What to do once you admit that decentralization never seems to work (2018)
1–10 of 148 posts
Re: What to do once you admit that decentralization never seems to work (2018)
#2> Berners-Lee’s World Wide Web Consortium (W3C), for instance, has been a critical governance body for the Web’s technical standards, enabling similar user experience across servers and browsers
I'm all for real standards, and believe W3C has acted in good faith, but in what shape is W3C's standardization process now (hint: they cancelled the HTML 5.3 and SVG 2 spec) and where has it lead us?
Re: What to do once you admit that decentralization never seems to work (2018)
#3> Berners-Lee’s World Wide Web Consortium (W3C), for instance, has been a critical governance body for the Web’s technical standards, enabling similar user experience across servers and browsers I'm all for real standards, and believe W3C has acted in good faith, but in what shape is W3C's standardization process now (hint: they cancelled the HTML 5.3 and SVG 2 spec) and where has it lead us?
As its name hints the Consortium is intended as a way for organisations to work together. This seemed like a natural fit for the Web which had all these exciting startups in the Valley and around the world. If you look at something like the IETF where notionally all participation is individual and organisations are acknowledged only as the employer of an individual contributor you see plenty of stuff getting done that is blatantly, even explicitly, done on behalf of that employer. So why not cut out the middle man?
But it turns out that's probably the wrong shape. While some stuff at IETF would get done by "any Google employee" or "any Microsoft employee" or "any Cisco employee" plenty of stuff that actually gets done is because one or more specific people cared about it, regardless of their employer or occasionally even if they don't have one.
For years it was literally impossible to work on W3C standards as an individual. They eventually fixed that, but the message sent echoes forever, "Go away, we don't care".
I like Tim (he is a nerd, he used to derail our apparently serious board-level meetings with his IT problems), but my impression is that one of the things that lead to the W3C was that Tim could see from pretty early on (certainly by the time W3C is founded) that there's a lot of cash sloshing about for his Web, and of course it's the nature of the beast that Tim doesn't automatically get a slice. That's part of the reason other systems didn't get traction - nobody wants to pay a toll. The W3C is a way for Tim to get some of that money for his pet projects.
Re: What to do once you admit that decentralization never seems to work (2018)
#4> Berners-Lee’s World Wide Web Consortium (W3C), for instance, has been a critical governance body for the Web’s technical standards, enabling similar user experience across servers and browsers I'm all for real standards, and believe W3C has acted in good faith, but in what shape is W3C's standardization process now (hint: they cancelled the HTML 5.3 and SVG 2 spec) and where has it lead us?
Where we are now is not necessarily where the standardisation process has led us. There's a very powerful tech oligopoly that has taken over some time ago.
I'm not sure there's anything the W3C alone could have done to prevent that, even if they hadn't made any of the mistakes they have made.
Re: What to do once you admit that decentralization never seems to work (2018)
#5Decentralisation will never happen on a scale that will affect significant web traffic.
However, there is another model for web apps, not purely decentralised, but a practical option where traffic is not funnelled through a small number of big players. Let's call it the WordPress model.
Consider the following: WordPress is an example of a profitable open source app that can easily be installed on countless shared hosting platforms or on a VPS. It's easy to switch hosting providers when you want to (rather than be locked into one provider). You can take your data with you when you switch. The popularity of WordPress means that one-click installs are widespread. Many providers also offer a 'managed' service where they take care of back-up and updates. But crucially you're still not tied to a single provider.
Now imagine if all web apps were as easy to install on the server via a one-click installation like WordPress. Unfortunately, there is no common standard or API for software installation on the server side, and this lack of an easy installation process for everyone severely limits self-hosting websites and apps.
Many developers think deploying a server-side web app is a non-issue, or they erroneously think that installing Cloudron/providing Docker instances/typing command line instructions are all "easy". Have you seen the server deployment instructions for "web friendly" languages like Ruby and Python? They're ludicrously complicated. Developers are completely blind to the complexity and see nothing wrong in such installation procedures.
The result is ever-growing centralised software through the SaaS model which beats the self-hosting alternative in simplicity every single time. Imagine if your local desktop apps were completely controlled, tracked and even terminated remotely by one provider - don't like the though of that? Yet, that is exactly what the centralised Saas model enables and nobody blinks an eye.
