Deprecating outdated issues on the GitHub public roadmap | Hacker News Reader

Deprecating outdated issues on the GitHub public roadmap

github.com

116 pointspreya2k90 comments

This is why you don’t announce future changes until they’re essentially ready to go out the door. Apple used to understand this better than most. Until you build the feature, you’ll have people asking you constantly when it’ll be done. After you build it, you’ll have people complaining that it doesn’t work exactly like they imagined it in their head. If you end up not building it and say so, you’ll get attacked for it, even by people who never knew about the plans before the cancellation.

Keep quiet about internal plans only you can affect. Let features requests come to you, monitor interest in those, and reply if they’re interesting or not feasible, so you can discuss and figure out what would work best for your users. Engage but don’t commit unless you’re certain something will happen.

I’m surprised the conversation hasn’t devolved to a bigger mess yet (maybe it’s being well moderated). It’s a shame they’re having to preemptively lock issues, but I completely get it. It’s exhausting having to deal with abuse on public forums when you’re on the receiving end and always have to keep your conposure.

I think I agree with you for discrete product releases, but for ongoing SaaS and moreso PaaS, it's helpful for integrations to have some visibility on your roadmap. I'm might just write a hack workaround for some issue if I have a belief that in 2024 you'll solve X -- I've got bigger fish to fry. But if I know you've removed it from your roadmap and it was planned by me, I can put it on mine to resolve. Surprising me with features means more often than I care to admit I write something to solve a problem, use it for a few months, and then vendor comes out with a solution to it that's close enough and better supported so I toss out code -- with a roadmap I could have avoided overlapping efforts.

But I'm more on the operational side - so I care more than most about the integrations with lots of vendors, different PoV's may of course differ.

I feel you. You seem to be reasonable and empathetic and understand that sometimes priorities change and that an imperfect roadmap can still bring benefits over a hidden one. On the other hand, there are a lot of people who feel entitled to get anything you ever mentioned, even (especially) if they’re getting it for free, and managing those people when they complain is time consuming and drains your sanity. Both of which take your time and energy away from improving the product to all the reasonable people.

But that was the whole point. You never had any visibility into their roadmap.

You only evet had visibility into a daydream that changed. What good is, no scratch that, there is no good in knowing something that you only think you know. It's negative value. It's worse than knowing nothing at all. It doesn't matter how much you want to know, and how good it feels to have that want pacified.

Similarly, this is also why even if someone does publish a road map, you should ignore it and only deal with whatever exists as it exists right now. Make all decisions, including looking ahead, based on nothing but the current state.

By that same logic, if somebody tells me they will pick me up on the way to a movie, I should go there on my own because they aren’t here already. I should walk to the movies, because my car is not currently running and whether or not it starts is in the future.

There’s such a thing as trust. Tools, people, and groups that have shown themselves to be trustworthy may continue to be trusted. It is also reasonable to point out these shifts, so that others know when announcements should no longer be trusted.

That is hyperbole not exprapolation.

There was a time when the dream of open source was healthy and strong. GitHub was a big part of the movement. So it shouldn't be surprising that they did more of their planning in the open than a company like Apple.

There are tradeoffs with secrecy. Trust me when I tell you, secrecy comes at a cost. It limits the flow of ideas, even within your own organization.

I'm tempted to agree with you anyways, since I've been in the situation before where we wanted to change things we were making while we were making them. You don't want to give off the impression that you don't know what you're doing.

Honestly, this kinda goes back to agile vs. waterfall too. It's all about defining requirements and meeting them.

> There are tradeoffs with secrecy. Trust me when I tell you, secrecy comes at a cost. It limits the flow of ideas, even within your own organization.

To clarify, I’m not advocating for active secrecy as goal, especially not inside the company. I’m more arguing for not actively sharing with the outside what you don’t have to because they don’t have the same context you do and will judge your decision out of their own biased feelings. Which then leads to you having to waste an inordinate amount of time explaining yourself.

