A proposal for dropping Microsummaries
Microsummaries were suggested as a new Web feature many years ago, and have been supported since Firefox 2.0. In case you were wondering, they are an auto-updating short summary of a page stored as a bookmark (the page itself provides some basic additional information to power this).
Unfortunately, after the initial launch, the excitement around the proposal and the implementation started to fade. Others suggested alternative proposals with different behaviors, like Web Slices. Microsummaries still have their place, being conceptually different, but as a Firefox-only feature and due to the low advertising received after launch, they are rarely supported on modern pages and apps.
I can sum up reasons we should drop them from the main product:
- Poor support on the Web
- High code maintenance cost (70KB component is pretty large, especially compared to its usage)
- Mostly unmaintained (no resources to look at the almost 60 bugs already filed and at improvements)
- Has surprising user privacy implications (just looking at a bookmark causes a background visit)
- Largely undiscoverable
- No active owner and proposal to revamp the feature
I think the main reasoning is that this feature cannot grow without someone actively pursuing it, unfortunately I feel that right now we miss this person. We'd want to see Microsummaries growing, and going back to be a first-class citizen of the Web, but till then, having unmaintained code is a cost larger than the benefits.
Thus, here is the plan to bring on this stuff:
- Drop Microsummaries on mozilla-central now, so it ends up in the next Aurora merge (targeting Firefox 6)
- Build a basic Microsummaries extension using the current service code
- Put the extension code on a public repository to grow a community around it
I can't give a good timeline for the extension, ideally we hope someone will make a step toward us and offer to inherit this code in his extension or to make himself the extension, that would largely speed up this step and be greatly appreciated. Otherwise we'll try to put up something usable, but can't ensure it will be ready before the feature removal, nor we'll have enough resource to maintain it competitive.
I'm trying to collect feedback AGAINST doing this in order to figure out how disruptive this change could be, so feel free to object and bring your points. Some feedback has already been collected in the bug tracking this removal and I'll shortly create a thread in mozilla.dev.apps.firefox to bring on the discussion.
(Thanks to Dietrich and Sdwilsh for invaluable feedback on this article.)
