Another interesting read in which Moxie Marlinspike essentially accuses Web3, as it currently exists, of being Cloud2, a platform controlled by a few with ‘decentralisation’ aspects largely smoke and mirrors. He also bemoans the lack of cryptography as a way of distributing trust, and the lack of attention to the interface between clients and servers - clients being users of blockchains, servers being the machines running the blockchains. They are two very different things.
It also looks like holders of NFTs will lose their shirts because of the lack of cryptographic safeguards.
As I see it, the article mostly vindicates the Safe model because with Safe:
Everything to do with trust and consensus is based on cryptography rather than infrastructure
There is a concerted emphasis on usability, aka the client server interface
The one caveat being that Elders do not become de facto servers by the barriers to entry becoming too high.
The hard parts are done! Now just aligning mirror segments which will take a while as they move incredibly slowly.
The really great news which came a few days ago was that the launch was soooo perfect in terms of getting just the right amount of velocity. If there would have been too much then telescope would have been lost, so they under-pushed, but too little pushing would mean that JWST would have to use more of it’s own fuel to get to the right spot. Well turns out that JWST won’t need to use much of it’s own fuel at all and the rocket gave it the best possible amount of thrust they could hope for.
The upshot is that JWST will have plenty of fuel for station-keeping – meaning it’s lifetime at L2 will be much much longer than anticipated - barring other problems in the future, JWST will be taking photos for perhaps a decade, or even longer!!
Initially 4 years then 10 years and now 20 years. Possibly irrelevant but an idea association came to my mind: isn’t JWST propellant reserve similar to Maidsafe cash reserve? It is in bitcoins and maids. A relative crash happened recently but if you zoom out you see that they have increased a lot since ICO.
Similarity doesn’t stop here because Maidsafe even had a refueling mission!
No criticism, lifetime increase is a good thing in both cases: It allowed more experiments with safe network, the same as it will with JWST.
There were some in the scientific community saying that NASA or others should now duplicate the JWST since the cost would be a fraction of the original cost. Maybe one or two billion, but the benefits to cost of having multiple Telescopes would make the project many times more profitable. The amount of available ‘telescope time’ multiplies for each telescope that is put up. And there is plenty of room at the L2 spot.
But Safe somehow fails with Moxie’s ending thoughts:
It tries to decentralize infrastructure, although it still seems to allow users’ machines to participate in the network (elders barrier to entry seems important here as you say)
It’s written in language, that has opinion of being not very easy to write software in.
That would be fantastic - won’t hold my breath though. If multiple scopes, then having one at opposite side of original’s L2 orbit looking at the same spot for space-diversity, then wow, resolution would increase dramatically.
What’s the alternative here? Needs to be participants hardware for ultimate trust. Adults and Elders also - what is an alternative? If there is a better one, please please provide!
Keyword there is ‘opinion’ … Rust is growing in popularity but no language is perfect nor will any ever agree on what’s best. IMO, it’s far far from the worst language they could have chosen.
This article seems useful to those looking at system design and software architecture in general, and particularly to how to integrate payment systems and the transactional framework which implements them. Quite a short read if you skim.
Probably it means that author is in contradiction to himself, stating that we need centralized infrastructure, and on the other hand – complaining about blockchain being too centralized. Perhaps adding proper (client-side!) encryption and anonymization to centralized side-services to b-chain like Infura, Etherscan would be the way to go? Or doing, what Solid does: relying on commodity hosting / VPS, but with some layer of additional anonymity like Tor on top? These are alternatives I can think of, but I also think Safenet is doing ok, and Moxie is not 100% right about this point. I believe decentralization can also be made right, if people don’t actually run servers and are truly able to fully participate in the network by running node on their existing hardware / internet connection.
True, but it also is evident, that ease of development wasn’t a key value for Maidsafe, when choosing the language.
Not sure that’s accurate. My understanding is that they chose Rust to reduce the overall code complexity and to make it easier for others to read and develop.
Correct. It was the reduction in sources of bugs in code and particularly mutli-thread handling. EG race conditions, memory mishandling, overloading of variables, and so on
RUST is great for OS and similar programming since a lot of the nasty tiny bugs are very difficult to introduce into the program since the code constrains so much sloppy handling.
There are a number of projects (including Linux, Windows) that are taking up RUST because it is better than other current languages.
The difficulty comes from the requirements to not take shortcuts etc.
IIRC David said Rust was a much better “systems language”. Before that people generally chose C++, and before that C, then Assembly, then machine code.
To say Rust is hard to program is naive IMO, it makes it harder to write incorrect code, and easier to write correct code than any of the others I listed. When programming systems which lots of other code, services and users rely on this is more important than when writing a single app, especially if the app isn’t critical like a game. You probably should consider Rust for critical apps like a wallet though.
It might be more correct to say it’s easier to write a game app than a wallet app, because you need to be able to achieve a much higher standard of reliability in the wallet. Doing that in an “easy to write” language like Python is going to be much harder than doing the same thing in Rust IMO.
So why write vdash in Rust rather than say Python? I think it would be perfectly fine and maybe easier for some to use Python for that, or nodejs etc. In my case I wanted to learn Rust, and tbh I prefer systems languages and now that I’ve got the basics I’d choose Rust again because it’s actually high productivity - far harder to introduce bugs - which to my mind is easier than something like Python. More head scratching to get the code written, less to get it working!
Ok, I get it, it’s easier to write some sort of software in Rust, than in other languages. Hmm fair enough for me, I was probably unnecessarily generalizing.
Essentially the idea here is that the semantic web didn’t take off because people don’t bother translating the data in databases into the html markup, too much extra work. But if the database itself is directly queryable on the client, as it would mostly need to be on SAFE, then you don’t need special markup to add the raw data to the html, because the database is already available for anyone who wants to use it (though by itself doesn’t help with the issue of data from different sources being in structured in incompatible ways and using incompatible formats)
