A Seat at the Table and the Art of Business Value by Mark Schwartz

Author:Mark Schwartz
Language: eng
Format: epub
Publisher: IT Revolution Press

* * *

There is a kind of dishonesty in the idea of triaging known defects.

* * *

One small problem: I have learned the hard way that another strategy developers can use for eliminating failing tests is to ... eliminate failing tests. Developers have been known to comment out or delete tests that are not passing, figuring that they will come back to them later to make them work, but in the meantime, the code is passing. This sounds pretty sneaky and terrible, doesn’t it? Well, then, it bears some thought that this clever technique is really equivalent to the technique of triaging defects that have been entered into the bug-tracking system: the developer is making a conscious decision not to fix something right away. There is a kind of dishonesty in the idea of triaging known defects.

Not all testing can be done immediately by the developer. Some defects might be discovered further downstream as users get their hands on the feature, or as other developers integrate their code with this developer’s code. Now we need our defect-tracking system, right? Maybe, but there are some other strategies we can try first. We want to move all downstream testing as far to the left as possible—that is, we want it to happen as soon after the developer finishes his or her code as possible. This is facilitated by Continuous Delivery, since the feature will be making its way through all the remaining steps quickly before deployment.

If someone does discover a defect downstream, one strategy he or she can employ, rather than entering the issue into a defect-tracking system, is to use “wideband communication”—a fancy term for talking. That is to say, he or she can simply show the defect to the developer and have a conversation around it. The developer can ask questions and gather more information from the discoverer. The advantages are a more effective process and faster communication; the disadvantage is that the tester might have to interrupt the developer, who by now is working on the next feature. But if we commit to the idea of fixing defects as soon as possible, then we might need to accept that trade-off.

Finally, we might have to accept that some defects will get entered into the defect-tracking system; at the very least the ones discovered in production. Here, I offer my final suggestion on the matter: defects should never be prioritized or triaged. There are only two types of defects: those that will be fixed immediately, and those that will never be fixed and should not be entered into the defect-tracking system. There is no “later” when it comes to defects.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.