Jeremy Morgan

The key problem here is "lots of programmers". Even on a big project, there should only ever be "enough programmers", just enough to get the job done under budget (hopefully) and on time. It's easy to put the burden on the programmer to carry on soldier! But a lot of the burden of productivity is often on the managers themselves. Mangement needs to manage, not cause strife and yet that's what some do. Erik Lukas seems like a guy who's had his act together from very early on and whether he read up or not, he knew how to manage his people. That's a key skill. There's also been a complete cultural overhaul in the community. Programemrs have been spoiled, for better for worse, and maybe we're expecting too much. Jeremy Morgan wrote a post last year that really illustrates this. Well worth a read : http://www.jeremymorgan.com/blog/programming/the-programmers-before-us…

Might be this one? http://www.jeremymorgan.com/

If you’re looking to migrate away from Twitter to Mastodon, check out this article: https://www.jeremymorgan.com/tutorials/mastodon/migrate-twitter-mastod…

This isn't even a link to an article. It's a link to a category on Lifehacker, which then links through to an article, which then links through to another article at: http://www.jeremymorgan.com/tutorials/raspberry-pi/how-to-raspberry-pi…

Looks like you are talking about a "distributed data system", where the OP (and me) were heading to the idea of a "personal website with social functions". Those are very nearly the same thing. Most websites and online services (and I've engineered several of them) are basically: 1. An application engine, which is a fancy term for "database wrapper script". Much of this can be dispensed with if you build a largely static site, though access control and some scripted / scheduled events are likely still necessary. This talks to ... 2. A data tier. Often multiple levels of caching and / or distributed data and or map reduce and / or nosql and / or database. Again, largely dispensible in a static site. 3. A lot of front-end caching and load balancing. The less content you actually have to serve, the better. It's better to think of a personal cloud + feed syndication (pretty much my formulation as well, BTW) as a cache-tier seed. That is, whatever kit you're running pumps data out to a distributed caching framework (think DNS or bittorrent). If a request cannot be met from cache, or it's expired, then your origin gets a request. But if it can be, you're serving content without actually taking the hit. And if the caching is demand-responsive, then you actually have the situation where you get better at serving data the more requests there are. A DDoS would simply result in more caching peers feeding your data. You also get the benefit that should your residential (or colo)…