Thanks Gitlab | Hacker News Reader

Thanks Gitlab

cmpwn.com

146 pointseitland83 comments

Open articleView on HN

My Mastodon is being hugged to death, so here's the content:

@sir@cmpwn.com

Thanks Gitlab

image: https://sr.ht/qAVP.png

@sir@cmpwn.com

If my competitors keep making stupid decisions then making SourceHut successful will be pretty easy

@sir@cmpwn.com

Hi HN, the missing context is that this is signups for sourcehut.org, a GitLab competitor, in the wake of news that they'd be adding third-party telemetry to their software.

---

Note, I mostly use my Mastodon account for shitposting, this isn't a serious marketing avenue or anything. This is my personal Mastodon instance, its infrastructure not holding up under load has no impact on SourceHut availability.

> If my competitors keep making stupid decisions then making SourceHut successful will be pretty easy

I know it is early days for SourceHut but this sounds realistic on some conditions. Due to the closed-source nature of Github + MS and this magnificent own goal from GitLab which disqualifies its claim to supporting "free and open source" projects, it should be only "easy" for SourceHut to be sustainable and supporting free software ethics as long as it resists VC funding and is self-funded with paid support.

Without more convincing features and a matching similar interface rivaling GitHub/GitLab/Bitbucket it may go otherwise, but it already has some of essential features already and I hope this true free-software alternative succeeds.

This is as close as it gets to creating a service that truly supports free-software with paid support.

It will probably never have anywhere near feature parity with GitLab or GitHub in the browser. Part of it's goals are to remain simple and minimize JavaScript usage. Which is noble in it's own right IMO. But it's straight up impossible to do a lot of what Git{Hub,Lab} do without JS.

> And Drew DeVault is kind of an asshole

This sentence was totally uncalled for.

His software is fantastic and he's an exceptional engineer. But I also personally know people that were turned off of Sourcehut because of the way he interacts with the community in some cases. When your project is a one man (now two) show, that kind of thing matters a lot.

C'mon with this gossipy rumor-mongering.

Here, let me fix that for you:

"And I think Drew DeVault is kind of an asshole."

Calling someone an asshole is inherently opinionated. No need to add "think" to the sentence.

lol

I disagree, but I don't have time to elaborate this morning. FWIW:

https://en.wikipedia.org/wiki/General_Semantics

https://en.wikipedia.org/wiki/E-prime

Cheers.

I don't know.

On one hand, I agree that people being less-than-super-kind might be off putting, on the other hand I'm kinda tired about this age's push to always be super condescending about pretty much anything.

Some people have a tough/tough personality, get over it.

>His software is fantastic and he's an exceptional engineer.

At the end of the day that's all that really matters. Look at Linus's many rants. People who get mad at him usually tried to assert their <insert special status here> as a reason for why their code/ideas get rejected. Reality is, most of the code just isn't up to quality standards and maintainers don't have the time to hold your hand and explain why it isn't.

>But I also personally know people that were turned off of Sourcehut because of the way he interacts with the community in some cases.

He made someone mad in one of my channels because he cracked a 10+ year old IRC joke and that person didn't get it. So now of course they are going to boycott everything he does and smear him as a misogynist.

> When your project is a one man (now two) show, that kind of thing matters a lot.

No, this means you can tell people to pound sand and not have to deal with HR and politics. It's the exact opposite of what you think.

We have real problems in the world, like bigger than a bad joke or a rejection letter. We should absolutely shame/ridicule businesses that have revenue models but then choose to open these shady channels on the side (like this telemetry). We should be suspect of VC firms that take <insert brutal regime here>'s money.

None of these statements are what I'd expect from a company. A product is more than the engineering that goes into it, so good engineer or not, if they suck at the rest of it then it doesn't matter. Linus is an exception to the rule, not the rule and lol at comparing people to him.

Maybe stop making misogynistic jokes? Doesn't matter how old they are, not even sure why the age of the joke is relevant.

Lastly, telling people to pound sand has to be one of the worst business strategies I've heard of.

> But it's straight up impossible to do a lot of what Git{Hub,Lab} do without JS.

Like infinite incremental loading that randomly stops working and also breaks browser's built-in search functionality because the thing you want to find is not loaded yet until you scroll down? Yeah, I'm glad they can't do that BS.

Making a good service isn't only about what you can technically do. I'd say it's much more about what you choose to do with whatever tech you have at hand. My experience with GitLag is that it's slow and the UI is generally clumsy (I have to use Google so often to find my way around), and you run into antifeatures like infinite scrolling and this: https://i.imgur.com/Qmtec2b.png

Sadly I think all these sites (SourceHut, Github, Gitlab..) have things to learn from cvsweb. That is, if one could ever convince them to prioritize convenience over prettiness. That is, probably never.

