Re: [rdf4j-dev] backlog refinement: closing stale issues

["Jeen Broekstra" <jeen@xxxxxxxxxxxx>]

On Fri, 9 Apr 2021, at 09:28, Håvard Ottestad wrote:

I think we should keep any issues that are actual bugs, regardless of if anyone ever intends to fix them. Same for PRs that demonstrate bugs. 

The issue backlog is not an archive, it's intended to reflect things that we plan to address at some point. 

An issue that is never looked at, never picked up, never commented on is just dead weight - even if it is a real bug[1] or a really good improvement idea, apparently nobody cares. Closing them with a "wontfix" or "stale" tag is the clearest signal we can give that it's just not going to be addressed unless someone shouts out and takes ownership. But keeping it open is just fooling ourselves - and frankly having such "dead" issues in the backlog gets in my way when planning things. 

I want to get to a stage where we can look at the backlog with confidence that most issues are at least somewhat relevant, current and likely to be picked up at some stage. I also want to encourage people to step up and take ownership, and not just wait for someone else to "hopefully" fix a thing at some unspecified point in the future.

The same holds for pull requests. I don't mind draft PRs or PRs that demonstrate a problem, that can be useful. But code grows stale. A lot of old (draft) open pull requests and unmerged branches that no-one takes ownership of just creates clutter - actual code clutter (branches with merge conflicts, badly written "temporary" code, etc) as well as mental clutter ("is this still relevant? who is actually looking at this? should I start a new branch or use the existing branch?"). 

If no-one is willing to pick it up then at some point it just has to go. The pull request tab is supposed to show "this is being worked on", not "here are all the old broken toys we have in our attic". What's the point of having a PR if no-one is willing to fix it? 

Note, by the way, that closing an issue or PR is not deleting it. You can still look at them, you can still find them with a search. If someone jumps up and says "But this is really important!" it can be reopened, or a new issue with more up-to-date info can be created. But until then there's no point to having it open forever (as for branches associated with such PRs, we'll take that on a case-by-case basis, but at some point we should also just get rid of old unmerged branches). 

Also, for what it's worth: we're not just closing everything old of course. We are talking, in most cases, about issues that have been untouched for the last 3-4 years or more.  We also have the intermediary solution of marking issues as "help wanted", which I am using for several issues where I am still somewhat hopeful that somebody might step up.

TL;DR: don't be afraid to close an issue or a PR. It's not irretrievable if needed, and it creates clarity about what our current priorities are. If we close something that is still relevant, someone will speak up. In fact, that's what I'm hoping to encourage: people speaking up about what's important.

Cheers,

Jeen

[1] If a bug happens but nobody is around to complain about it, is it a real bug? :)