rakoo

Indeed, import/export is one of the most important design point centering control back to the user, and even better is abstracted storage of data: if I had a single OPML file accessible from everywhere (with auth, of course) and all tools took the universally accessible path, I could change feeds in one tool and automatically see the result in another without fussing out import/export buttons and procedures.

For reference, and not implying it's better or worse than your work OP, I've pleasantly used Fraidycat (https://fraidyc.at/) in the past. It's a webextension, so completely local, and also incorporates the idea of having a "calmer" experience: no infinite list of links to check, different update rates, ...

I love your philosophy page, OP ! (https://jamesg.blog/2024/11/30/designing-a-calm-web-reader/)

That's exactly why pubsubhubbub, now websub, has been created. Around 10 years ago.

So you don't test your backups to Glacier ?

> while Elon Musk created Tesla, SpaceX and Starlink the EU [created] some regulation that mandates that bottle caps must hold to the bottle

At least the EU made something useful

We value your body integrity. We and our 1492 partners would like to stab you.

Use ublock origin with the "Cookie notices" custom lists. Not explicitely accepting cookies is legally the same as refusing them (now, whether websites actually respect that is the opening keynote of the Naiveté conference)

The big central place is still the PLC directory that effectively means all accounts are centralized at BlueSky, even if your posts are not. They haven't planned to make it any decentralized in the future.

The problem seems to be that each individual split part is valid in itself. This means that the entire file, with the central directory at the end, can diverge from each entry. This is the original issue.

Because, as said before, zip managing splits ends with two sources of truth in the file format that can differ while the whole file still being valid

That's a very bad way of solving that issue. If transmission is a problem, either use a proper retry-friendly protocol (such as bittorrent) or split the file. Using hacks on the data format just leads to additional pain

The ironic bit is that conceptually Bluesky is RSS with WebSub <https://en.wikipedia.org/wiki/WebSub>

- PDS are websites with an RSS feeds, each a publisher

- Relays are WebSub hubs aggregating many sources into a central host

- App views, labelers, Feed Generators, whatever are subscribers being alerted when a new entry is received, making their internal sauce. They're also hubs for pushing content to any step that comes after

- PDS are at the end, subscribers of labelers, app views and feed generators. They make their internal sauce to have a nice social-oriented UI.

A properly decentralized, boring-tech Bluesky can take this form. Steps are additive, not all of them are needed

1. A single, simple server that can receive and emit RSS feeds. Emission of RSS feeds must be able to include content from received entries. Directly subscribe to the people you want to follow, you're done. Social readers <https://indieweb.org/social_reader> do something like that with indieweb formats.

2. A network of WebSub hubs to more efficiently get and distribute everyone's content. superfeedr.com is one of them, https://switchboard.p3k.io/ is another one, but we need many more. Also RSS feeds need to have a server-side that sends a notification to the hub when something is published. Having hubs make searching and having a general view easier. In fact we already have something like that, it's called Planets <https://en.wikipedia.org/wiki/Planet_(software)>

3. Any filtering can happen on top. "People with blue hair", "Posts without tuna in them", any algorithm is anyone's to build

All the layers already exist, but like any social endeavour it's all about the network effect. Having a simple thing to install where you can post, follow and search will make wonders

This is exactly why WebSub, formerly Pubsubhubbub, was written. Publishers can push to a relay only when something new happens, never having to be polled.

It was created more than a decade ago.

I don't see the fundamental difference. Both systems work under expected conditions and will crash parts of it if the conditions don't happen. The scales (and thus the visibility of bugs) change, the technologies change, but the architecture really doesn't. Erlang programs are not magically devoid of bugs, the bugs are just not creating errors

To be fair this is what the BEAM vm structures everything on: If something is wonky, crash it and restart from a known ok state. Except when BEAM does it everyone says it's brilliant

"I tried this thing 12 years ago and it sucked, therefore it still sucks today and will suck forever"

Market share only matters if there's a market, ie concurrent proposals. In an open world, concurrent proposals make no sense: let all applications talk with any other, let all protocols take inspiration from each other to build the best infrastructure.