If you consider anything such a big issue, then your suggestion solves nothing - people still complaint, and now they also complaint about no transparency (see Apple)

> It’s exhausting having to deal with abuse on public forums when you’re on the receiving end and always have to keep your conposure.

What's so exhausting about ignoring like all of these corporates do?

> What's so exhausting about ignoring like all of these corporates do?

Maybe have some empathy for the people who have to interact with individual customers. There’s not an amorphous blob of “corporates” who deals with everything, there are real people with an life just as rich as yours on the other side.

Let's test your hypothesis on this issue: a "real person with rich life" posted a corporate statement and interacted 0 times with responses. Was that too exhausting?

We don’t need to go that far, we can test my hypothesis with this conversation. You are becoming exhausting and absolutely proving my point. I explained at length something that is clear and obvious to anyone who does customer support for any popular product. Instead of making an effort to be empathetic to the other side, you’re being contrarian and dismissive. Fortunately you’re not my customer and we’re not in my repo, so I have no obligation to continue this conversation.

The fact that the person on GitHub has not replied and the other conversations have been locked demonstrates that they are preemptively avoiding the exhausting interaction by limiting replies. Those are the actions of someone who knows the cause and effects of burnout. It is the opposite of ignoring, it is actively defensive. If you’re unable or unwilling to understand that, there’s nothing else I can say to you.

For a mature product, I think it makes sense to set an expectation of stability - what you see is what you get and if it’s not good enough as-is, don’t use it.

I don’t think closing off all internal discussion of future improvements necessarily the way to go, though? Sharing things like draft standards and programming languages enhancement proposals seems to work out pretty well.

Issue trackers, not so much. I think part of the problem is that the reply box is right there, inviting drive-by opinions, and makes it hard to keep things in perspective.

As the post suggests - the GitHub community discussions are still the place to talk about these items. This is just closing the sometimes literally years old public roadmap posts.

What’s preferable, clarity on what’s actively planned, or ambiguity with dozens of features languishing on the public roadmap without any updates?

Video game companies used to be like this. There was no community engagement, and the companies were essentially walled gardens. One day a feature (or even whole game!) would appear, and that would be the first you'd heard of it.

The only one that directly annoys me is not being able to have threaded comments at the PR level. https://github.com/github/roadmap/issues/552 You can do it with "quoting", which is fine if there are two of you but turns into a mess if there's more than that.

They've said that they're watching the discussions for feedback, so I hope they listen and implement that one.

Happy that they are being transparent (rather than letting the issues rot), annoyed that they appear to be prioritising marginally useful AI stuff for basic UX.

That is surprising to me. I am so tired of commenting on random lines with the usual "commenting on a random line for the sake of threading" apology.

Also commenting on random irrelevant changed line because you can't comment on unchanged ones (but they might need changes due to rest of the PR).

Github code review is horrible and I think it's actually worse than not having anything at all because now it's hard to convince people we should switch to something better. "Oh I don't want to learn another tool, github is good enough"

> Also commenting on random irrelevant changed line because you can't comment on unchanged ones (but they might need changes due to rest of the PR).

Gitea does this as well, it's a real bugbear.

Shameless plug but this is one of the many GitHub code review limitations I set out to fix when I created CodeApprove (https://codeapprove.com).

You can comment on any line, a whole file, or a whole PR and all comments are threaded and tracked to resolution. They are never hidden because they’re outdated or because the discussion is too long.

Going to stick my neck out here.

A lot of these "improvements" fall into the following 3 categories:

1) More complexity around issue tracking

2) More complexity around permissions

3) IDE-ness and general visual-studioification of the web interface.

Since many of the issues make GitHub bloated and more difficult to use for general use cases, they _should_ be removed.

Yes, the last thing they need is more bloat. I wish they removed achievements as well.

I agree, also the source code viewer (not sure what it's called) has regressed tremendously. A few years back it was simple, quite snappy and just worked. Now it has become slow and bloated and barely works for me on the desktop and the mobile version is full of rendering bugs rendering the whole thing rather useless. (Using firefox both on mobile and on the PC)

All they had to do, literally, was to do nothing but no, can't have that. MUST ADD BLOAT.

