Hi Christoph --

Thanks for sharing you thoughts. Let's see how this new method of issue tracking gets accepted. Then we can surely also talk about integration into squeak-dev.

Keep it simple and agile. :-)


Am 02.04.2022 00:25:02 schrieb christoph.thiede@student.hpi.uni-potsdam.de <christoph.thiede@student.hpi.uni-potsdam.de>:

Hi Craig, dear board, dear all,

wow, I seem to be a bit late on the party, but nevertheless, I wanted to share some thoughts about this - I have been advocating for GitHub issues for a longer time already, and I'm really excited to see some progress here. Thanks!

I have quite a few ideas, trying to be as brief as possible:

1. Connection to mailing list

As mentioned by others already, advantages of the ML include a high reach within the community and (basic) independency of other platforms (at least, all posts will be preserved in our archives). I would strongly opt for setting up the equivalent of OpenSmalltalk Bot [1] for the new issue tracker, too. Three questions:

i. Who are the developers/maintainers of this bot? Would you mind open-sourcing the scripts/infrastructure for it? If it is a simple process, I would volunteer to try my luck if no one else is available. Maybe we could just extend the scope of the existing bot?
ii. What are the current limitations of the bot? I think metadata such as labels/assignees will not be reported to the mailing list, but we might renounce that. Ideally, there will be a repository to which everyone interested could contribute in order to improve i.e. the formatting of bot messages.
iii. If the issue tracker will be used more frequently, I can put integration of GitHub Issues to the backlog of Squeak Inbox Talk (volunteers would also be welcome, of course!). This might be a helpful combination.

2. Design of the issue tracker

I'm glad that a first step has been taken, but my personal demands on the appearance of the issue tracker would still be higher. Some proposals:

i. Could we please rename this repo to just "squeak-issues" or "squeak-image"? I understand the idea behind the title "object-memory", but IMHO it is a bit too abstract - when I stumbled upon the repo for the first time, I would rather have expected a VM component or something like that. Our goal should be the easiest recognizability that is possible.
ii. I would volunteer to extend the repo structure with a few labels, issue templates, a nicer readme, etc. Also, I could try to maintain the issue list a bit (assigning labels, closing issues etc.). However, all of this requires write access to the repository - would it be possible to grant write access to all (interested) core developers? Of course, we can discuss any proposals for a labeling scheme or issue templates before.
iii. Do we need any guidelines/etiquette for the issue tracker, for instance with regard to relevance or closing reasons for issues? I might be not the only person to have dozens of bug reports, feature requests, etc. lying on my desktop. Would it be helpful to copy all of them to the new issue tracker or should we restrain ourselves by some criteria? (What I would dislike is something like auto-closing of issues older than x days. IMO this not much more than a PR trick for organizations that have lost overview of their issues, and it is very discouraging to bug reporters.)
iv. Are we also intending to task planning features of GitHub (such as milestones, assignees, projects)?

3. Notes

- As the question has arisen about individual email notifications: If you have a GitHub account, you can watch all activities on the repository via the Watch button in the upper right corner. You can also subscribe to individual conversations via the notifications widget in the right sidebar of any issue. You don't need GitHub Actions for this. However, I do not think that this is an alternative to a mailing list connection for everyone - we should really try to merge all artifacts into a single place.
- I also like the new "Send feedback" item in the help menu. How would you think about improving automation for this? We could try to build a mix between Squot's feedback dialog and the solution I built for Squeak Inbox Talk. Ideally, everyone could file their feedback without leaving their image or creating a GitHub account, and each feedback could contain the basic environment data automatically.

Looking forward to your replies! Have a nice weekend! :-)


[1] https://github.com/OpenSmalltalk-Bot

Sent from Squeak Inbox Talk

On 2022-03-01T21:21:53-08:00, craig@blackpagedigital.com wrote:

> Hi all--
> The Squeak board is trying an experiment. We've created the GitHub
> repo "squeak-object-memory", and are curious to see how useful it is to
> track issues there.
> With this message, I can close the first issue[1]. :) Please
> bring on the other thirty gazillion of them...
> thanks!
> Craig, on behalf of the Squeak board
> [1] https://github.com/squeak-smalltalk/squeak-object-memory/issues/1
> --
> Craig Latta :: research computer scientist
> Black Page Digital :: Berkeley, California
> 663137D7940BF5C0AFC :: 1349FB2ADA32C4D5314CE