensitive Networks with timestamped everything sounds like great way to approach this problem :blob_gnikniht: @mouloud @zig

Follow

@theruran In such paradigm, simpler protocols, with fully distributed small mining capabilities in each node of a truly mesh cyberspace architecture, and within each microprocessor, can bring the time sensitive trust chain needed to replace the current blockchain implementation, with its energyvore mining. That's what I curently think.
@mouloud @zig

Β· Β· Web Β· 1 Β· 1 Β· 1

@stman @mouloud @zig a trustworthy distributed time server will be important when solar storm activity picks up and our GPS goes out for extended periods; and in scenarios when that GPS infrastructure rots due to lack of funding and the satellites fall out of the sky.

@stman yes you are right! I guess I meant server generally as in an entity that serves something. "global synchronous clock" maybe more accurate. in blockchain-speak it's a kind of oracle. @mouloud @zig

@stman The closest thing to this that I know of is Google Spanner which uses GPS and atomic (hardware) clocks to ensure global synchronicity. Its FOSS equivalent is CockroachDB which only emulates this functionality through the system clock synchronized by NTP. @mouloud @zig

@stman @theruran @zig

I agree that CPUs (about which I am clueless so far) might and certainly are geared toward some particular use, that do not always intersect the interests of the commons. And more generally, available tools are nurturing a particular social system of domination.

@stman @theruran @zig

What I want to write has two sides a) I believe there is a slow, smooth path toward libertarian socialism. Stronger claim, it is the only viable path. I believe a brutal change is not possible. Feeding the "revolution" meme as in the french revolution of 1789 is counter productive. If anything like that would happen, that is again the people sitting at the first row that will be taking advantage of it like bourgeoisie has taken advantage of french revolution.

@stman @theruran @zig

b) I agree there is a need to THINK about the new system but also the think about path toward that system. It is necessary to DOCUMENT and prove, if not necessary, at least that there is sufficient alternatives. And toward that goal, I am confident that one can not just reboot the system with a new operating system. That is, one need an upgrade path, and step by step hotfix, and swap existing cyber infrastructures with better ones.

@stman @theruran @zig

The upgrade path I am thinking about is similar to the one that happened with Linux: at the beginning is was toy, then it was sufficient, now it necessary.

@stman @theruran @zig

I might be wrong, but I consider code to be XXI literacy. We can live without code and go back to caves or something. But if we want to human science, say medicine, to progress with need computers and code.

In other words, yes you can try to convince people with free-form text, and think about new socioeconomical systems. But eventually, people will go back to their usual routine business as usual privative tools.

@stman @theruran @zig

I think one need to both document and code alternatives software that gives a cyber body to alternatives socioeconomic systems.

There was much work that has been done, but there is more that remains to be done.

@stman @theruran @zig

Toward that goal I have been working on alternative "operating system" that is supported by a DHT.

It will allow to not only copy but also improve the user experience of the www.

So far my work has been concerned about offering better alternatives to mastodon, mediawiki and to some extent google search engine.

I have not touched yet, other aspects like factories, production chains, logistic and many others...

@mouloud amazing! please share

and yes, I like this diagram. it's important to have an upgrade path or else the new system will be too great of a leap to adopt.

@stman @zig

@theruran

I already tried to share my work and I had little or no feedback.

I am working on new stuff (based on previous work) but it is not ready yet, like not ready at all.

@stman @zig

@xj9 @mouloud @stman @theruran @zig
I suspect it started with GNU trying to demand recognition as an OS, and claiming the kernel/drivers are just a component thereof
I'm not sure what it was like historically though (i.e whether there was a clear separation between userland and os/kernelland)

@xj9 @zig @theruran @stman @toast

The idea of what is an operating system is completely skewed by the domination unix. What I know of plan9 does not bring much new things in that space.

Things like ocaps or distributed ocaps makes me think we, humans, could be locked up outside the "OS".

Like the rest, it is built on what sort-of works (coming from the past), that is itself biased toward reproducing or implementing known systems like hierarchy (conway law).

@xj9

When I wrote unix, I meant inspired from unix, not necessarily forks. That is user-space, POSIX threads and processus, hierarchical file system, sh-like programming language to input commands, volatile vs. persistent memory etc...