I wish it would remain as a basic viewer with syntax highlighting. If you want an edit/cursor mode you should have to click a button to enable it. At the moment, clicking on thepane places a cursor at that location which breaks using the arrow keys for scrolling the window -- if the cursor is off screen it will scroll to the cursor location not where the viewport is currently located.

GitHub is primarily as issue tracker though... and code runner I guess.

I was writing about this. GitHub is essentially a bloated git plus keys plus email list.

But then I remembered about search. I use and enjoy GitHub code search a lot, and it makes me a better developer.

Yes. And maybe this goes for features that are already implemented as well.

Ever since GitHub was bought by Microsoft things have got worse for me as an end user -- browser compatibility is worse and I run into bugs frequently when on not-chrome; their academic programme now requires an insanely invasive localisation check that instantly fails for me on Linux and their support couldn't advance the process as I hadn't submitted an application yet; and their tooling slowly pushes people away from FOSS and towards proprietary methods. I wish they'd figure out how to make it a hacker friendly place again.

out of curiosity what browser do you have in mind?

I use Firefox on Linux and I didn't have any issues annoying enough to stick in my memory.

I use Firefox on Windows (on multiple machines), and when somebody has an image embedded in an issue or comment in Github, and I click it to see the original sized image it will not load, and simply shows a blank page instead. It's been like this for two years or so.

Can you link a specific (publicly accessible) example? I want to try reproducing.

Using Librewolf 131.0.3-1. Going to the GitHub homepage makes the tab freeze and I have to go to /login or a repository page directly.

Same here.

(I think I've ever seen one minor bug: for any repo in the clone dropdown, the "SSH" tab was hidden for me a while ago. Unfortunately I don't remember if this was browser related.)

No op, but videos/gif (used a lot for demos) don't work , you have to refresh the page

> and their tooling slowly pushes people away from FOSS and towards proprietary methods.

Can you explain this one in a bit more detail?

They said they are pushed away from Linux and non-Chrome

Translation: here's a bunch of useful features requests customers asked for, but we don't have engineering budget to build them and not enough customers asked for them, so we're not building them. Plus they don't say "AI" so our executive team said to axe them.

At my project we are forced to use Enterprise GitHub including Actions. Often 80% of the build time consists of up- or downloading intermediate or final build artifacts. This can take more than 40 minutes which is extremely painful. I was really looking forward to finally getting a speedup here. The promise:

> GitHub Actions: Artifacts v4 available in GitHub Enterprise Server #930 … We will be extending support for v4 of the actions to upload and download artifacts to GitHub Enterprise Server (GHES). This new version improves artifact upload and download speeds by up to 98%.

I don‘t understand at all how this is not a priority anymore. :-(

I believe the future is GitHub enterprise cloud. https://docs.github.com/en/enterprise-cloud@latest/admin/ove... all the benefits of GitHub enterprise, with none of the cons

> all the benefits of GitHub enterprise, with none of the con

To some organizations, having to use a cloud SaaS instead of hosting the application on-prem is a con.

Isn't that the value GitLab is trying to provide?

Perhaps. The point remains that GitHub currently offers a self-hosted solution (GitHub Enteprise Server), and saying that the cloud version is the exact same (no cons) isn't true.

Why bother with Github Artifacts at all, especially for intermediate steps?

At our self-hosted GHE instance, we use either (self-hosted) Artifactory, or raw S3. Fast, and easy to work with.

"After an in-depth review, we’ve identified a number of open issues that have become outdated over time—some for several years."

Sounds fair enough to me.

"Leave them until they're outdated" isn't a valid way to deal with issues in my book.

Plenty of large projects need to go through and do a bit of a sweep of old issues every few years. If your projects have the discipline to never need that I congratulate you.

But "Don't sweat over any little issue, especially if they have gone multiple years without major complaints from a sizeable part of your user base" is.

