Internet for the offline or marginally online

A.K.A. “the internet as if the entire world did not live in Seoul and Silicon Valley”.

If you are experience generally bad connectivity because you live in an area with bad wireless internet, or you occasionally go inside buildings or experience topography in your landscape, you need to think about this. OSX is particularly horrible if the internet is intermittent.

Also you should have the internet cached for offline use even if the net is nice right now, because nation states are playing at war games to destroy the internet, and us little people will suffer when that happens and we can't get our Youtube instructional videos on how to survive the apocalypse after it happens.

Gettings bits of the online internet offline

This is a tricky problem in general, and specific, partial solutions abound. Especially unidirectional solutions. For example, you can get read-only versions of wikipedia for offline use in your remote mountain village; but there is no easy way to contribute your updates back to the version on the main internet.

Offline automatic filesync

Some of the file sync options I mentioned also work as sneakernets, e.g. git-annex, syncthing, and retroshare, which also does a lot more besides.

Technical stuff which handles intermittent connectivity

The command line might need some upgrading

{language=bash}

function wget_non_california () { until wget -c --timeout=10 \$*; do sleep 10; done }


Backed by a major player: mars, the network stack of Wechat.

Offline wikipedia

AFAICT there is no way to contribute upstream. But a reasonably simple and well-curated option is to use the Kiwix offline wikipedia, which can give you everything, everything minus pictures, or only “medical” articles, or only “school” articles and so on.

Offline manuals

devdocs.io is an excellent offline cache of API docs that works from your browser.

Even more comprehensive are the specialised apps dash (OSX, USD25) and zeal (Linux, Windows, open source).

The whole internet offline

Amber Case argues for IPFS, which at a glance sounds great for reading content although I don't understand how it generalises to writing content.

Practical solutions are about bringing alternative bits of the internet into decentralised availability. Surprisingly, this is popular for social networks.