We have discussed providing mgmt. gestures that easily allow developer to subtract from current analysis result set (by disabling rules, by auto-fixing, etc.). What about approaching the problem by thoughtful onboarding of new analyzers/supporting remediation of existing debt?
Current thinking involves better features for baselining existing issues or simply disabling high-volume rules and requiring developers to opt-in to new analysis. Other ideas:
1) allow users to create a 'queue' of checks, where new rules are enabled as engineers make progress driving current enabled set to zero.
2) given a good analyzer 'min-bar', the IDE could automatically elevate all rules for which a code base is clean to 'error'
3) is there an on-boarding experience that can lead engineers through the result set which can be easily/immediately addresses (perhaps by offering to automatically fix all issues addressable with extremely high confidence)
4) what about simple date-driven remediation plans? I.E. developers reduce existing warning set to stay close to a burn-down, the slope of which is established by specifying a specific end-date or a per-day budget (reduce warnings by X each day)
5) how about assisting users with identifying valuable diagnostics by giving them access to collected 'useful/not-useful telemetry'?
We have discussed providing mgmt. gestures that easily allow developer to subtract from current analysis result set (by disabling rules, by auto-fixing, etc.). What about approaching the problem by thoughtful onboarding of new analyzers/supporting remediation of existing debt?
Current thinking involves better features for baselining existing issues or simply disabling high-volume rules and requiring developers to opt-in to new analysis. Other ideas:
1) allow users to create a 'queue' of checks, where new rules are enabled as engineers make progress driving current enabled set to zero. 2) given a good analyzer 'min-bar', the IDE could automatically elevate all rules for which a code base is clean to 'error' 3) is there an on-boarding experience that can lead engineers through the result set which can be easily/immediately addresses (perhaps by offering to automatically fix all issues addressable with extremely high confidence) 4) what about simple date-driven remediation plans? I.E. developers reduce existing warning set to stay close to a burn-down, the slope of which is established by specifying a specific end-date or a per-day budget (reduce warnings by X each day) 5) how about assisting users with identifying valuable diagnostics by giving them access to collected 'useful/not-useful telemetry'?