Tag: computing

Categories
Uncategorized

“It feels fragile because of the Internet”

This blog post about using a 14-year-old PowerPC Mac Mini running 12-year-old OS X Leopard has this musing towards the end:

Something about using this feels very fragile, and not because of the machine itself, the operating system, or even the interface. It feels fragile because of the internet.

The internet has so aggressively taken over our lives that we can’t imagine a computing experience without it. And when it’s no longer there on a platform that didn’t really work properly without it, it becomes impossible to use in many ways. One has the feeling that even older operating systems won’t feel this broken in retrospect, because their experiences are otherwise separate from the internet and work without it being at the center of the experience.

I remember this period of “even older” operating systems in the 90s, of coaxing my Pentium desktop PC to run Red Hat Linux 5 and work on it for a whole day – tinkering, programming, writing, simple gaming – without connecting to the internet, or for that matter without any part of the OS being internet-first. I’d “go online” by dialling the modem via a shell script, look up what I wanted to all at once, then disconnect. The internet was like a trip to the library instead of being the environment itself. No chat client ran perennially, no mail client polled for new email. No iCloud Drive synchronised silently with the Cloud (the term didn’t exist then) in the background.

It wasn’t necessarily a better or worse time, just that it definitely was different.

Categories
Uncategorized

Raspberry Pis as low-footprint single-purpose machines

I’m excited about the new, more powerful Raspberry Pi. This Hacker News thread describes several interesting hardware projects that the people have used the Pi for. And the Raspberry Pi website positions the new, more powerful Pi 4 as a Linux desktop machine.

I’m interested in its use as a low-power, low-footprint single-purpose machine, and having a few of them across the house.

For instance I have a privacy-focused Pi that’s physically tied to my router and serves as a network-wide ad- and tracker- blocker via pi-hole. It also runs cloudflared to encrypt my DNS lookups over HTTPS. I’m also setting it up as an OpenVPN server.

Another one is about to enter service as a XBMC/Kodi machine for movies and TV shows. This lives behind the TV. I’m also considering a third to serve as a Time Machine via netatalk and general-purpose backup machine via Syncthing. This will probably be behind my desk.

Taken together they’ll draw less power and run quieter than a desktop machine that does everything, and will be near-invisible too. It’s a new class of personal computer.

Categories
Uncategorized

Owning your data, offline

Google’s services including Gmail, Drive etc went down for over 4 hours. I did not even know about this until now, nearly 24 hours after the incident; most of my documents are available offline, including email. Which made me articulate what I have long realised:

This incident should be a wakeup call for owning your data and certainly having offline access to it. If you can’t even ‘reach’ your data because it’s in the Cloud, having perfect sync between all your devices does not mean much. Even if it’s a rare event – what if you need your own information right then?

The other risk of having all your data on the Cloud without a pure-offline copy (such that files are first-class citizen – offline google docs files that need the docs.google.com Chrome client or app to open, do not count) is that you could be locked out of your own account. The journalist Mat Honan had his accounts broken into and held hostage/wiped out in 2012. Even if you’re not digitally attacked, it’s easy to be locked out.

Maybe it’s worth giving up cross-device syncing and moving to a single-device setup like Jason Fried of Basecamp all the way back in 2010. He does not have a work computer and a home computer:

> One powerful, portable, fast, machine with a high-rez screen and a clean desktop. I don’t really believe in dreams when it comes to hardware. These are the tools you use to do your job – you should have the best you can afford.

If you must have multiple machines, sync them peer-to-peer using something like Resilio Sync, or its open source alternative SyncThing, both based on BitTorrent. Back up using Time Machine (if you’ve got Macs) or if you are even slightly technically inclined, via unidirectional rsync to an external hard drive connected to your machine.

Above all, keep your data in open formats. Using a single backed up machine is no use if all your files live in, say, Evernote or OneNote. Can your files be read natively by open-source equivalents? Can they be _conveniently_ exported in bulk? There is no shortage of open formats for multiple types of data: CalDAV, CardDAV, Markdown, ogg, mobi, PDF, mbox – even Microsoft Office is a mostly known format. Do you store your data in a simple folder-tree structure instead of in a proprietary library – your photos may all be PNG but they may be in Apple Photos’ binary library format.

You’ll give up the glamour and some ease of cloud-based, real-time collaborative, unstructured write-anywhere apps, but what you’ll gain is a lot more valuable – the ability to have anytime anywhere access to your own data, for years on end, worry-free. No one will be able to force you to continue using their software, pay a subscription, lock you out from your own files, leak them in a security breach, or ‘go down’ in an outage.

When it comes to your files, your memories, your life, sticking to the basics is a good idea.