We are biased toward what we already know.

@zig @theruran @stman @toast

@xj9

What is an hypervisor if not a wanna-be secured sandbox? Isn't sandboxing the purpose of rings, kernel space vs. user space separation, process, jail or even browsers?

@zig @theruran @stman @toast

@xj9

I just feel like there won't be feasible ways of using this hardware in the future. Energy will be scarce, and manufacturing facilities will stop producing critical components such as the RAM. We will have piles of electronics with no manuals/schematics and not enough power to run them but for a short time anyway.

@zig @stman @mouloud

@xj9 @zig @theruran @stman

I lost the URL, but there is at least one project that aims to build a minimal OS that would be very easy to port to any hardware. Well, now that I re-think about it there is many such OS.

The conversation went into several direction. One of them is to create new hardware with the idea of supporting a non-kyriarchy.

@xj9

I gotta think about that some more, but it falls within the scope. The silicon wafer manufacturing process being developed by #LibreSilicon is a key component.

I'm not sure what to do with the old hardware yet. The best I can think of right now is to use it to virtualize the system we propose.

Consider the path of no action - all that hardware will go to waste anyway. e-recycling is shit. :flan_shrug:​

@zig @stman @mouloud

@xj9

My use of the words "operating system" may be misguided, even if in the grand scheme of things I would like to get rid of the linux kernel or bsd. I do not have a completely clear idea of that yet.

> Plans are only good intentions unless they immediately degenerate into hard work.
> Peter Drucker

@zig @theruran @stman

@mouloud @theruran @stman TrueTime requires special hardware it was (somewhat?) replicated by cockroachdb and TiKV without the specific hardware. Apparently, foundationdb bypass that requirements completely.

Speaking of hardware and OS, read rumpkernel.org/

@stman

What is the most important constraint to ensure an alternative cyberspace architectures belongs to, and serves, equally, all the people on earth?

Thinking about this at first it seemed that decentralized governance of a copyleft architecture specification and implementation would be the deciding factor its sustainability. I don't think this is what you're asking though.

Even with a successful global multistakeholder cooperative that provides architecture governance, the architecture itself could still be shit. We see this today already and we should be aware that people like to pile on their CVs because they're careerists and don't care about doing great work. A co-op would look nothing like the Linux Foundation of today though, but that's a topic for another time.

The deciding factor in ensuring the alternative cyberspace architecture remains free to everyone forever -- I am convinced -- is that the architecture model is fully self-documenting. We don't see this anywhere except in aerospace / military industrial applications which are obviously classified and proprietary. Even there it is extremely expensive and not widely practiced. Fully model-driven engineering is extremely powerful but again, every model element needs to be richly documented so full documents and books can be generated from the model.

Part of the problem is the garbage tooling available. The rest is the capitalist culture of withholding information for profiteering. When only an elite cadre of engineers understand the system, everyone suffers. We see this today with our computing systems both proprietary and FOSS. Public domain or strong copyleft are of course important but useless without the effort spent documenting the architecture.

We may understand constraints differently, so I still don't know if I am answering your question as you intended. Constraints have a specific engineering definition and are dealt with differently by systems engineers - I am just learning they may be recast as optimization goals.

What is most dangerous centralized aspect in software kernels like Linux in terms of cyber-security model?

The MMU is provided by the microprocessor, so do you mean the virtual memory / page tables managed by the kernel? Linux' fault is that it's a monolithic kernel of C code. But even advanced separation kernels are not saved from the vulnerabilities of the microprocessor microarchitecture. Then the idea of an Abstract C Machine is all about sequential processing so that's why things like the MMU exist and there's your architectural bottleneck - Achilles' heel right there! So yeah - there is software maintaining security barriers around memory and that software sucks; and the microcode and hardware architecture of the microprocessor also tries to place security barriers but due its unmanaged complexity creates more side channels instead.

@zig @mouloud

Sign in to participate in the conversation
Mastodon

Server run by the main developers of the project 🐘 It is not focused on any particular niche interest - everyone is welcome as long as you follow our code of conduct!