They didn't know their place, and still don't -- clearly. Dockerhub has very little to offer (basically hanging on by a thread) and doesn't deliver enough value to justify an expense for what was always claimed to be a free service. If you give something away for free, it better be just a starting point, for you to figure out how you're going to add enough value to justify people paying you. And then -- make your case and tell them why.
Docker didn't do any of this. They literally tried to take a free product and just start charging for it. It feels the same as a restaurant who gave free toothpicks away suddenly want to start charging for them. This speaks to how out of touch their senior leadership is, and the only possible thing they can do to recover is to part ways with the leaders who made that terrible business decision.
I've been in meetings like this before -- some analyst runs a SQL query and makes the case for charging for something that has always been free, and hamstrung leadership goes along with it because it sounds like such an easy way to "unlock value". What they really mean, is, easy way to make money off unsuspecting people via something that should probably be illegal, and would be in a different context (i.e. bait and switch).
Docker Hub has little to offer, and yet everyone was up in arms when what little they offered for free gets taken away. Don't you see the contradiction here?
No, because there is no such thing as "everyone complaining". Those are different user groups.
Docker Hub has little to offer to most commercial businesses. When we looked at it, it was hilariously slow and had basically no advantage over any other commercial registry. There was no reason to give our company dollars to Docker instead of Amazon or Google.
Docker Hub is pretty important to open-source projects, as there are very few providers who are ready to distribute large binary files for free without ads or even possibility of them. It was also pretty important to tech companies who use said projects but lack the resources to set up private registry and build docker images they need.
It's the former users who bring money to Docker, and it were latter users who were complaining. There is definitely some overlap but unfortunately (for Docker the company) it is not that big.
Yes and the fact that they don't see the difference between the user groups is symptomatic of the problems that have blighted their attempts to find a workable business model. They offer little to no value to the people who can (and will) pay for things and a lot of value to people who can only structurally "afford" a free service.
In that world charging is never going to yield success. The free-only users are going to find some alternative and in forcing them to do so you have sacrificed the goodwill from the free service in the first place, and the people who could pay won't do so in this case because you literally aren't offering them anything worth paying for.
The contradictions go away a bit when you relent from strawmanning, that second quote is half of a statement which doesn't say what you've clipped it to say.
The guidelines of the site specifically say to respond to the strongest argument, I'd recommend doing that instead of what you're doing.
How can I be making a strawman? I'm directly quoting what you wrote. Here, I can quote the whole sentence (I omitted the second half for brevity):
> Docker Hub has little to offer to most commercial businesses.
> It was also pretty important to tech companies who use said projects but lack the resources to set up private registry and build docker images they need.
So, docker has, at the same time, "little to offer to commercial businesses", but it it's "pretty important for tech companies who use [OSS] projects" but lack resources to not rely on docker hub. I still see the contradiction.
Spare me the reference to the site guidelines. You piggybacked off my comment and made an unrelated point (docker hub is too slow for your business - how is that relevant to my initial comment?). Kettle, pot, all that.
> So, docker has, at the same time, "little to offer to commercial businesses"
That's not what was said, you missed out the word "most", which changes the meaning of the quote:
> Docker Hub has little to offer to most commercial businesses
It can simultaneously be true that:
* Docker Hub has little to offer most businesses
* Docker Hub is valuable to businesses that don't have the resources to set up a private registry.
These can both be true if only a minority of businesses fall into the second category.
I don't think that's exactly true (we'd need to weight businesses by revenue or number of employees for it to maybe be true), but their comment isn't self contradictory.
> You piggybacked off my comment and made an unrelated point (docker hub is too slow for your business - how is that relevant to my initial comment?). Kettle, pot, all that.
I'm not the person you were originally replying to.
It’s a famous old joke, about how people just want to complain. The joke is that these are probably repeat visitors who make the same complaints every time. It’s bad, and we want more.
It has little to offer for paid users. It’s a great place to host your open-source images (you know; the community work that caused Docker to be so popular in the first place).
Forced to pay or leave, the OSS folks will all leave.
> What they really mean, is, easy way to make money off unsuspecting people via something that should probably be illegal, and would be in a different context (i.e. bait and switch).
When a service is free I don't think you can say it's a bait and switch when the one offering can no longer afford giving so much away.
Even for paid services, there is no guarantee prices won't rise once a contract expires, or at the next billing cycle. Now if they had promised free forever then perhaps a case could be made for false advertising.
Should free services be heavily scrutinized as an anti-competitive dirty marketing tactic, then?