Regarding that screenshot of GitLab:

https://git.sr.ht/~sircmpwn/linux/tree/master/arch/arm/boot/...

Thank you for posting that. Actually I checked it out and contemplated including that very link in my original post, but I wasn't sure if your service was facing the same hug of death that presumably hit your Mastodon. :)

SourceHut is on different infrastructure than my personal Mastodon instance, it can take a lot more heat :)

@clarry, thanks for your comment. It keeps us on our toes.

I'm a Frontend Engineering Manager at GitLab for specifically the area of Create, which is where your screenshot falls under.

We do value convenience and good experience. We're in the final steps of preparing a significant improvement to the Repository Browser [1]. We steered away from things like virtual scrolling exactly for the reason you mention: it breaks the browser's built-in search functionality.

We're currently putting the repository refactor up to feature completeness with a plan to make this as fast and useful of an experience as possible on a modern browser. Navigating between folders is much swifter than if we didn't use any JavaScript at all, just to name one of the benefits. Another benefit of using JS here is the fuzzy file finder (press t on any folder and it pops up). All in the sake of convenience. We hope you'll appreciate the efforts we've put into these details.

Feel free to drop your thoughts on the Epic or creating new issues. Tag @andr3 on them and I'll follow up. We're listening.

[1] https://gitlab.com/groups/gitlab-org/-/epics/159

> If my competitors keep making stupid decisions then making SourceHut successful will be pretty easy

Randy Bush had a favorite phrase: "I encourage my competitors to behave this way"

I'm curios why the post about Gitlab [1] was marked as a duplicate and thereby banished from the front page, without any other post in sight.

YCombinator favoritism? ( Gitlab was part of YCombinator 2015 )

[1] https://news.ycombinator.com/item?id=21343761

Edit: Number 22 yesterday. https://news.ycombinator.com/front?day=2019-10-23

Must have missed it. Seems fine to me.

I don't think so.

Personally, I did flag the post you mentioned, because that same thing (discussing introducing telemetry) was on the front page yesterday (Yesterday), for most of the day (or, at least, for most of the US/Pacific day).

I wouldn't be surprised if there is a policy where if an "item" (a topic, a post, whatever) hits the front page, then it should not appear again immediately after (unless there's some new development).

Ah, I missed it yesterday. Number 22. (https://news.ycombinator.com/front?day=2019-10-23)

It was marked as a duplicate because it was a duplicate of https://news.ycombinator.com/item?id=21344575, which spent over 14 hours on HN's front page and got over 200 upvotes.

Now there are three threads; those two and this one, which is currently on the front page. I'm happy to merge them, but am not sure which submission should win.

Edit: https://news.ycombinator.com/item?id=21343761 has the most substantive discussion of the fresh threads, so I guess we'll undupe that one and merge. Give me a few minutes because it's a bit tricky when there are more than two.

Edit 2: Ok, I've merged all the comments into the thread except for the ones which were explicitly about this post in a way that wouldn't make sense over there.

I ignored the original post due to the boring title, and only read the second post due to the more interesting title. I don't really have a point other than the duplicate was what I was following, while the original was not.

+1 Same here. There were also some interesting comments and discussion going on in that thread. Understand why it was marked as a duplicate, don't understand why the threads weren't combined

edit: Thank you @dang!

Knowing that it was on the front page yesterday makes the duping justifiable to me.

But that probably contributed a lot to pushing up this post, which will produce much worse discussion.

Moderation is hard.

> But that probably contributed a lot to pushing up this post, which will produce much worse discussion

That's exactly right! A subtle observation.

We tend to demote follow-up or copycat posts for that very reason. Not as much in a case like this, though, because of the rule that we moderate less, not more, when a YC startup is involved. I just posted about that at https://news.ycombinator.com/item?id=21346983.

Drew is always on spot with his marketing. :)

I think it would be good to link related blog post - https://sourcehut.org/blog/2019-10-23-srht-puts-users-first/ (maybe that would give some context)

great stuff, an idiotic play by a bigger fish likely provides the biggest marketing opportunity for sourcehut it has ever had up until this point, and this guy is taking advantage of it.

there's also something grand about the fact that this was posted on mastodon (instead of twitter) to complete the FLOSS cosmic justice cycle

OK, Context for people:

Apparently GitLab said it was going to telemetry, so a competitor, SourceHut saw a surge in signups.

To me, it's a sign that if someone is adding telemetry for on-premise software, it better be opt-in.

They weren't adding telemetry to their on-prem software, though.