Clutter accumulates in any bug tracker.

Ideally they'd've been cleaning it out more regularly and these feature requests would've been marked "not currently on the roadmap" and closed much sooner.

But I don't think I've ever seen a team do a perfect job of that, and I'm probably worse at it than they are.

If they were all truly outdated, then sure. But as the comments have pointed out, Github also closed at least half a dozen still-relevant issues like "Commenting on unchanged lines in a pull request" that would be significant UX improvements.

That's a big one. I need to deal with perforce which has plenty not to like, but I frequently attach comments on unchanged lines in code reviews saying something like "you'll need to change this too", and their code review tool supports this.

Where did the "basic functionality" quote come from? I haven't looked at all 42 in detail, but it largely seems like _advanced_ functionality to me.

Yeah I don't see the term "basic functionality" anywhere on the linked page, having that in quotes in the post title here feels misleading to me.

Renaming this post to match the original title would be appropriate: "Deprecating Outdated Issues on the GitHub Public Roadmap"

I've emailed @dang

This is exactly why HN discourages editorialising titles. This one was clearly modified to spark outrage. People click on it with a predisposition to be mad instead of being given the chance to reflect first.

How on earth is threading comments on PRs advanced in any way?

Maybe overcomplicated is the better term? Constantly adding "basic" features on top of each other might create us a Jira.

We've reverted the title now. (Submitted title was "GitHub removes 42 "basic functionality" features from their roadmap")

Submitters: "Please use the original title, unless it is misleading or linkbait; don't editorialize." - https://news.ycombinator.com/newsguidelines.html

The trouble with bug tracking / project management software is that everybody wants Just One More Feature.

But if you implement them all, it will become an overcomplicated mess, somebody will replace it with a simpler version, and the cycle will repeat.

Would I like some of the features they've decided not to implement? Yes.

Would I hate the results if they implemented everything on this list? Also yes.

And making everything configurable so you can pick the exact subset you want is (a) an incredible amount of work to make the resulting combinatorial explosion of possible choices all work nicely (b) tends to inevitably lead to something like JIRA.

I do appreciate people being annoyed about specific features they'd really like getting removed from the roadmap, but so it goes.

There's some weird stuff there.

> GitHub Actions: Artifacts v4 available in GitHub Enterprise Server

So they're deprecating Artifacts V3 next week, and now announced they won't upgrade Enterprise Server to v4?

This one doesn’t make any sense, it HAS to be a mistake

Probably a subtle push to sell the cloud-hosted Enterprise version. I bet the on-prem Enterprise will slowly stop getting all the latest-and-greatest features. Producing and supporting the on-prem version has extra costs, but probably doesn't bring in enough (extra) revenue to justify working on it as much.

Heh even CISA.gov is grumpy: https://github.com/github/roadmap/discussions/1014#discussio...

Personally I'm not a fan of "we haven't got round to it in ages, let's close it"

Issues at the bottom of your backlog:

a) Cost you basically nothing

b) Document previous demand

c) Can be useful tasks for new joiners who are skilling up on the project

d) Can be bumped if demand re-awakens

e) Documents known feature gaps

I assume none of their AI features were deprecated.

Whilst I'd still take it with a grain of salt, two have been. Mostly their code scanning AI stuff though (or pieces of them), which... Has had a lot of issues, compared to more traditional approaches.

+ Code scanning: AI-powered autofixes for CodeQL alerts integrated into VS Code

+ Secret scanning push protection for gists

CodeQL has never worked anyway and I’m unsure why it’s even a thing given it doesn’t work

Could you expand on what doesn't work for you?

all of the important AI features, like building models from the code, were almost certainly pushed.

customers don't care about those, though, and this is a good way to distract from any discussions about the implications.

This was a bad way to announce closing these issues. With a big list like that, everyone's going to find something they're pissed off about not having. They should have closed the issues slowly and quietly in the background - individually they're small enough that people probably wouldn't notice. But like this, many more people see what's happening and get angry together. Not a good way to do PR.

