JepZ

4 years ago, I assumed, that by 2021 we would have about 50% IPv6 adoption:

https://news.ycombinator.com/item?id=14855347

Now it looks like I was wrong and we got just about 33% and the curve seems to flatten already:

https://www.google.com/intl/en/ipv6/statistics.html#20

If you run a federated server, not all contacts might use the same server. With end-to-end encryption that doesn't really matter.

EDIT: Moreover, trust is not binary. So while your family might trust you, maybe your dad doesn't want you to be able to read everything he writes your mom or so (you get the idea).

Sadly I don't. I have mine running since a while, and some things might not be state-of-the-art anymore.

The server doesn't really have to support end-to-end encryption as that is part of the clients (in fact, there are some server-side extensions which have to be present, but those are mostly enabled by default).

Afaik, the default ejabberd configuration is very close to what you need, and there is just one part that you have to remove to enable OMEMO [1]. I don't understand why but recently the ejabberd devs introduced that part to their default configuration which makes it harder to use end-to-end encryption.

Nevertheless, if you are very interested in a detailed guide, I could write one as I am thinking about setting up a secondary server as a testing environment.

[1]: https://github.com/processone/ejabberd/blob/master/ejabberd....

Well, while I dream the same dream, it feels awkward asking my family and friends to pay 5€ per year so that we can share the cost of running our XMPP-Server. Currently, I pay all the bills and manage the server myself (so no cost/effort for them), and they still keep using WhatsApp with most of their contacts.

I don't know what the root of that evil is, but there are undoubtedly multiple factors involved. First of all, most people have WhatsApp already.

Secondly, it is effortless to use. With federated systems, you always have to choose a provider. Once you have overcome that hurdle, the privacy-sensitive people like us do not want to share their address book with the server so finding your people is a manual setup for everyone (another hurdle).

Last but not least, the client landscape of XMPP is still far from perfect. If you want to use end-to-end encryption (e.g., OMEMO) there are finally some clients which work with each other (Android: Conversations, iOS: ChatSecure, Desktop: Gajim), but configuring all that stuff (Server + Clients), is not as easy as pushing a button. Other features like video calls are still very fragmented and rarely work if different clients are involved.

I think it would take ten dedicated developers about a year to fix all those problems (if they would agree on common goals and focus on those) and even after that, we would still have to sell the product.

Yeah, maybe I am a bit too negative about the format itself. My problem with the Flatpak only approach was that it was kinda hard to get it to run at all and it didn't work particularly well, so I ended up with putting some work into it while ending up with an unsolved problem (stable Linux desktop SIP client).

So maybe this is not an inherent problem of this technology and will work better in the future.

Why is it problematic when people hide their identity? I mean, if the criticism is valid, what does it matter who said it?

I think hiding the identity is not the real problem here. To me, it looks more problematic, that the critique is not very constructive, one-sided and loaded with imputations.

This is especially bad for projects like Linphone (open source SIP Client) which exclusively provide Flatpak-builds for Linux [1].

[1]: http://www.linphone.org/technical-corner/linphone/downloads

Well, you are comparing apples and oranges here. The equivalent AWS service would be EC2 and not S3. I know that you didn't start with that (as gprasanth put them in the same race), but it should be clear those two have different redundancy levels.

In fact, I don't know where AWS nor Hetzner stores the 'disk' of the VPS or even the backups. And while those are undoubtedly essential attributes for enterprise-level services, I think especially for side projects the usability of the service is quite relevant.

> 7. Prompt if you can

Please don't. There is nothing wrong with interactive tools, but by default, they should not be. So instead of making non-interactive session possible via flags, the default should be to be non-interactive. If there is an option to start an interactive session, everything is fine.

Otherwise, you would never know when your script could run into some kind of interactive session (and therefore break; possibly after an update).

I still miss Wave... In my opinion, Googles biggest failure was the missing real-world federation. They promised us, that there will be a server to run on your own hardware and yet it took them years to release anything that was usable. Even years after the open source release the software was quite unstable.

Paired with the missing backward compatibility with email those two are the most important aspects of why Wave failed IMHO.

Nextcloud [1] (Federated Groupware) has the concept of circles [2], but yes, that's probably not what you meant with social network ;-)

[1]: https://nextcloud.com

[2]: https://apps.nextcloud.com/apps/circles

> [...] no language barrier because so many people here speak English.

I wonder how you came up with that one. From my experience (I am German myself), Germans are not good at English. It is not so bad within the large cities but in the countryside, good luck. Just take any of the Scandinavian countries and you will find that they have a higher percentage of people speaking English.

Living in Germany without learning German is a hard challenge. However, speaking German with an American accent is sometimes considered cool ;-)

I am sorry, but that is no BS, just pure facts. Granted the PC is not very powerful (CPU: AMD E-350), but the story is true. When the update started, I thought 'Shit that thing didn't just reboot itself. Okay here we are, that will take one hour max.'. But in fact, it took about 4 hours.

