The beta.rust-lang.org PR debacle


#63

Just wondering: would it solve most of the complaints here if we took the current home page and put it above the fold in the new website while keeping the rest of the new website exactly as it is right now? We’ll then get code samples, a better slogan (subjectively speaking) as well as all the goodness from the new website.


#64

Here’s another thought: Do we even need a slogan? The headlines Performance, Reliability, Productivity in the section Why Rust? already show Rust’s strengths. Also, these words are meaningful to both programmers and economists.

My idea is to remove the tagline without replacement. However, a code sample would be really helpful for programmers to get a first impression. If you can’t agree which code sample to choose, don’t choose. Instead, randomly pick a sample every time the page is loaded.

Here’s a layout you could use for inspiration:


#65

I really hope the core team goes back to the drawing board completely with the design. Everything from the font and color choices to the layout to the navigation is a downgrade and real departure from the established Rust identity. Examples of similar projects with more effective front pages:


https://rocket.rs
https://elm-lang.org


#66

This is way overstated, in my opinion. While it’s certainly not perfect yet, I think it does a far better job than some of the other examples linked in your post. For example, while the slogan is pretty content-less, it quickly gets to a “Why Rust” section is persuasive about Rust being unique – that’s far better than one of those other languages that starts with “we have syntax that looks like another language” (a claim that makes me think the language has nothing interesting if that’s what the page starts with).


#67

And locked as well. I’ve never seen a GitHub issue locked so quickly. There’s no way to upvote, discuss, etc.


#68

Summarizing the content feedback:

Don’t describe it — be it.

I.e., simply show me a documentation sample and give a link to the rest, and I’ll decide for myself if it’s quality. Don’t merely tell me, “We have great documentation”. :slight_smile:

Reminds me of dating profiles that say, “I’m intelligent, witty…” instead of actually saying something intelligent and witty. These seem to disprove themselves.


#69
Oh wow, that is very rough handling! :( I get that there's a deadline for 2018 edition launch, but immediately and categorically shutting down feedback, telling people to stop commenting, and locking down issues — that's the treatment that spammers and harassers should get, not someone who posted a mock-up of alternative design (and IMHO a very good one!)

I haven’t heard of the new site/design before it was effectively final, and I’m frequently reading all Rust-related news. I wish there was more outreach about it, and not “bold” (to put mildly) design set just before the deadline.


#70

My statement may have been heavy handed, my apologies if so. That said, you are arguing that the content of the linked sites is not up to the level of the beta Rust site. While I think your opinion of Crystal (my inference based on your comment) is debatable, it is orthogonal to my original statement regarding site design.

My intent was to imply a preference for:

  • Fewer colors
  • Focus on readability
  • Most important features most prominent
  • Restrained rather than garish

#71

I quite like the minimalist design of the current Rust page, sort of like Google clinging to the same design philosophy around their search page (some of us might remember the “portal” landing pages other search engines had back then). The current Rust page is nice and clean.

And I’m also confused on why the Rust team took on this project in the first place, at least at this point in time. Is there really a belief that Rust’s adoption is hampered by some perceived lack of a flashier landing page? Or any page at all, really. I can say that in my organization and people that I’ve talked to elsewhere, the landing page has never been mentioned, nothing even remotely close. There are much bigger fish to fry and barriers to adoption, and I’m just really surprised that this would be a project right now, particularly since one may, a priori, imagine the community bikeshedding this to death.

But the way the github issue that @dogweather linked was handled is just, in all honesty and with due respect to the relevant folks, abysmal. Of course, the way the page was “rolled out” wasn’t an RFC but request for bug reports, and that sets the stage for this type of thing. But whatever, it’s been done - I hope valuable lessons were learned and we can, collectively, improve going forward.

The Rust community and all the people involved in evolving the language, stdlib, tooling, outreach/conferences/videos, etc are superb - it’s a headscratcher to then see this type of thing pop up.


#72

The Rust community and all the people involved in evolving the language, stdlib, tooling, outreach/conferences/videos, etc are superb - it’s a headscratcher to then see this type of thing pop up.

It’s really not, though. Diversity of opinion has never been popular around the Rust community, which is a real shame. I mean, just look at the posts that have been hidden in this very thread. :frowning: I hope the teams take this as a wake up call and genuinely try to do better going forward.


#73

The PR is literally against the contribution guidelines. https://github.com/rust-lang/beta.rust-lang.org/blob/master/CONTRIBUTING.md#design-and-content
This is handled like in any other FOSS project. PRs with no expectation on being accepted or merged in their form should be closed immediately. Especially, the contribution guidelines avoid “proposal” PRs, which I find a reasonable thing.

