Email or username:

Password:

Forgot your password?
All posts Tobias's posts Post Back to profile
Tobias Bernard

Part 3 of my series on post-collapse computing is out: blogs.gnome.org/tbernard/2022/

This time: Concrete directions and ideas for how we could make our software more resilient. Some hightlights 🧵

24 Oct 2022 at 12:20 | Open on mastodon.social
11 comments
Tobias Bernard

Software should be local-first, i.e. the network is fully optional. Everything works without any connection, but when you have a connection collaboration and other fancy features are possible.

inkandswitch.com/local-first

24 Oct 2022 at 12:21 | Open on mastodon.social
Tobias Bernard

When there's no network, make it easy to use USB storage as a fallback.

24 Oct 2022 at 12:23 | Open on mastodon.social
Tobias Bernard

The file system has some very good properties for resilience (flexible, future-proof, interoperabile across platforms, familiar), so it's a better primary soure of truth for user data than custom content apps.

24 Oct 2022 at 12:24 | Open on mastodon.social
Tobias Bernard

Textbundle-style stuff is cool (extend simple text formats in small ways that allow them to replace much more complex apps/formats), and we should do more of it.

textbundle.org

24 Oct 2022 at 12:26 | Open on mastodon.social
Tobias Bernard

Don't use the latest, most powerful computers for development, to make sure you keep an eye on performance and support for older hardware.

24 Oct 2022 at 12:26 | Open on mastodon.social
Tobias Bernard

Development should be local-first as well, including dependency caching and offline documentation, to enable local repair work on apps and system.

24 Oct 2022 at 12:27 | Open on mastodon.social
Haelwenn /элвэн/ :triskell:
@tbernard Reminds me of how some smartphones, which are by far the worst in terms of data transfer, can't use USB keys (OTG keys or with a Y-cable) and it's probably only a software defect…
24 Oct 2022 at 20:41 | Open on queer.hacktivis.me
Dragan Espenschied

@tbernard Thank you for this interesting series.

It is weird, to me the conclusions on how software should be designed in face of climate crisis, environmental crisis, pandemic, and war, sound like ways to make generally reasonable software that doesn't suck.

I can imagine that your position is not very established, but it presents an interesting "softening" of more radical perma computing or collapse computing ideas.

Wanting to limit excess in all areas, including computing, is great. Yet in a way, a typical 738 MB Electron to do list app is as much overspecced as a project creating an operating system that can run on a Furby dug out from the rubble in a postapocalyptic world is "negatively overspecced."

The stylish cyberpunk doomsday scenarios are kind of easy to co-opt and fraudulently addressed by big tech products, while an established practice of making reasonable software is much harder to compromise. Continuity is the biggest enemy of disruption :) So I much welcome this investigation of photo apps and syncing data over unstable connections.

Related, here's my favorite critique of the "Arctic Code Vault" as a negatively overspecced project to save computing after the collapse: blog.dshr.org/2019/11/seeds-or

@tbernard Thank you for this interesting series.

It is weird, to me the conclusions on how software should be designed in face of climate crisis, environmental crisis, pandemic, and war, sound like ways to make generally reasonable software that doesn't suck.

I can imagine that your position is not very established, but it presents an interesting "softening" of more radical perma computing or collapse computing ideas.

24 Oct 2022 at 20:02 | Open on post.lurk.org
グレェ「grey」

@despens @tbernard z0mg, GitHub will *not* be all that useful in a disaster IMHO.

I've worked within the field of so-called "software escrow" and what many refer to as "software preservation".

It is non-trivial to understate it.

e.g. recreating prior art to help attorneys for Fortune 25 invalidate spurious patents, while my employer invoiced them $300/hour for my time and paid me $15-$30 if I were lucky. All while contending with brittle old mostly broken code and hardware.

Grueling work.

24 Oct 2022 at 23:02 | Open on mastodon.social
Tobias Bernard

@despens Yeah, I think there's a tendency to think of collapse as a thing in the far future that will affect other people, but the timeline for climate collapse is so much shorter than that.

I don't think bootstrapping a computing stack on microcontrollers is going to ever have much practical relevance, but keeping your current laptop going 20 years down the line very much will.

25 Oct 2022 at 12:54 | Open on mastodon.social
OrX_Qx pirateradio
@Tobias Bernard @Dragan Espenschied

Thanks for great article! These principles are not only for doomsday, I see it reasonably practical in many off-grid scenarios, which are here now. For example field research in distant areas, in nature labs where there is no other option than local energy and no stable connection, or you want to deal with research carbon footprint, so you design livinglab infrastructure including data processing not dependent on any external "cloud" to reach carbon balance.

The field research is a key to ecosystem sustainability and more than human naturecultures research which is in fact fun and breathtaking, quite an opposite to dystopia. With some well balanced computing thrown into forest/garden it just explodes with food and biodiversity. Sorry for disappointment, maybe it will still come when the extractive urban drones steal the harvest.
@Tobias Bernard @Dragan Espenschied

Thanks for great article! These principles are not only for doomsday, I see it reasonably practical in many off-grid scenarios, which are here now. For example field research in distant areas, in nature labs where there is no other option than local energy and no stable connection, or you want to deal with research carbon footprint, so you design livinglab infrastructure...
25 Oct 2022 at 10:27 | Open on webs.node9.org
Go Up