Day 2 of the Old Computer Challenge, 60 minutes of Internet per day. Yesterday I said it was easy. I changed my mind.
I think my parents switched their Internet subscription from RTC to DSL around 2005, 17 years ago, it was a revolution for us because not only it was multiple time faster (up to 16 kB/s !) but it was unlimited in time! Since then, I only had unlimited Internet (no time, no quota), and it became natural to me to expect to have Internet all the time.
Because of this, it's really hard for me to just think about tracking my Internet time. There are many devices in my home connected to the Internet and I just don't think about it when I use them, I noticed I was checking emails or XMPP on my phone, I turned its Wi-Fi on in the morning and forgot about it then.
There are high chances I used more than my quota yesterday because of my phone, but I also forgot to stop the time accounting script. (It had a bug preventing it to stop correctly for my defense). And then I noticed I was totally out of time yesterday evening, I had to plan a trip for today which involved looking at some addresses and maps, despite I have a local OpenStreetMap database it's rarely enough to prepare a trip when you go somewhere the first time, and that you know you will be short on time to figure things out on the spot.
Ah yes, my car also has an Internet connection with its own LTE access, I can't count it as part as the challenge because it's not really useful (I don't think I used it at all), but it's there.
And it's in my Nintendo Switch too, but it has an airplane mode to disable connectivity.
And Steam (the game library) requires being online when streaming video games locally (to play on the couch)...
So, there are many devices and software silently (not always) relying on the Internet to work that we don't always know exactly why they need it.
While I said I wasn't really restrained with only one hour of Internet, this was yesterday. I didn't have a feeling to work on open source project in the day, but today I would like to help to review packages updates/changes, but I couldn't. Packaging requires a lot of bandwidth and time, it requires searching for errors if they are known or new, it just can't be done offline because it relies on many external packages that has to be downloaded, and with a DSL line it takes a lot of time to keep a system up to date with its development branch.
Of course, with some base materials like the project main repository, it's possible to contribute, but not really at reviewing packages.
I will add my counter a 30 minutes penalty for not tracking my phone Internet usage today. I still have 750 seconds of Internet when writing this blog post (including the penalty).
Yesterday I improved my blog deployment to reduce the time taken by the file synchronization process, from 18s to 4s. I'm using rsync, but I have four remote servers to synchronize: 1 for http, 1 for gemini, 1 for gopher and 1 for a gopher backup. As the output files of my blog are always generated and brand new, rsync was recopying all the files to update the modification time, now I'm using -c for checksum and -I to ignore times, and it's significantly faster and ensure the changes are copied. I insist about the changes being copied, because if you rely on size only, it will work 99% of the time, except when you fix a single letter type that won't change the file size... been there.