They probably just want to push their cloud services, so that you 'Never lose a file due to some broken update!' :-D

If you want to use Windows, just do it. But if you had to experience what it feels like when Windows thinks it is time for an automatic-forced-reboot-update which takes 4 hours the day you have to submit a thesis, then you might know why I am not so fond of Microsoft products anymore...

By the way, later that month I learned that I had been lucky as some people ended up being stuck within the update.

A great tool for choosing harmonic colors:

https://color.adobe.com

There are others out there, but that URL stuck to my mind.

Well, just ask the Toibe index:

https://www.tiobe.com/tiobe-index/

It looks like Elixier ist by far less popular than the other discussed languages.

Recently I have become a fan of the Hetzner Cloud:

https://www.hetzner.com/cloud?country=us

I don't know how competitive their prices are, but I like their easy to use interface which is complemented by and also easy to use API.

Adding a 7-day automatic backup history is just a matter of about two clicks, and the additional costs seem reasonable to me.

I wonder what JS dependencies FTP has...

Yeah, I really like that part:

> A cloud-based brain-to-brain interface server could direct information transmission...

They can barely transmit 1 bit per second with no prospect of increasing the data rate but start talking BS about the cloud.

Well, yes, in fact, I was complaining about the Same-Origin Policy, and CORS is just the consequence of the way the SOP works. Nevertheless, this doesn't really change the situation.

If the browsers separated the session by origin (as blauditore wrote), the whole problem space would look very different.

I never really understood why we have CORS. I mean, the problem with CSRF is that some random page can trick your browser into adding its authentication token to a request which does not originate from the authenticated page. So why the do we need the server to tell the browser that it should not send requests from other origins?

In my opinion, it would have been much better to improve the browsers to not include cookies in 3rd party requests automatically (only when they are explicitly specified via JS for example). It should have solved the issue equally well, without introducing some bulky server-side security feature to remote control browsers.

In general, I agree that juniors tend to skip useful information, but it depends a bit on the technology you are working with. PHP, for example, was known to report the error for the line after the one you forgot the semicolon at or at the line after the last line of the file when a parenthesis was missing somewhere.

While there is a reason why it did so, those are error messages which are quite hard to understand the first few times they occur to you.

Nevertheless, those lousy error message should be exceptions nowadays, and if you are using decent technologies, the chances are much higher, that you just didn't read the error message not careful enough to understand the problem.

While I like your way of thinking, I don't think the argument applies. The game itself doesn't possess any kind of AI and is somewhat static, more like a sudoku book: There are lots of puzzles, but you know it when you have solved one.

The one-on-one tutor idea is that you have a master who sees the mistakes the student makes and gives him an exercise to target precisely the misconception the student might have in his head.

The Witness, on the other hand, doesn't possess such intelligence. Instead, it is a carefully crafted series of puzzles which slowly broaden the possible moves. Most of the time every next puzzle requires you to learn a new part of the rules. Sometimes you assumed that part anyway, and the puzzles are easy. But sometimes you have to find out that misconception in your head and replace it with something correct which makes the puzzle harder.

So one concept includes an intelligent observer while the other is more like a perfected text book.

There are other ways of making sure the student understood the concept. Recently I played the game 'The Witness'. The whole game is about learning new puzzle rules, and yet there is not a single dialog within the game or even text explaining those rules.

I am not saying that their technique is the most efficient (e.g., adding hints would undoubtedly increase the efficiency, but also ruin the game experience), just that there are other methods of making sure a student understands a concept. You don't necessarily need the one-on-one conversations. Those conversations are mostly useful to round up incomplete teaching material (again, I am not saying that creating perfect teaching material is easy).

I think the One-on-One Tutoring things is something the AI current movement could make a difference. Everybody seems to be obsessed with building really cool stuff, but our teaching system is quite obsolete and could get better by adding smart systems.

That said I must to add, that I am referring to teaching humans who are 12 years and older (IMHO young kids require physical interaction if you want to avoid psychological conditions).

There are very few people who want to read many books that are 1000 years old. The few books that interest more people can be converted to newer formats, the few people that want to read all old books will have to use extra tools to do so.

That said, I am not a particular fan of books. They take a lot of physical space, are heavy and age. So as long as we stick to reasonable formats (e.g., text-based, non-binary), it should not be too hard for future generations to use our books.

Using DRM, on the other hand, might make things complicated.

Two years ago I wrote a little Progressive Web App which stores its data in an offline cache (browser) and syncs to my Nextcloud via WebDAV. IMHO that has pretty much the same advantages that Solid proclaims.

I mean, in spirit I seem to value the same things as Solid (decentralized, own your data, etc.), but what I don't understand is why it has to introduce so many abstract/new names when it wants to be 'simple'?

Building on top of the Semantic Web concepts isn't going to help either as it has enough disadvantages of its own (e.g., complex standards without adding any real value).

I just discovered a directory with all those 'goodies':

https://github.com/duckduckgo/zeroclickinfo-goodies/tree/mas...

To use them you can take a look at the file and search for the term 'triggers' which should give you a hint.

Edit: Goodies seem to be just one category. The other categories can be found here: https://duck.co/ia