Yet.

They weren’t, but they also were leaving themselves the ability to without informing their customers.

What am I looking at?

The number of daily sign ups for sourcehut.

This is billed as a competitor to GitLab.

GitLab announced something about allowing some sort of telemetry on its site -- I didn't look into the details but its all over HN -- and people have complained that this might imply privacy violations.

So this alternative to GitLab got a bump in signups.

Thank you for providing context.

We also called out our signups at GitLab when GitHub got acquired, so it is more than fair that Drew is posting this.

BTW As Drew said https://gitlab.com/gitlab-com/www-gitlab-com/merge_requests/... we're pumping the brakes on the telemetry changes.

> [...] we're pumping the brakes on the telemetry changes

So you revert (don't implement it)? Thx! Reverting it is the right decision (I hope that means "pumping the brakes").

From what I can tell they're still going to do it, but in a different way, e.g. it might be opt in (hopefully not in a fake-opt-in style like a GDPR popup) or opt out or whatever.

Fundamentally they want to do telemetry it seems and are trying to figure out a hack to get people to accept it.

(GitLab employee) we're discussing other options too, I'm personally advocating strongly against just going ahead with the current plan. My preference would be for no third-party involvement and just better ways to analyze the data we already have

I got an email from you guys yesterday and I have to say I was quite confused. My company is an EE customer and we pay for our seats. Why are you including this telemetry? For the free version I understand, but you should not be treating your enterprise customers like this.

I'm sorry our plans were so far off the mark.

Pump the brakes? No. Reverse course, before your product ends up in the same place as github.

I got Gitlab's email and immediately knew people would be freaking out about it.

The fallacy of transparency: The more transparent you are, the more people think you're up to some shady shit. It's crazy.

That's not the fallacy of transparency. The real fallacy is when companies think that doing shady shit is OK as long as they're transparent about it.

Someone's going to have to explain to me how Gitlab is doing something shady with this stuff.

Unless I'm misreading all this, it all looks like very standard telemetry, the same kind Github is doing.

Part of the issue is that proprietary software would be running from a company that advertises itself as being oriented around free software.

In addition, it would apply to the enterprise deployments, which is bad because enterprises can’t allow telemetry for regulatory and security reasons.

> Part of the issue is that proprietary software would be running from a company that advertises itself as being oriented around free software.

They're also running on Google's cloud, which is mostly proprietary software. And I can tell you right now they use the most-definitely-closed-source Salesforce for a lot of stuff.

Do you think there's actually a significant amount of people who care about the license type of the internal tools of the companies they use?

None of that is frontend though, which is an important distinction.

Can you explain to me why that's a distinction?

If anything we're talking about frontend JS code that can be blocked by most adblockers/tracker blockers and thus will never touch your data, rather than the uncircumventable "your data is hosted on Google servers".

And the latter does raise alarm bells for a tiny minority, but why would frontend be worse in that aspect?

With frontend code, it’s code running on your machine that you can’t personally audit written by a company you don’t personally trust. For some situations (enterprises) and people (FOSS enthusiasts, paranoiacs), that’s not at all okay.

I'm not claiming that Gitlab is doing anything shady with the data, I'm making a general comment about the notion of "transparency" as a panacea.

I do object to what Gitlab proposed, though, because I object to all telemetry that I don't specifically opt in to.

Do you think that if they had just put the tracker in nobody would have noticed? I think it would go silent for a month, and then become the subject of a top-of-HN blog post about how some enterprise security guy discovered it. Gitlab would write a blog post about how sorry they were, nobody would believe them, and there would be no way to un-do the breach of trust. By allowing everyone on HN to tell at them not to implement the change before doing it, they have made themselves able to avoid the catastrophe by cancelling the plan. Although the best thing Gitlab could have done for their public image as a trustworthy service provider would be to never have even let on that they would consider such a thing, this is the next best option, because whoever wanted to do this now has lost some internal political credibility.

I'm sure people would have noticed. Hell, some may have noticed immediately. But it would have been a non-story unless everyone got an email and had to google around for what's going on and figure out whether they need to be outraged.

If you're transparent about doing some shady shit, you're still doing some shady shit

I think it's a lot more simple than that and has everything to do with the privacy environment today. People in tech really, really, really dislike it when the products/services they use attempt to track them (generally speaking). There is very little trust these days on such matters, the default setting for tech people is to not trust companies that go anywhere near privacy violations, surveillance, tracking, et al. rather than giving the benefit of the doubt. Gitlab should have seen this coming a million kilometers away.

> People in tech really, really, really dislike it when the products/services they use attempt to track them (generally speaking).

