User:Emilio/Offline

These are notes taken at a meeting with Eric Nitschke, Julian Stirling, Felipe Schenone, Antonio Anaya and Emilio Velis around our interest in putting Appropedia offline and general offline content distribution.

How large is Appropedia?

• Main page
• Around 15k pages
• 2000 project documentation pages

Previous Attempts?

https://github.com/openzim/zim-requests/issues/701

• Takes lot to receive answers
• We haven’t tried creating a ZIM file

iFixit Example:

https://github.com/openzim/ifixit

WebRecorder

https://webrecorder.net/browsertrix/

https://github.com/iipc/awesome-web-archiving

https://github.com/iipc/awesome-web-archiving

https://github.com/datatogether/research/tree/master/web_archiving

https://docs.google.com/spreadsheets/d/1FqxwaZnIhhQ7jDCC-W64NMRf5rDeh2Shx3u01MsBmTQ/edit?gid=0#gid=0

https://github.com/edgi-govdata-archiving/awesome-website-change-monitoring

https://github.com/webrecorder

https://github.com/webrecorder/browsertrix

https://github.com/iipc/openwayback/wiki

https://www.httrack.com/

https://github.com/skallwar/suckit

https://www.gnu.org/software/wget/

http://pages.interlog.com/~tcharron/wgetwin.html

https://zimit.kiwix.org/#/

https://wiki.openzim.org/wiki/OpenZIM

https://github.com/openzim/ifixit

• Splitting into different sections (solar, water, etc).
• We do have videos on Appropedia but most are embedded on YouTube
• Raspberry Pis where a bit slow but now we have GigaBit connection and m.2 connectors
  • https://learningequality.org/
• The actual documentation for is a static page to bring offline
• Analytics
  • Internet in a Box are using Matomo analytics
  • AWstats
  • https://prometheus.io another metrics tool.
• What is Gitbuilding doing?
• How do we bridge the gap from pageviews to actual devices built and things done in the real world?
  • Antonio: For offline collection of metrics
    • once this content is deployed, you work with implementers and streamline information back to Appropedia.
    • Ensure that Appropedia grows in a different way with feedback.
    • Ensure that people contribute but are not being forced to use a telemetry system.
    • Measuring in person during events and other in-person gatherings (fablabs, republica). You need to do these things in person.
    • Appropedia can have contributions from local deployments, synching data to learn about the impact.
  • Integration with GitBuilding is the best way to go
    • Import/export option with Appropedia
    • Wikimedia could support with Wikifunding
    • Think about synching the changes back to Appropedia
    • Wiki fund/ei? (Eric?)
    • Creating a localized version of Apropedia content
• Tools?
  • Pandoc?
  • Jeff from OSHWA did this
  • Manyfold.app
  • https://weasyprint.org/
  • https://pagedjs.org/
• Final thoughts
  • Consider offline first / offline open source?
  • Get a docker contaniner and create a zim file?
  • Follow up to get gitbuilding into Appropedia
  • Discuss Fab25 / re:publica in Berlin

Julian has got zimit running in docker. It does create zims of websites that work in kiwix.

If we can create zims of Appropedia this means we don’t need to wait for the offical kiwix libraries team. We could create zims.appropedia.org.

Wakoma could perhaps zim up a load of websites that the kiwix folks don’t want to archive.

No success yet with zimming appropedia using a simple:

docker run -v $(pwd):/output ghcr.io/openzim/zimit zimit --url https://www.appropedia.org --name appropedia

this probably would have worked eventually but was trying to zim all the past history.

There is also: https://github.com/openzim/mwoffliner which is mediawiki specific. I am not sure how cusomised Appropedia is and if this will work. It doesn’t seem to have a docker image linked on the GitHub (there is a dockerfile to build one). I tried to install mwoffliner in a node docker container, but there are issues.

Options for zimming:

• Use Regex excludes to exclude a number of the special pages and past history in appropedia
• Try to build a docker image for mwoffliner and see if it runs

It would be good to have a focussed technical call about a pathway to think about a minimal implementation. This could be as simple as people build their docs and serve them on their own page (or on GitHub/GitLab pages), and then we have an easy way to link from appropedia. This however is not compatible with offline Appropedia. We could also think about something a little more complex where the static pages gitbuilding creates are hosted within appropedia in some way.