View source for Feature planning and better involvement of external contributors

[[Category:QtCS2023]]

==Session Summary==

Explore options for how external (not employed by The Qt Company) contributors could join grooming/planning sessions for their areas of interest. Vladimir recently started this effort in the context of Qt Network, but other domains like QtCore may benefit from it too.

==Session Owners==


==Notes==


This is a raw dump, Vladimir will tidy it:


<nowiki>#</nowiki> Common communication

We were all on IRC, now we're not.

We used to use email to mailing lists. Are we mssing using media

Ulf: We need to commit to one and use it.

Christian: We did set up Matrix, no-one used it

Tukka: We need to understand the purpose of any messaging. There's a need for instant and storage.

The official thing for decisions is email.

Are we replacing or enhancing what we currently have.

Volker:

Email works fine big tasks. It gives people chance to write things up properly.

Not a problem of tools.

Problem is conversation that have happened internally, and it stays internal.

Ulf: A lot of discussions don't take over place email. They're happening in person or over IM.

Actual discussions happened over IRC - and it had open active 3rd parties there.

Albert:

I make discoveries *in* the release notes and there are surprises. Sharing roadmap ahead of time would be better.

Nico: We had lots of announcements and promises about "QML2" and then it didn't materisalise, we were in the dark.

Discussion about communication media and formats.

Tukka:

We have: blog, forum, wiki

Guiseppe:  we have too many communication channels, but no standardisation of what is where

ML is good for announcing. ML is bad at brainstorming. Forum is good at that?

In general ML is bad at brainstorming, and then were no contributions.

Thiago: but the topic you had was super complex. People assume cleverer people are handling it

<nowiki>----</nowiki>

Volker: We should announce emails saying "we will discuss XYZ at XYZ on this platform". Anyone who's interested would get involved regardless.

Ulf: It misses spontanaity.

Vladamir: We should have more meetings on specific topics, like platform people are doing.

<nowiki>----</nowiki>

Some projects use Youtube for the public meetings. There's less typing.

Even with Qt not all people know everythign that's going on

30 people in a call is useless.

<nowiki>------</nowiki>

There is a call for sales, product, somethingelse, where they share what they have been up to.

Tukka: Should we put discussions on forums, instead of the ML.

Albert: it's not about the media of messaging, it's having the content

Marc Mutz: "What happens in the cmopany, stays in teh company"

Having it in the team targets would make a difference

<nowiki>------</nowiki>

Jupyter moved from having meetings to a "collab cafe" where people come with topics and then people just discuss the things they want to talk about and breakout.

"it's not the tool that matters". Topics are made on the fly. It feels more engaging, and it's the official meeting for people.

Thiago: "I don't know what I don't know" - "but more is bad".

Practice "office hours", then we can deep dive into it

hearing from 3rd parties before people do big patches would be good. It makes life easier.

<nowiki>--------------</nowiki>

Lets have a channel for "first patch". It bounces people off. 'No first patch dies silently'.

"Stage button ettiquite" can be confusing. It needs something, maybe a "first commit" badge

<nowiki>------</nowiki>

Utilise Qt Acadmey(!) on how to contribute to Qt.

<nowiki>-----------</nowiki>

Some way to know if a patch is going to be accepted *before* writing it is important.

The ML is the answer.

(Ulf is still there on IRC)

<nowiki>---------</nowiki>

We can start saying on a per module saying who the maintainers are and who to talk to in the modules.

<nowiki>--------------</nowiki>

Concrete actions:

- Get some forums specfically for contributing

- Some other very specfic domains will start having meetings (hardcore C++ people) using platform as a reference

- TQC developers should use ML more (to announce and then turn into disucssion. Including roadmap)