P3: Handle deprecated features gracefully

[maxh]

e.g. we should know the milestone the feature landed and the milestone that it was deprecate.

[foolip]

Is the idea to have this information in https://www.chromestatus.com/data/featurepopularity? If so what's needed to make it happen?

[ebidel]

We need an automatic way to know when a feature has landed or be deprecated in Blink. Is there anything like that for each milestone?

[foolip]

It can be extracted from the UseCounter.h/cpp from the release branch with relative ease. I've done it for UseCounter.h using JavaScript for a personal dashboard thing, but for chromium-dashboard doing it on the server side would make more sense. Or possibly having a python script that does it and checks the result into Git just like for the list of features itself currently. What do you think?

[ebidel]

Yea, that would be great.

On Thu, Nov 20, 2014 at 1:11 AM, Philip Jägenstedt <notifications@github.com

wrote:

It can be extracted from the UseCounter.h/cpp from the release branch with
relative ease. I've done it for UseCounter.h using JavaScript for a
personal dashboard thing, but for chromium-dashboard doing it on the server
side would make more sense. Or possibly having a python script that does it
and checks the result into Git just like for the list of features itself
currently. What do you think?

—
Reply to this email directly or view it on GitHub
#22 (comment)
.

[foolip]

Are simple removed/deprecated flags matching the stable channel for each bucket sufficient, or would a lists of existing/deprecated features for individual milestones be worthwhile? It really depends on how you want to present the information I think.

One idea is to mark the stable channel milestones on the time axis. That would make it rather apparent to a human which milestone a counter was added or removed in, without needing to generate per-milestone feature lists. Would getting that into the timeline view be hard, if the data is there? If you don't have an internal feed for it, one can get the information from http://googlechromereleases.blogspot.se/

[jpmedley]

#1471 is a version of this issue more related to the current interface.