If you have deeper pockets than your competitor you could bleed them dry by offering it for free until they're dead and then "coincidentally" not being able to afford keeping it free after that.
I distrust free services as a consumer at this point, but I think it's unrealistic to expect enough people to do that to keep it from being a way of distorting the market and stifling competition.
If it's less of a dirty trick and more of a "shit, we thought we'd figure something out, but we didn't" then I still think there being more scrutiny/regulation around free services could be beneficial purely to push companies away from stupid self-defeating and market-harming business models?
So not a service, rather software surrounded by unaffiliated services? This post is about the official service part of Docker. To my knowledge there is no official Linux service.
Read what I said again. It's not about what you know or not. This kind of precedent could easily be applied to things it shouldn't be, or rather things you don't think it should be.
try explaining all this to someone uninterested or uninformed about technology. A cubicle worker or something.
No, the thread stems from above where it's basically indicated that free things should be illegal if there's things being sold off the back of it.
The precedent things like this would create could easily be extended to other things. Good lawyers could hold things the OP isn't think about to this standard. Especially if theres a jury uninformed about the topic at hand.
> Should free services be heavily scrutinized as an anti-competitive dirty marketing tactic, then?
Perhaps if you offer a free service, you should be legally required to either post a notice "this service may switch to a paid version at any moment with a 30 days warning", or make a public promise "this service will be provided for free until 2030" (or any other year). You are allowed to increase the year later.
In the former case, you could be sued if the notice is deliberately hidden. In the latter case, not providing the free service would be considered a breach of contract.
This is an interesting point. If the alternative to charging for a free service is to cancel it, which should one do? This of course is usually a false dichotomy. An alternative is to set it free for an exchange of goodwill and perceived loss of potential monetary value. But hey, that's what a true friendship and family entails. Professional dishonesty is pernicious. Hand back your PR degrees, people, if you want to make the world a better place.
I find their $60/year plan okay for unlimited private registries, OCI registries, vulnerability scanning, granular permissions with app tokens, and builds. I am fully aware I can do some of that with GitHub, I am also paying for it, it costs me $12/month with a couple of orgs and an additional contributor.
What other alternatives exist? I can do a registry with object storage backend, but I still have to host the registry container/VM itself, and I have no scanning, no ha, no permissions. I could use GitHub for everything maybe but who knows if they don’t jack up their prices? I’m also not confident about putting all my eggs in that basket. I could use quay, but that ain’t cheaper…
This assumes one knows how scalable such a generous free offering will be forever, on day one. Much more likely they were basically dumping to undercut the competition and gain market share.
Though I suppose some startups are genuinely naive enough to just launch and hope they'll find a revenue stream rich enough later.
I'm not saying they're right or wrong, but they believe so. You need to count this when writing down your monetization plan. People nowadays are generally extremely entitled.
I’ve run a service that pissed off free users by charging for things that used to be free.
It increased revenue and decreased costs from the tiny percentage that actually left. I think it’s win/win for any business that doesn’t rely on network effects.
Part of Docker's issue was the executed it absolutely horribly. They gave a short suspense, very little detail, and few options for migrating under such a deadline.
In my personal experience? No. I asked for 15%, knowing my co-workers made more than that, and inflation had been 9.1% YoY. I was told to be happy to get 3%.
What did I do? Found a different job and employer. Jobs are like insurance rates: you should shop around occasionally. For businesses it's a cost of doing business when trying to compete to lure new workers.
I still don't understand what the supposed value add of Docker Desktop really was for developers who just needed to run Linux Docker containers (I know there was some value for those needing Windows or Mac containers probably). A nice UI?
At my company, we all had to uninstall Docker Desktop while they sorted out licensing and billing or whatever.
While that was happening, everyone just switched to using Docker in WSL, almost overnight. For me the setup involved changing some settings in my IDE so that it used WSL, it was one drop-down in IntelliJ. I also used VS Code and that was even easier.
The whole thing was very puzzling for me and thousands of other engineers. Many business decisions Docker makes are puzzling. But it seems to be working out for them.
What I mean is that their drive to monetize has actually produced a lot of revenue, i.e. was a success. People were predicting some kind of mass migration away from the now paid Docker products and it didn't happen.
The pressure to monetize further is just the usual endless growth stuff, it doesn't mean their previous monetization efforts actually went badly.
There’s nothing inherently wrong about an initial free period for a product. In many cases it’s a rational launch strategy.