Maybe the idea was to rip the band-aid off and hope the outrage burns out quickly.

"At GitHub, transparency and clarity are at the heart of our relationship with the community."

When these kind of carefully crafted prety slogans has to be a prime statement before anything is told then my suspicious mind gets very alert. Probably even overcompensate. Do people take these kind of forefront self evaluations as facts or have suspicion when this has to be announced and highlighted, stated (instead of being obvious). I do not know why (of course I do!) but these kind of self admirative statements have the opposite effect on me. Even when they are true.

I mean, it was clearly phrased by the marketing team, but it's on a post where they actively close a lot of bugs and clarify they won't be doing them, instead of doing what most companies do and just ignoring them for years.

I'm confused -- are they just closing the issues because they're outdated for various reasons, or are they yanking those features entirely where there's some attempt already made?

For example, the top one in that list is the "Command Palette" -- but it's already live and working fine! And I'm pretty sure "Precise code navigation" also already exists for TypeScript.

So are these features that are already GA going to be removed..?

Yeah, I hope it doesn't mean that they will remove Command Palette. Unfortunately, it's still an opt-in "feature preview", so who knows what this means. For me it's a key navigation tool in GitHub, but I can understand that it's more of a power user feature.

> "Precise code navigation" also already exists for TypeScript

We had conflated JS and TS support into a single release issue. JS support never landed, but TS did: https://github.blog/changelog/2024-03-14-precise-code-naviga...

When we first created the release issue, the thinking was that we wanted to launch support for those two languages at the same time, and for that support to be compatible with each other. (So we could correctly follow e.g. a TS library referencing a definition in an upstream JS library.)

Unfortunately we never got JS support to the point where we could GA it. Largely for scaling reasons — there is a HUGE amount of JS on GitHub, and JS is a dynamic enough language that we have to do more per-file processing on it than for TS or Python. We decided not to wait for JS to be GA-ready before releasing TS, but then never corrected the release issue to account for that.

I wish they put their focus more on GitHub Projects/Issues, it is just too slow.

no idea how [this](https://github.com/github/roadmap/issues/552) was closed as "outdated"

Is IPv6 reachability on some other issue tracker or why am I not finding it?

Could Github developer workflow metadata (e.g. issues) be exported/serialized into a git repo for decentralized replication and/or import into an alternative?

not to git but there are bots to export to ie gitlab

Yes. I built a tool for that a few months ago: https://github.com/simonw/fetch-github-issues

I have it running in one of my repos using a GitHub Action that's triggered when an issue is created or updated - it commits a JSON export of that issue back to the repo.

Then I can git clone the repo and get all of the issues data.

Perfect, thanks!

What the fuck? A casual browse of that list and these are very real features that appear to already have had some development.

No command palette? JS and TS precise code navigation being cancelled? SSH connections to GitHub Actions?

What on earth is this new “roadmap” then? More AI garbage slop and less focus on developer tooling and source control?

This is a very dark day.

> Commenting on unchanged lines in a pull request

This is infuriating that it's missing. Not being able to just say "Hey, missed updating this line" is just an insane oversight

I wish they would put more effort in Github Issues (the project management product). They are so close to have something that just exactly what I need. But it seems like they haven't touched it in a couple of years. There are still many rough edges that has been there since the beta.

Like issue search being not instant, it's impossible to find duplicates if it takes 5 seconds per search and not 500ms.

The inability to comment on any line in a PR is a real pain. I’m disappointed to see the plans for this abandoned.

Guess who turned GitHub into a SPOF?

We did.

Death by committee or death by microsoft, take your pick.

I am going to call it, Microsoft has not been particularly good stewards of GitHub.

They have over complicated so much of it to please corporate customers that it has really lost what made it great to begin with.

It used to make everything seem simple and manageable. Changes were slow, sure, but they felt like they were at least thought through. The docs used to be simple and easy to navigate. Everything is about 10x more complex now.

Reply on news.ycombinator.com