Followup on Website Concerns

Part of the Rust 2018 release was a new web site, overseen by the Core Team, with help from external designers, as well as content work and feedback from all of the Rust Teams.

While the site did go live last week, there were some concerns about the process leading up to its release, which was different from how other parts of the project are typically developed. In particular, there was relatively little room for community-wide input in the process.

The process we ended up following was not what was originally intended, but we failed to communicate well with the broader community about what was going on. Due to the timing around these steps and the Rust 2018 release, we’ve also been unable to respond quickly to these concerns when they were raised. But we do hear the concerns, and we want to work through them!

Our plan is to put together a retrospective for the web site work (and to highlight the various concerns) early in 2019. We’ll open a thread here on internals, with a link to the retrospective post when it is posted, and work together as a community to determine how to move forward. That gives everyone time to enjoy the holidays — and play with Rust 2018 — before fully engaging in discussion about the process.

In the meantime, the new web site will stay in place as-is, aside from ongoing tweaks and small improvements. Along those lines, we ask that people with process concerns or proposals for major overhauls to the website hold off on opening issues on these topics for the time being.

Thank you!

The Rust Core Team

31 Likes

As a note, most of the core team will be away until the new year, so folks will mostly not be responding on this (and other threads). We will read comments here when we get back, and will try to respond, either collectively or individually.

I for one appreciate this. Although the discussion about the website has sometimes gotten heated, overall it’s a pretty typical example of a group of people feeling unheard and ignored. The website itself will be fine: it’s already substantially improved in the aspects people were complaining about, and hopefully will continue to evolve now that the time crunch is gone. But the community reaction clearly indicates a process failure, which needs to be addressed for the future. I was worried that the complaint thread would just fester indefinitely without any real “official” response, until it inevitably lost steam, leaving the issue unresolved. Guess I should have had more faith.

This is not the first time we’ve been through this. Aaron Turon’s blog post series about “listening and trust”, written this May in response to a different set of controversies, is highly relevant here as well. That controversies will pop up from time to time is an inevitable consequence of a continually growing community, but I’d say Rust has a proud history of doing community right, and that includes treating each one as a learning experience. Whenever there’s an “us vs. them” feeling between elements of the broader community and the teams, something has gone wrong. Instead, as he wrote,

If we can trust each other to listen and take concerns seriously, we free ourselves to be uncertain about those concerns, and open to possibilities that superficially work against them.

Hopefully we can allow that to happen here too, and have a meaningful, constructive discussion about the website process and what can be improved. :slight_smile:

25 Likes

If not everyone likes the new website then that’s a pretty good sign that the rest of the release of Rust 2018 went smoothly!!! - Well done all.

4 Likes

On the other hand, this is also what I find a little worrying about the website episode. It seems like the problems with the website process have been similar to earlier instances (I think the one that keeps popping up is an example about the integer type names from the pre-1.0 days, and maybe the early versions of the module system); as such, it feels like the learning experience still needs improvement.

6 Likes

As this was cross-posted from users, and might also be better followup to The beta.rust-lang.org PR debacle, I’ll offer one point of user perspective here, which I hope you will consider in the new year:

The moderation measures applied during this and other incidents, like locking of topics and lots of hidden (flagged) comments, appear in most cases cases to go beyond any tangible enforcement of the code-of-conduct. The practice extends to github issues as well.

At the very least, these moderation measures have the optics, to outsiders and new-ish users, of censoring dissent and insincere openness.

Personally I think this heavy handed moderation is more worrisome for the future of Rust than any shortcomings of the new website.

21 Likes

Otherwise, what is happening here is analogous to pushing code intoproduction, and then doing review and tests.

This is the part that alerted me a bit, even ignoring my general design and target audience concerns.

While compiler had most of the edition work done in advance and had freeze in couple of week before release, on the website and rustup sides most of the bugfixing work seemed to happen in the last few days before the release, and then after the release and some parts are still broken.
I'm not even responsible for this, but still had a fair share of secondhand embarrassment when release announce praised rustfmt/clippy becoming stable, but they couldn't actually be installed without workarounds.