On the other hand, it’s both irrational and incredibly entitled to expect that anything anyone ever gives you for free will be available for free in unlimited quantities, in perpetuity.
My formative experience of their backwards compatibility was being unable to open old word docs in newer versions of word. Open office could. Go figure.
As for how long their tech lasts, I used XNA for a game jam in 2012, and had an alright experience. I could no longer download and use XNA by the time the 2013 game jam came around.
> My formative experience of their backwards compatibility was being unable to open old word docs in newer versions of word.
From experience: Word's still dropping support for older documents, and LibreOffice still supports them. (The only LO incompatibility I know is missing Math formulae sometimes, and that might have been fixed; if not, it's a fairly straightforward code change, since OOXML Math tags contain textboxes.)
However, their websites have not held to their software standards. Links break all the time in my experience, especially for older documentation (msdn, etc).
I dont get this argument. Docker didnt break backwards compatibility. They literally just wanted big projects to pay for the storage and bandwidth they were using.
They said: "You have 30 days to move away or pay. If you don't do anything, you will lose access to your published images, but we won't remove them. So they will rot on our registry and confuse your users."
At least that's what I understood. And it really sucked. I would even understand if they did not want to pay for open source images. But 1) give me more than 30 days 2) either remove my images or leave me in control to remove them later, not something in between.
It's a really bad move to give something away for free for years, and then suddenly announce you will charge for it and everyone has 30 days to figure it out.
Over a shorter time scale that would be called "bait and switch" and it would be illegal.
Over a longer time scale it's still a massive dick move and destroys customer trust.
> Over a shorter time scale that would be called "bait and switch" and it would be illegal.
No, it wouldn't. It would be "bait and switch" if you had committed to use Docker Hub for the next year, and they announced after two months that now you have to pay for it and you can't back out. If I offer you a piece of gum for free today, and tomorrow you come back and I say either pay $1 or you don't get my gum, have I committed any kind of fraud or whatever?
> Over a longer time scale it's still a massive dick move and destroys customer trust.
The bait and switch argument is so disingenuous. As if they don’t have the right to make money out of their product. Payment models change with the world and economy.
when most customers trust the tech giants like google, payment processors and such, they have already been taught not to apply scrutiny to whom to trust
If all they wanted was the whales to pay up, why did they not target the big projects, using some objective metric like pull count or actual bandwidth consumption rather than literally everybody? Team projects are often used for namespacing reasons and teams of one are not uncommon.
>The company with the best track record of maintaining backwards compatibility in the entire tech industry owns Github.
They also own TS, which seems to have major breaking changes each year and necessitate annual churn and refactoring of entire codebases. Windows having remarkable backwards compatibility does not mean the rest of their products do, too.
EDIT: I could also bring up one of their other products, Lean, whose 4.x versions have so much breakage that mathlib has been on a community-maintained fork of Lean 3 for as long as I can remember.
Backwards compatibility buys enterprise licenses. These are for profit corporations - even their free tier has you somewhere in a marketing funnel or a revenue structure. Today they have a “good” CEO, tomorrow it could be another Steve Ballmer in charge.
The bottom line is, one shouldn’t rely on a single corp for something that seems like “critical infrastructure”.
For me it’s because they are very careful not to screw over customers that rely on certain behavior.
Having said that, Microsoft does sometimes break backwards compatibility.
I sell a Windows app (a battery meter) that cannot work on Windows 11 because the new Taskbar doesn’t expose an interface for plugins to render data on the taskbar (IDeskBand is the old interface).
So I can display a small window to the user but it’s far less convenient that showing it persistently in the taskbar.
There's nothing wrong with no longer wanting to distribute a 25 year old application that you no longer support development tools for. However your comment is completely disingenuous. One of the first links I ran into was a massive support article from Microsoft about how they will still allow VB6 applications to run on Windows and gave detailed instructions/tools on upgrading a Visual Basic project so you can continue development and don't lose all the work.
> While the Visual Basic 6.0 IDE is no longer supported, Microsoft's goal is that Visual Basic 6.0 applications continue to run on supported Windows versions. The resources available from this page should help you as you maintain existing applications, and as you migrate your functionality to .NET.[0]
This is much better support than other companies provide for discontinued platforms. Do you personally want to be on the hook for a version of your software you wrote/released 25 years ago? Especially if you've released what you consider to be improved upgrades of the software? Would you even be willing to help users upgrade from your 25 year old software? If the answer to any of those is no, then you're providing a very skewed perspective on what reasonable support for a product looks like. And once again, Microsoft is still writing articles 20 years after the initial release to help users upgrade. For all the bad crap Microsoft does, completely abandoning widely used frameworks is not one of them.
