Firefox/Sync/Progress Report/2017Q3

< Firefox‎ | Sync‎ | Progress Report
Revision as of 22:38, 9 August 2017 by Juliemccracken (talk | contribs) (Finished updates)

August 2, 2017

OKR Progress

Objective Key Result Confidence Features Notes
Jul 6 Jul 25 Aug 2 Aug 16 Aug 30 Sep 13 Sep 27
1. To enable bi-directional bookmarks on iOS by Firefox 57.

User story: As an iOS user, I’d like to see my existing mobile bookmarks on my desktop.

1.1 Users can see their mobile bookmarks on desktop. 50% 65% 65%

Bookmarks are Sync'd from iOS to desktop.

  • Went unexpectedly well. Has landed.
1.2 Users can delete their mobile bookmarks from their iOS device.

We don’t want to prevent users from deleting their mobile bookmarks from the mobile device.

50% 50% 50%

Ability to delete mobile bookmarks from iOS device and Sync those deletions to desktop.

  • Also going well.
2. To have our data provide valuable insights about Sync engine errors and bookmark validation. 2.1 Validate our validation data.

(Desktop, iOS and Android)

We don’t have confidence that our validation data query results and if they tell us what we want to know. Our analyses tell us different things depending on how they’re run. Let’s get a third party to review the validity of our queries and results.

50% 50% 75%

Validated validation data

  • Work is progressing and results are looking good.


Risks:

  • Assume we won’t hire an analyst by end of Q3 to support our team.
2.2 Stand up dashboards with accurate bookmark validation data from Android and iOS.

Make sure that the Android and iOS pings are coming in and land the data in Presto in order to create dashboards.

50% 50% 60%

Bookmark validation dashboards for iOS and Android

  • Still blocked on the pre-requisite of landing a bookmark validator for Android, but this has been reviewed and is almost ready.


Risks:

  • Android validation data aren’t in yet.
  • What’s the status of iOS sending pings? Landed, but we need to make sure they’re sent/received.
  • Is our analysis correct? If not, we risk creating dashboards that don’t answer our questions.
2.3 Flatten Sync engine data.

The engine data are currently unusable for creating dashboards. We need to flatten, have confidence in our results so that we can prioritize our fixes for Desktop.

50% 50% 60%

Provide better sync engine error data for dashboards.

  • A lot of work has been completed but with mixed results.
  • Have permission to do it in Scala and look at why it’s crashing the way it is.
2.4 Stand up Sync Engine Dashboards.

Using the flattened Sync engine data, stand up dashboards and have confidence in the analysis.

40% 40% 40%

Bookmark Validation Dashboard

Dependent upon 2.3.
2.5 Formulate concrete action items from the validation and sync engine data.

Create a backlog of prioritized bugs to be addressed as a result of our data discoveries.

60% 50% 55%

Implementation of a Repair Failure Strategy

  • Have been conducting some analysis of the validation and forming some ideas.
3. To prevent bookmark corruption. 3.1 Land the bookmark Sync tracker on Android.

In support of Full Bookmark Management that’s being introduced in 56.

Needs to land in nightly by: July 31st (56 goes to Beta on August 7).

50% 70% 80% Version-based syncing of bookmarks
  • Agreed to move bookmark management to 57. Work itself is largely done. Some remaining changes will be done this week.

This work will make the work for 3.2 easier.

3.2 Prevent re-ordering of bookmarks on Android.

There is a recurring problem where users bookmarks are re-ordered on their mobile device. Fix this.

50% 50% 60%

Prevention of bookmark reordering on Android.

  • Have made progress on 3.1.
3.3 Analyze and prevent erroneous bookmark deduping on Desktop.

There is a problem where users’ bookmarks are incorrectly deduped or reordered on their desktop device. This is a form of data loss.

60% 60% 75%

Feature description(s)

  • Understand how iOS merge works now. Ported it to desktop. Merged trees can apply back to Places now.
3.4 Batch upload API enabled for all users, with more than 99% of uploads fitting into a single atomic batch. 50% 75% 80%

Batch upload API enabled

  • Work is progressing well.