5 Likes

I deleted my previous post before because in hindsight I could have put things in a more constructive light. I found this old page which I think is a very good model for how effective organisation of a community project should look. This looks very different from the recent commit stream with individuals commiting without any point of communication, plan of action, resulting in zero discussion. There needs to be one person not coding and only coordinating.

Generally, these updates have been discussed on Discord. For example, the blog link thing came up in an infra team meeting, and then was discussed with other people, and then the PR sent.

Thanks for that information. I want to think about if there are any constructive ways I can contribute to the organization of Rust’s online presence.

This highlights everything that I'm uncomfortable about regarding the website, and the situation with Discord in general. My own understanding of the migration away from IRC and to other chat platforms such as Discord and Zulip is purely because the language/compiler teams wanted to have more organization in their discussions, and #rustc and #rust-internals were too noisy for that purpose. I believe Zulip has cleanly fit the role as a replacement for these two IRC channels, but what is Discord? I can see that it even has its own #rust-beginners channel, leading me to think that it is trying to replace the entire Rust IRC community altogether.

This is what's causing the disconnect and miscommunication between the core team and the rest of the community. Though it may appear to the core team that they have solicited community feedback by having their discussion on Discord, in reality Discord does not equate to the entire Rust community -- a substantial amount of people is still on IRC and other platforms. I would really encourage the core team to consider not splitting the community by trying to replicate the IRC community on other chat platforms.

3 Likes

To be clear, I did not say that we think we accepted feedback on discord and that's the cause of this. I'm saying that sometimes, bugs get reported on chat platforms, and then get fixed without a ton of in-repo discussion. That would be true no matter the platform, it happens on IRC too. Or even Twitter. Not everyone reports bugs by opening issues, and sometimes fixes are small enough that it's not worth opening an issue just to immediately open a PR fixing that issue.

The Discord situation is entirely separate.

@KiChjang, I completely agree with you. Also, using Discord excludes people like me and many others who are not in the United States timezone. This is contrary to Rust’s stated goals of being international. Also, rather than chat rooms, written records are very important so that Rust teams or anyone can clearly concretely demonstrate community involvement if there are any concerns about this. It is a very good way to learn from the past and improve.

If you're talking about the blog link removal, that was an urgent need from the release team, since we needed to get 1.31.1 out and the link was still pointing to 1.31.0. After the fact we opened a thread here on internals to ask for help/suggestions on how to add it back without requiring us to deploy new changes to the website each release. There is still no comment on the post, so if you have any idea/suggestion please post on it!

While the core team initially proposed the move to Discord, every single Rust team decided in autonomy if they wanted to move away from IRC and where. For example, the Infra team decided to move away because we were getting a lot of spam in the IRC channel, and Discord provides better moderation tools to prevent that kind of stuff. Other teams are still on IRC, and other migrated to Zulip, all in full autonomy.

Using Discord/Zulip as a chat platform instead of IRC can actually help more people outside of the European/US timezones to be involved in the discussions, since you don't need to have a bouncer up and running somewhere to be able to read all the messages when you log in.

Are you suggesting the Rust project should stop using chat and just use issues for every discussion? That would kill the productivity of a lot of teams :(

4 Likes

Moderator note: As always, if you have questions about moderation then please email rust-mods@rust-lang.com.

1 Like

Disagree with this. Well yeah, you end up missing meetings, but most of the other communications can be followed async-ly.

I’d also like to add that a substantial amount of teams are based in Europe.

2 Likes

I think the rust team do exceptionally well in involving the community. When most people say consultation what they really mean is they’ve already decided. With rust, decisions are generally made by consensus. This is a testament to the quality of the people involved in the project and its organization - it is much much easier to make decisions as an individual than it is to take all the different views into consideration. Add to that the fact that many people involved in rust have never met in person, and use text to communicate. It’s a staggering achievement.

I do happen to think that there are some problems with the website. I do worry that people will go on it and not know what rust is (a programming language), but I’m sure these issues will get fixed, and there are things I like about compared to the old one. Everyone should just spend more time appreciating how awesome rust and the teams behind it is. Theoretical research right through to great end-user experience.

We can always improve - no community is ever perfect, but given how dysfunctional other parts of society are, we really should spend more time marveling at the achievements rust has already made.

7 Likes

Has the retrospective been published somewhere?

8 Likes

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.