I repeat. Their backward compatibility isn't amazing. It may be good but let's stop pretending Microsoft deserves special kudos for what's actually a mixed bag in terms of backward compatibility. And look at the sibling comment to mine about bad backward compatibility with data, in the case of Microsoft Word.
Why would you want to buy the VB6 IDE? If you’ve got legacy VB6 stuff to maintain, you’ve probably got a 20-year-old license for the IDE already. If you somehow do not, you can buy a Visual Studio subscription [0] and get it. Supporting something does not necessarily mean they sell new licences.
well, if it's so great that you can run old apps, what if one of those old apps you run, you want to make a tiny change to it?
but there's gonna be security issues.
still it shows that Microsoft's legendary (albeit narrow) backward compatibility doesn't really mean a whole lot, except for sounding cool
> you’ve probably got a 20-year-old license for the IDE already
As for this, no, someone could pick up an old app written by someone else, or lost their copy of the license, or maybe the license belonged to a company they no longer work for. Also it apparently doesn't apply to the person asking the question...
You can get a copy of the VB6 IDE with a Visual Studio subscription if you really need it. It works on Windows 11 — you can search for "visual basic 6 windows 11" on YouTube to see it in action (I’m not linking to any videos, since those videos tend to suggest less legal ways to acquire software).
I don’t understand your problem with the tool being available in a VS subscription.
> Because of Microsoft. Microsoft is being unhelpful here as usual
Windows 11 can be bought in many places, including microsoft.com. All you need is a spare $140 or so. People pirate it anyway. Is Microsoft being unhelpful by expecting payment for software?
Whole point is that Dockerhub dared to ask for money instead of hosting your stuff for free and paying for the bandwidth.
The reason why Github is free for normal users it that Microsoft extracts value out of the code you host there, by at least training their LLM. Same will apply for your docker containers.
This is part of the Embracing open source.
Now they have already started selling Copilot and with CopilotX bringing new AI assisted stuff to code reviews and pull requests even more people will migrate over.
This is part of the Extending.
Then after they have everyone they can get to use Github and they have extracted all value from your code they will move to the last step and ask you to pay for hosting your stuff and while git is free to use and you can setup your own servers the general UX will be so much better that even hobbyist will pay.
GitHub has always been free for normal users, even before MS bought them out. It is a pretty classic loss leader to get developers so comfortable with it in their personal work that they push their employers to pay for organization accounts.
Neither your extend nor extinguish meet the classic definition of either. They aren't changing the spec of git. Having a nice UX doesn't count as extinguish. Literally anyone can set up gitea or another git server. They might not get the cool toys like copilot, but that isn't git.
This is not EEE, this is watering down EEE to look not nearly as terrible as it actually was back in the day. If anything, it looks like a shill for MS trying to get people to forget what EEE actually was.
> Whole point is that Dockerhub dared to ask for money instead of hosting your stuff for free and paying for the bandwidth.
I would be fine with that. What I hated it that they gave me 30 days, and not doing anything seemed to result in my images staying on their registry, but out of my control. So that my images would rot there but still be available, and I would have to tell my users that I moved to another registry which is not Docker? What the hell was that?
> The reason why Github is free for normal users it that Microsoft extracts value out of the code you host there, by at least training their LLM. Same will apply for your docker containers.
This was true when Copilot only used the CODEX model, which was only trained on Github code.
But now that, they are using/going to use GPT-4, which you bet includes all code sources possible, GitLab, Codeberg, probably even your private gitea service!, it really doesn't matter anymore, they are going to train on your public code no matter what.
The only thing left to stop it is the U.S. Supreme court lawsuit, it will probably decide the Fate of modern generative A.I.
Or they could make it so if someone pays $30/mo they can get a notification every time someone visits their GitHub. This is a reality for LinkedIn which is also owned by Microsoft.
Don't let Microsoft ever get away with saying they respect privacy...
My big concern wasn't where my projects are going, it's other projects I use or build upon. With absolutely zero indication of who was coming and who was going, it was just .. wait to see what pipelines explode?
Docker, Inc. is in a tricky position. They've raised a lot of capital, and continue to raise capital. According to Wikipedia, they took $105 million from Bain Capital in November of last year. So Docker, Inc has Mirantis on one side, and a whole bunch of VC on the other side.
I'm sure that Docker Hub is expensive to run, and the core Docker app isn't free to maintain either.