This isn't even sort of generally speaking true. Nearly EVERYONE on HN is aware to some extent how incredibly massively they're tracked by their ISP, by Google and Facebook, by their phone, by their phone's manufacturer, by their operating system, by the security cameras in all the shops they visit, by proxy through all their friends who aren't as privacy-conscious as they may be, etc.

I see very few people "hating" this. And I say it as someone who does hate it. Most people here are aware of it, and they all have the "tingly feeling", but most actually do choose to actively ignore it. "I don't have to deal with the idea that my phone is tracking me until there's public outrage abou tit" sort of thing.

And I'm not stubborn enough to put "gitlab wants to know how their product is being used" and "samsung wants to know what I'm watching and when" in the same bucket. And given how much of an improvement gitlab is over the status quo, I also know not to make enemies of my closest allies.

Have you considered that often the reason why we don't complain about the worse alternatives are that we already are avoiding them/mitigating them?

Not always, but often.

Well, when you're transparent about adding features that users don't want, I don't think it comes as a surprise that users express they don't want those features.

Granted, Gitlab wants telemetry in order to understand how Gitlab is being used. I'm sympathetic to their goal. But I'm also sympathetic to a sort of "NIMBY" response from users even if overall telemetry might improve the ecosystem for everyone. It's a sort of "tragedy of the commons" situation.

There's also elements of selection bias as well. If you're highly transparent it's likely that people who value that will be drawn to your product, so when you subsequently take steps to reduce that transparency somewhat (such as roping in a third party from an industry known to engage in shady things), and make it mandatory for users in the most intrusive way possible, it's likely to garner a bigger backlash than for a less transparent product because your user base cares about that sort of thing more.

someone's git hosting service gets more signups because of bad PR by gitlab.

Sorry, but that's an understatement of what sourcehut.org is.

And it's being hugged down after getting to the top.

Unfortunate, I'm sure SourceHut is seeing a nice bump, and it seems like quite a well built product.

Uh, I just signed up today. After gitlab stupid decision sourcehut looks a lot more interesting and not I want to see it in action.

This is kind of a shitty, disrespectful thing to post. Really it's just insulting a competitor. I'm not a fan of GitLab's decision either, but this shouldn't be frontpage HN.

We called out the increasing signups at GitLab when GitHub got acquired, so from our side it is more than fair that Drew is posting this.

BTW As Drew said https://gitlab.com/gitlab-com/www-gitlab-com/merge_requests/... we're pumping the brakes on the telemetry changes.

I think OP was referring to this specifically:

> If my competitors keep making stupid decisions then making SourceHut successful will be pretty easy

He could have posted the same thing with the same effect without being outright rude.

Ah, thanks. I only looked at the first post https://cmpwn.com/@sir/103018233505800721 not the rest of the thread.

Yeah, Gitlab has indeed been known for some shitty PR. Like repeatedly accusing Github of copying features, even though Gitlab as a whole started out as an almost identical clone of Github.

Drew treats his Mastadon as a somewhat personal place and had not intended it to be posting here. So yes, perhaps it shouldn’t be here, but it’s not really insulting a competitor.

See:

https://cmpwn.com/@sir/103018388460813380

https://cmpwn.com/@sir/103018399211308183

https://cmpwn.com/@sir/103018430813804744

Pouncing on weakness is the essence of competition. As long as nothing is being misrepresented, taking full advantage of missteps is practically an obligation.

Sure... although there is also the "don't interrupt your competition when they are making a mistake" angle which would engender less vocal protest

You're joking right? This is being posted on a personal account by one half of Sourcehut against a company that has had over $400 million in funding. Respect has to be earned and Gitlab have lost a lot of respect over this.

What's more rude, the comment from DeVault or gitlab clearly acting against the interests of their users?

I know which one offends me more.

Can somebody give me some context? OP is just a Mastodon post showing signup numbers (for Sourcehut?).

https://news.ycombinator.com/item?id=21346456

I posted it for three reasons:

- directly related to another huge topic

- maybe more people will notice there are alternatives to Twitter

- to send more business in the way of sourcehut (completely unaffiliated, just felt like a nice thing to do)

This reminded me that I've been putting off trying sourcehut. Gotta get on that.

Passive aggressive title, frivolous outrage of the day, Top frontpage seat confirmed.

"Please don't complain that a submission is inappropriate. If a story is spam or off-topic, flag it."

https://news.ycombinator.com/newsguidelines.html

what is the context?

https://news.ycombinator.com/item?id=21346456

What's the context here, exactly?

100 signups a day? sourcehut is already dead, they just don’t know it yet.

Guys, your mobile view really need some care.

Reply on news.ycombinator.com