Hello Again, World!

Here I am, back at blogging again!

The story so far

Back in 2001, I had a small blog that was using FogCreek CityDesk. I was using the blog to share random thoughts (as one was wont to do), including programming ideas. I was in my mid twenties back then and eagerly looked for opportunities to make a dent in the world :-).

I wrote my last blog there on 2002/04/22, leaving behind plenty of unfinished work like this Back to Basics series of articles about XML. I was reeeaaally into XML back then!

I restarted blogging on 2004/10/26. This time I used DotClear. I had a pretty good run on this platform. Between 2010/01/04 and 2010/03/09 I switched over to WordPress.

This article from 2009/04/16 shows how much I was into Google Reader back then, reposting stuff to Twitter in a bit of a dismissive manner.

I really had a love/hate relationship with Twitter. I joined the service with a fanfare on 2007/05/04:

I wrote a case against Twitter on 2009/08/12. But on 2011/12/27 I eventually recognized that Twitter was a good enough platform for me :-).

Finally, some time between 2013/08/17 and 2013/09/17, due to the loss of Google Reader, I threw the towel and redirected all the traffic from my blog to Twitter.

A new start

So why start again now? Well, Twitter again… While Twitter is now a great place to discover new stuff on the web, provided you take care of following the right persons, it is still an awful place to write and have discussions. The fundamentals are all wrong.

So here we go again. This time I’m starting on a new platform: Hugo.

After self-hosting my previous blogs, I was fed up with the admin chores and the security risks. I briefly considered using Medium, which looks good in theory, but has a pretty annoying monetization scheme, and actively participates to the current website obesity crisis.

Hugo is a static web site generator. Once the site is published, it comes with zero runtime dependencies1: no backend code, no database, you just need a HTTP server that can serve static files. The sources for the site are written in a dialect of Markdown.

This seems to be like the most basic setup that can stand the test of time. I did keep some backups and git repositories of previous version of this blog - but in the end this relies on a lot of legacy software (like antiquated version of MySQL) and the only pragmatic way of finding my old writings was to use the Wayback Machine2.

My hope is that software archaeology will be simpler using Hugo - everything you need to read the original article is a dump of the content folder. The weak point in Hugo is the Markdown dialect - there are a few extensions like front matter and short codes that are custom and will require Hugo or some reverse engineering to make sense of in 20+ years3. But the formats are pretty simple, and who knows, since Hugo is OSS written in Go, it might still be able to run it then?

I will actually try and restore old articles here. Not that they are so great, but it’s part of my digital identity that I want to preserve.

So with that all being said, let’s see how things go in the future!

  1. I have chosen to introduce two external dependencies: Google Analytics to get an idea of how many people are actually reading this (not much, mind you), and Disqus to get comment threads.

  2. I just donated to the Internet Archive, and you should do the same!

  3. Funny how 20 years in the future seem so far, while I’m looking at articles I wrote in 2001 (17 years ago) and it seems much closer to me. Life is short, I guess.