There's nobody to talk to because companies explicitly make it so, not because a protocol is better than another

Cycling is not dangerous because of nature, it's dangerous because of cars. It's not the same as "skydiving is dangerous" where the nature itself goes against you.

What matters is what we want to do as a society: leave the cars where they are as some kind of unmovable force of destiny, or actually manage them to not make them dangerous.

Unsurprisingly this is something you could do in Plan 9:

    /path/to/file:/from.*to

opens the file and sets the selection to the range of text between "from" and "to". Bonus: you can use regular expressions.

It pains me to see the web having strayed so far from its focus on interlinking. It's easy to make a parallel when on one side device makers have taken control back by removing access to files and putting applications first (this is the only model on mobile), to the point that those makers have enormous control over apps' existence, and the web where applications are now the only way to access content, to the point that content may or may not available if the right API exists.

Seconded, and I'll even add that you can buy second-hand laptops, they even have an included battery that will help you for short outages.

But more importantly, buying second-hand anything is orders of magnitude better than a new Raspberry Pi, because most of the environmental impact is on the building of end devices. Re-using an old laptop you don't use anymore furthers its life. The impact of usage, whether it draws 2W or 50W, doesn't change much on that front.

In any case, please stop buying new things, and re-use existing hardware instead of inducing even more pollution.

Raspberry Pi at the beginning: we want everyone to have one !

Raspberry Pi now: only geeks actually buy this

It's perfectly legal if you install a device the owners fully agreed to install and signed a 3-days long licensing text explicitly saying it that they didn't bother to read. Which is exactly what Facebook is doing.

That's the typical reply from someone who doesn't have a life on Facebook, and it's nice for you but you can't extrapolate the lives of literally billions of people just from your own, singular example.

People have all their friends and family on Facebook and still want to talk to them. People have orgs, businesses, support groups, anything on Facebook. They don't accept the garbage conditions because they decided, but because it's their only way to do it.

If anything, lazygit has shown it is possible to have more natural workflows on top of git as demonstrated by https://www.youtube.com/watch?v=CPLdltN7wgE, so it is indeed possible

I still don't think capitalism will help in solving this issue, if only because putting a cost on the environment will always be fought against and never integrated: there is no interest in this. And I also don't think that a strong State will implement strong checks on capitalism, if only because States are inherently dominated by capitalist powers.

If we want to take it into consideration, we have to take the matter in our own hands, away from the State, away from capitalistic interests.

Your monitor manufacturer isn't on the list and also wants a piece of that fresh brain meat

There's a big difference with a full distribution like Tails and using a private-by-default browser like Librewolf.

> Serving websites is an area where capitalism’s promise of achieving efficiency of resource utilization through economic incentives probably actually works, via shared hardware.

Most of the environmental impact happens at the building of hardware, so putting even more websites in datacenter only increases environmental impact. All Capitalism has done here is completely forget that impact by telling you it's ok to do always more.

What we collectively need to do is not find a better tech, it's reduce the total amount of tech we use, and that starts with questioning our actual uses. We don't need 24/7 available websites, 90% availability is more than enough. In fact, more sites should only target 90% availability, and if we were serious about tackling this we'd look towards 1. using old computers to the death because they still can and 2. doing more offline-first. Putting your website in a datacenter is exactly what the Jevons Paradox (https://en.wikipedia.org/wiki/Jevons_paradox) warns about. Its environmental impact keeps exploding year after year and it's not with that mindset that we'll make a dent in it.

(Remember, the environment is not a resource, so Capitalism doesn't care about it)

So they'd violate your privacy and re-centralize the web for free ? That's not exactly a rebuttal

From TFA:

> You can be finished with your project whenever you decide to be done with it. And "done" can mean anything you want it to be. Whose standards of completion or perfection are you holding yourself to anyway? Forget about those! Something is done when you say it is. When it's no longer interesting. When you've gotten a sufficient amount of entertainment and experience from it. When you've learned enough from it. Whatever, whenever. Done is what you say it is.

If you already say the goal is fitness, you're not doing SOFA. The whole point is to accrue experience, experiment, discover, not a predefined state. You don't SOFA a specific sport to lose weight, you SOFA when you want to find a sport you like enough that you will be able to do it regularly.