There’s no point in keeping it open or have it linger as a discussion spot. This spams people issue tracker inboxes which we all have to use for work. It adds nothing but noise there. Discussion about the design is fine and accepted and if you want to present an alternative design, please do it here or at other spots.

There’s tons of high-quality design feedback here, which is discussed and listened to. https://github.com/rust-lang/beta.rust-lang.org/issues


#74

There’s not a belief, we are sure of that. The old page was already a liability for 2 years. It was hard to maintain, at many places factually wrong and we have frequently noticed that people couldn’t place the language.

I do a lot of booth service for Rust (as in: I stand at conferences taking questions of people for a whole day) and “I checked the website and found none of what you tell me” was very usual. The most common question was “Yeah, it’s a programming language, alright, but what is it used for?”. That was an extreme problem for embedded folks, because at no place, embedded workloads were covered.

We have been carrying that page around for ages, with no clear strategy. The old page was so useless that we usually didn’t even point people to it.

The planning around the page has started at Rust all-hands 2018, with feedback from all represented working groups and people that want to publish on it. That should be taken into account. If you argue for not showing aspects like network programming and embedded upfront like the current page was, you are also asking all those working groups to take a step back to have their work presented.

I have some more thoughts collected here: https://twitter.com/Argorak/status/1069195278350200832


#75

The unfriendly contribution guidelines that outright reject all design feedback are part of the problem.


#76

I don’t agree, otherwise we’d have everyone all over the place with design. We’re accepting a lot of feedback on the design (see the issues lists), but outright full change requests are at the discretion of the responsible people. The guidelines specifically don’t accept any kind of out-of-the-blue PRs without previous discussion issues.

These are - by the way - the guidelines to which everyone working on the page during the closed phase was bound as well.

Also: you are putting words in our mouth here. The guidelines reject “large” changes, everything else is fair. We don’t reject all design feedback and the guidelines do not even hint at that at all. Specifically, the guideline is:

Design and content are complicated thing, and many folks have lots of polarizing feels about it. For the moment, large design changes are made exclusively by the team (listed above). If you have concerns or thoughts for small changes, please file an issue so that we can discuss it first!

It gives a direct and clear guidance on how to start with that change and the PR ignores all of that.


#77

Ah, this finally gets to my question. I assume the big-picture design decisions, like removing the code snippet or choosing the color palette, happened during this “closed phase”? And were so convincingly decided that broader community discussion is actively discouraged (in the place the decision-makers actually attend, not here on the sidelines)?

So, with one week left to have input, the community is asked… for what, exactly? Typos and broken links? Why is this even on GitHub?


#78

I agree. It’s very tempting to say “everybody must be involved at every stage”, but in practice this is impossible. Especially in an issue as divisive and bikesheddy as this, there must be some line drawn which says “this is reserved to people who actually work on and maintain this”.


#79

At this point, a fully closed process and simply unveiling a new design would have been preferable, yes. This instead feels like “we’ll do it closed, but pretend it was open for appearance”. I’m sure that wasn’t the intent, but asking for feedback on a big change, then categorically rejecting all substantial feedback, is going to upset the people whose time and effort were solicited.


#80

This sounds like a content issue, not a visual design one.

Same - sounds like content.

It’s a landing page, it’s not supposed to be extremely useful - it’s meant to be (I think) a portal with curated links for further reading, support/discussion channels, and so on. It’s not a manual so it’ll never satisfy every conceivable visitor.

Again, content.

I think most of the negative feedback on the beta is around its visual aspects, such as the chosen color palette. Was there really no way to improve the content but leave the basic visual cues and styles as they were?

And I’ll reiterate my belief that the landing page is not holding back adoption - I say this as someone evangelizing Rust to my org, and fielding many questions about Rust in the process from people of varying background. People that visit the Rust site, don’t find answers to their questions and then proceed to ignore Rust because of that, wouldn’t adopt Rust anyway because the path to adoption requires a lot more effort than they’re clearly willing to expend.

I’d love to discuss ideas for improving adoption, perhaps a separate thread, if folks are interested in doing that.


#81

The Rust-the-language manages to include community in the design just fine:

  1. Ideas are floated and discussed (in the open),
  2. best pieces are collected together and polished in RFCs,
  3. and then the team makes executive decision to ensure the final design is sound and coherent.

There’s nothing wrong with the third step per-se. However, the redesign of the website seems to have skipped steps 1 and 2 entirely, which left the community out of the process.


#82

Okay, I’m out. I gave you a pretty extensive response and you cherry-pick one of the sentences that suits you while not addressing that I feel you misrepresented the contribution guidelines. I’m okay with you feeling angry and respect your work, but this is not a basis on which I want to interact with you.