Beaker 0.5 Released

[Paul Frazee]

Release notes: https://github.com/beakerbrowser/beaker/releases/tag/0.5.0

Big thanks to Poga Po for contributing the History view on Dat archives.

About bkr

I'm happy to have the bkr command-line tool added in this release. (https://www.npmjs.com/package/bkr). Previously I had been using the dat cli, which this improves on in two ways:

 1. Key storage.  By using the bkr cli, your sites' private keys are always kept in beaker's storage. That way, they're unlikely to get lost, or accidentally committed to a git repo.

 2. Publishing flow. The bkr cli lets you choose to explicitly publish, instead of instantly publishing any changed files.

The bkr tool should feel familiar to anybody who has used npm and git. The most important commands are:

bkr init - create a new site

bkr co - checkout a copy of an existing site

bkr fork - duplicate, and then checkout, a site

bkr publish - publish a new version of a site

Changes to the Dat API

This shouldn't affect too many people yet (if any) but I removed the dat.serve() and dat.unserve() commands, in a collection of changes I made to simplify site-management. It's now up to users to manually host/unhost sites using Beaker's builtin UIs.

Previously, I had created a claims system for sites to register their desire to have a site hosted or not. If the number of claims was > 0, then the site would be hosted. It became pretty quickly obvious that this was going to be too complex and unpredictable to really work. So, I got rid of the claims, and decided to remove the API entirely for now. (The same system was being applied to keeping a dat saved. Those claims were also removed.)

There's also some new metadata being tracked: author, version, createdBy, and forkOf. Those last two will be handy for tracking the lineage of a dat. The createdBy is especially nice in your sites listing, because it helps contextualize a dat. "Which one is this? Oh yeah, it was created by Twitter."

Finally, I spent a little effort on the dat networking reliability, and managed to uncover some issues that have now been solved. There's a standing issue that shouldn't be too dire, and I think this will be an ongoing project, but things are definitely better now.

What's next

Plans are to keep working on application tools, especially the Web APIs. Discussions lately have been focusing a lot on the Fork button, and what we can accomplish around the concept of "Social Hacking." More on that in the near future.

I should be standing up a Public Peer service soon (currently codenamed Fuego). Look for that by February or sooner.

prf