His How to Fold a Julia Fractal (http://acko.net/blog/how-to-fold-a-julia-fractal/) is an example of someone who understands how to teach a topic, teaching a topic. And to his point, the visual and interactive pieces definitely help with regard to learning. I believe that textbooks, based on html/epub, with this type of interactions would help with education.
I'm pretty sure you can download a Prezi presentation. It's ugly (upwards to 200mb for something simple), but it can be done. Or at least could, I really don't like it, but a friend of mine uses it and he always downloads the presentations.
it feels a little silly to say this out loud, but I'm convinced that what Steven Wittens is doing on the web is some of the most advanced and incredible stuff I've seen since the original chrome experiments came out. Every time I read something on acko.net I am flush with both envy and wonder at how someone can command a browser with that kind of deftness.
Not a quick read, but very interesting and with impressive live demos.
EDIT: I really wanted to give a star to the source code, but it's not on github - it's on gitgud [1]. Honestly, for your private projects, use whatever you want, but if you have an open source one and you want discoverability and stars, use github.
Git is decentralized and it works perfectly well for what it needs to do. But git isn't a system to share open source projects and get social validations from other people. Github is though, and it's excellent at that - and the network effect is there too.
Git is an area where I'm surprised a DHT system for open source hasn't yet materialized. I'm writing dockerfiles lately where I realize what I really want to do is just go 'connect to the global DHT and clobe the tree with hash (sha1 here)'
The security model for that might be, um, interesting.
You don't want to clone a particular revision ID, you want to clone a particular branch. If it "lives" in one known place, then its home can provide access control. But if it doesn't have a home like that, then everyone needs to have consensus on what's included and what's not. And you probably don't want to have to publicize a new repo/branch/whatever ID every time your committer list changes.
> You don't want to clone a particular revision ID, you want to clone a particular branch
Actually, that's exactly what I personally want for a bunch of things. I want to grab a very specific version of an item and know that I've got the one I expected.
Yes I can. I can register, give a star, and it won't make a difference. You can like this or not, but it's how it works when network effect is at work.
I highly dislike centralisation, especially if it is centralisation on proprietary services. The more the merrier, people should feel comfortable using different sites and not chastise the underdogs.
But in the reality of things, having a central repository with a voting system that is significant and trustable enough is very important - for example, when you need to choose between apparently similar libraries. And a star on Github - rather, a good number of stars - does count much more than any number of stars on any other service in the programming community.
I say this not because I have a particular love for Github - I also use Gitlab for some projects, and it's very good - but just because it's the way it is.
Yes, they may count more on the most popular site but it doesn't have to matter. When enough people see it on Hacker news or Reddit it will be known anyway and it might get start on the site it is on. You can just judge it from a different scale.
And if the most popular should always win there would be no Linux, Windows is still much more popular.
I'm not sure your dislike of hosting the code on gitgud is bona fide a desire to have the project on the most popular platform; rather it might be a political move to discredit an alternative because it uses gamergate-inspired artwork on its front page.
