Comment by kkfx | Hacker News Reader

Comment by kkfx | Hacker News Reader

Parent Full thread

IMVHO:

- all distributed solutions show too much overhead to perform well for casual users, some (now abandoned) like ZeroNet was performant enough for some personal blog hosting, but barely and still no "indexing" solution stable enough;

- classic decentralized social like Usenet seems to be abandoned by most people (except for commercial piracy) so well, most people do not really care enough to take this root if they do not feel something at hand immediately...

Long story short: we "just" need IPv6 with a global per host AND a "family" domain name (with all relevant subdomains) per home, a homeserver (like common domestic "routers", just a bit more open/powerful in hw terms) and that's do the base to rebuild "the internet" with a connected desktop model where we no not need third parties or a VPS to traverse NAT, where we do not have to deal with strange long IDs instead of something like phone.bill.myfriendfamily.tld to make an IP2IP audio/video call and so on.

All other approaches ESPECIALLY those who try to mimic big platforms will fails.

I admire this vision, and also appreciate the fact that you quoted "just" to underline a recognition of how much of a fundamental shift this is :D

Thanks, and yes it is, but technically we have all the pieces since decades, so while it would be a revolution it's not really something new to be created with uncertain results...

The real revolution is a paradigm shift form the actual development focused on big of IT interests toward a classic desktop centered development, the needed ground is there and battle tested, the rest is "just" to be done on scale. A simple example: we have emails, almost any human on the planet have one, at least between those who use computers in a way or another. Today for 99% mail == webmail, something offered by some third party. We have some desktop MUAs but they are stuck in the '90s at best. However we have for instance OfflineIMAP, notmuch, storage normally big enough, so all we need to have emails on our iron is a MUAs friendly enough for end users, based on notmuch/mu like nomuch-emacs, mu4e, neomutt with notmuch backend and so on, that offer IMAP sync built in so and end user have only to enter relevant credentials and wait for the initial download time. The rest follow naturally.

Similarly you buy an IoT sensor? Your homeserver speak MQTT, CoIoT etc all you need is connecting to the sensor via LAN or to the AP offered by the sensor and then configure your WLAN. You see the sensor pops up in HA or something similar, that's is. No need of a third party "cloud". And so on.

The homeserver is simply a machine on 24/7/365 like most home "router", witch are actually embedded appliance, we only need them more powerful and open to be used by the end user, not only by the ISP. This "center point" offer "sync/services" to all families devices, that's is. The "cloud" became a simple backup/mirror of the homeserver for redundancy, availability, backup.

thats an interesting perspective. I am personally much in favour of reducing unnecessary interdependence in society, specifically to big corps: google can really screw me up just because they provide an email service for me. questions that come to mind though:

- why would someone do that? its not hard, it can be made easier, but what would be the tangible, low-level gain for people to do that? I can think of examples like safe storage of photos and documents (my old dad already does that, he buys hard disks and stores stuff), easy and safe sharing of stuff in the home (airdrop does this tbh), and secure smart home controls. but I can't generalise the pattern to properly understand the underlying benefit here.

- in communications, there is an issue of trust. like how can I trust email coming form a new personal server, to not spam or scam me? I think that can be solved separately though, just can't formulate the solution clearly in my mind atm.

IMVHO because of material and monetary interests: so far Google have not let down too many, but horror stories of people screwed by a Google ban are more and more common, "the cloud" start to be expensive not only for companies but also for individuals and for companies horror story about "economical dDOS" start to be real, Netlify have just paid 104.000 USD for a spike demand for a single mp3 file for instance, once such things became spread enough and that will happen soon, people start to look for alternatives. So far they feel to have no choice, but if the FLOSS world will came to rescue the choice would be there.

About emails, the shot is far longer: we start to have digital IDs, after decades finally the idea of a digital signature usable with a smart-card start to be the norm, at least in EU (eIDAS) but I think in the USA things are not much different, at maximum few years behind. With "family domain names" we start to "know each other" enough to filter safely our contacts and bayesian spam filters, mail aliases wisely used are more then sufficient for the 99% of usages.

Another VERY important thing such model can offer is automation, a thing 99% of users miss simply because there are no solutions designed for them. Now imaging a world with OFX/OpenBank mandatory available to all customers of any bank/financial institution, imaging a set of FLOSS clients to import all their feeds from different banks, cards, ... in a single place and auto-categorization of most transactions. This alone is a thing most users today do not ask for because they do not even know that's perfectly possible today. But with a resurgent desktop model such things would easily became common and at that point the big corp model would fail crushing on it's own technical limits.

To me the only point is "where to start", the classic answer was from universities labs where students start outlining and implementing their own world, due to the sorry state of unis these days that's can't happen, but in the FLOSS world it can. If enough "seasoned enough" devs start to say "dammit, I'm tired of the actual crap, instead of trying mimicking big corps platforms let's made something the classic desktop way" in few years such ideas and simple software would spread and following the same FLOSS spread "digital sustainability" where nowadays most devs works on *nix FLOSS environments even if many inside some proprietary walled gardens such comfy and simple solution would grow enough to be unstoppable. The only point so is having a not-so-small set of devs who want to start such change, the rest is already there...

Reply on news.ycombinator.com