I wish there was some momentum or traction in making server-side web app installation as universally simple as a one-click WordPress install for all for all web apps. It would unlock countless opportunities for developers to reach more users or customers. I'm not one for conspiracies, but I wonder sometimes if developers actually prefer the complexity of server-side installation because it makes selling a SaaS solution much more attractive over the ludicrously complicated self-hosting option.
Re: What to do once you admit that decentralization never seems to work (2018)
#6> decentralization never seems to work
Not true, and there is many technologies, both new and established, which are fundamentally decentralized. Bittorrent is the first example that springs to mind. Cryptocurrencies are another example, etc.
Re: What to do once you admit that decentralization never seems to work (2018)
#7Decentralisation will never happen on a scale that will affect significant web traffic. However, there is another model for web apps, not purely decentralised, but a practical option where traffic is not funnelled through a small number of big players. Let's call it the WordPress model. Consider the following: WordPress is an example of a profitable open source app that can easily be installed on countless shared hos…
Web 2.0 has been in use by a majority of people for maybe ten years now and the internet shows no signs of going away.
To say that decentralization will never happen is nuts.
Re: What to do once you admit that decentralization never seems to work (2018)
#8Decentralisation will never happen on a scale that will affect significant web traffic. However, there is another model for web apps, not purely decentralised, but a practical option where traffic is not funnelled through a small number of big players. Let's call it the WordPress model. Consider the following: WordPress is an example of a profitable open source app that can easily be installed on countless shared hos…
I find the steps I have to go through to install stuff that I expect thousands of other people must be installing at the same time on a cloud service ridiculously complicated. The same with most tasks.
I do however find them doable and maybe also a kind of job security for that reason. But I don't prefer them. I want to build stuff, not do a bunch of prep work every time I want to build stuff.
Re: What to do once you admit that decentralization never seems to work (2018)
#9Decentralization is good when there is no need for coordination. We sometimes like to collaborate - but we don't like the constant requirement of mandatory coordination. Unfortunately technology connects us more and more and that means that coordination and un-coordination brings more and more externalities (positive and negative respectively).
Decentralization works in cases that are like we were back to the near empty savannah and where everybody could just go in his own direction. The Internet was once like that - but it is not any more, it is rather the reverse.
By the way, just a few hours ago I submitted to HN this article: https://www.interfluidity.com/v2/7230.html - it is about these tensions between our freedom to do whatever we like and the need for coordination in a more and more connected world.
Re: What to do once you admit that decentralization never seems to work (2018)
#10Decentralisation will never happen on a scale that will affect significant web traffic. However, there is another model for web apps, not purely decentralised, but a practical option where traffic is not funnelled through a small number of big players. Let's call it the WordPress model. Consider the following: WordPress is an example of a profitable open source app that can easily be installed on countless shared hos…
I find the steps I have to go through to install stuff that I expect thousands of other people must be installing at the same time on a cloud service ridiculously complicated. The same with most tasks. I do however find them doable and maybe also a kind of job security for that reason. But I don't prefer them. I want to build stuff, not do a bunch of prep work every time I want to build stuff.
Decentralisation isn't a collection of technology. Rather, it's a set of principles or tenets. It's above all the by-product of the choices one makes.
Sure, you can install a gazillion packages, use serverless, static site generators, SaaS platforms, install CMS x, y or z.
But then again, you can still use a simple FTP program, copy HTML files over TCP/IP to a remote server with a process running in the background that listens for incoming HTTP requests. Your browser will still be able to from the HTML it receives via a HTTP response. In fact, you don't need a browser. Curl or Telnet will yield a string of text characters any human can read.
Those basic protocols still exist after 30 years and underpin the Web. Nothing changed fundamentally, when you dig deep.
As such, you could lease a shared hosting somewhere and put up your own set of HTML pages and slap a DNS domain to the IP of the server. Boom. Congratulations. You own a website now.
Decentralisation - at it's core - is about ownership. If you put content and data on a host, who owns that? You or the owner of the host? This is where the entire discussion becomes a legal debate.
The big shift of the past decade is that social media offer easy access to centralised hosts they own. This has caused a massive influx of users to publish stuff on hosts that are owned by others.
Meanwhile, there's this perception that you need a gazillion tools to build a website. That you need deep knowledge of languages and such to publish HTML formatted content on a host you own.
This couldn't be further from the truth. Unless you are a business, or an organisation and you need to manage tons HTML formatted content, you can do away with a text editor, basic HTML and shared hosting you pay a tuppence for.
The only reason why things got complex because of business reasons.