Consider opening date of issues, upwards trend

[doniks]

It seems the chart takes the total number of issues as the starting point in the time line. However, that's not accurate when issues are created over time rather than all being defined at the start.

It would be really helpful to see the upwards trend of new issues coming in. Right now, if new issues come in, I do not see the chart go up, but rather it looks like history is rewritten and we always started from more issues.

[radekstepan]

Interesting. Can you describe how your process works? Usually people allocate a fixed amount of issues to a milestone/sprint/release and the point of the chart is to show you a progress for that specific milestone. Are you adding more issues into a sprint when the chart shows you are ahead of schedule?

[doniks]

The most important difference is that while approaching the release more regression tests are performed and issues are raised.

There is also a secondary effect of reconsidering whether a specific issue really can/should be targeted for this milestone or moved to a later milestone.

The particular project/milestone I was trying to chart is this one:github.com/ubports/ubuntu-touch/milestone/7 . It's an operating system for mobile phones preparing their next release. The milestone starts off with a selected set of new features/ old issues, but over time issues are added (additional features or new regressions) or pushed out to a later milestone.

[radekstepan]

I see. So basically you'd like to adjust the trendline when new issues are added and on the day/time that an issue is added to add a vertical line up and then start trending downward again.

If you change the code in here:

https://github.com/radekstepan/burnchart/blob/master/src/js/modules/chart/lines.js#L100-L150

I can provide a config and some test cases for this so it's part of master. One way to go about it would be to recalculate the trendline fn every time a new issue has been added and push it to the array of points returned at the end of the function trend().