This TOS change was a fiasco, and absolutely didn't succeed in doing anything other than losing a lot of community goodwill. But I do understand why they're trying to squeeze some money out of Docker Hub. I'd be surprised if it's even break-even for them right now, and VC funds aren't exactly known for their acts of charity.
I get what you are saying, but I gave little sympathy for their situation - they used the VC tactic of "give it away to gain and monopolise the market, then do the old switcheroo once we're #1", so they had to know what was coming.
Not really what happened — they planned to build a cloud platform on top of it and probably subsidize Docker Hub using those profits. But they got their lunch ate by Kubernetes.
Which is crappy, because while I use k8s, it's actually poorly written, half baked and kinda non-portable. The fact that it depends on ELB when you use it at Amazon (just to extract more money), is super sad.
Docker Swarm was a pure technology that didn't have such shitty half thought-out dependencies.
Personal thoughts and feelings about k8s aside, please don't conflate the requirements of EKS, Amazon's managed k8s offering of which they are fit to do what they do best and tightly knit their own products to come up with a solution, with k8s as a whole. You're perfectly welcome to deploy your own k8s with static IPs for ingress to nginx if you'd like on AWS, as you can anywhere.
I wonder how Docker survives after this, this action now rolled back was clearly a cost cutting exercise. Im still surprised Google hasnt acquired them.
Maybe: Push more Docker users toward Kubernetes -> Push the concept that Kubernetes apps are easiest to run via GKE -> Google Cloud profit?
Also: the default location of Docker images would become Google's registry -> drive more businesses to pay for Google's container registry. (It fulfills more business use cases than Docker's, so it should be easier to sell than Docker Hub.)
With these things I always wonder how many people they ignored to push this through. I am pretty sure there are plenty of people at Docker who 100% accurately predicted this, spoke up about it, and got ignored.
I still refuse to use Dockerhub and I cannot wait for Podman (and building multiarch containers) on Mac to get better so I can absolutely ditch anything Docker Inc or anything Mirantis touches all together. These are two companies I will NEVER trust again and will use whatever power I have to convince others to do the same.
Same! Can't wait to just get rid of it completely. Red Hat also hosted Container Plumbing Days this week and was pretty good! Looking forward to what they will do with Podman (specially on macOS)
Probably not better in so far as they're also trying to make a buck.
But IBM has pretty deep pockets - the kind where they could feasibly write off the bandwidth of a container image registry as a goodwill gesture and not bat an eye. Same with GitHub/Microsoft.
From recently history, the same cannot be said about Docker the company.
Experienced sell-support-for-open-source business conservatively funneling enterprises into its support network vs increasingly desperate VC money young-ish startup doing things without being able to predict consequences.
Lots of replies related to image hosting. We still base our containers off the public images provided by docker hub, but our customized images are stored in a private Azure container registry. This is more due to the high security of our environment than anything else. We cannot deploy from external sources.
This was about public images for free docker organizations (teams), used by many open source projects for branding / naming consistency.
I maintain one of the projects which moved to ghcr.io and have no interest in moving back, only further away. They have shown their hand and can no longer be trusted as good stewards or ecosystem participants imo
You do realize that you moving away is actually a win for Dockerhub, right?
Whole point of removing the free plan wasn't to squeeze money out from you, but to make you pay for the free storage and bandwidth. If you don't take up that free storage and bandwidth that is literally the same as removing your free plan.
I don't know how podman does container storing, but if everyone moves over you probably shouldn't expect them to have free tier forever.
EDIT: I would love to hear your counter arguments since you are downvoting. Why am I wrong?
You've fallen into the trap of an overly simplistic and narrow view.
When projects and people move away from DockerHub, it hurts Docker-the-company because it further decreases their already tenuous and evaporating relevance.
Reducing a fraction of storage and bandwidth costs isn't important compared to retaining mindshare and the perception of relevance.
Edit: @nextlevelwizard The explanation you requested has been provided. Now it seems you're disagreeing with a common, widely accepted phenomenon. I was hoping for a more collaborative interaction.
My team has certainly accelerated the move away from Docker Hub towards our internal registries. And I don't see a reason to stop now. Just one fewer unreliable vendor for us to care about.
But you are still using docker. My company has always hosted our own repository for containers, that wasn’t the argument being made - it was that people and companies would move away from docker
Companies will leave if they cannot trust Docker Hub to keep their CI systems working. Scaring decision makers with technical issues and instability is a pretty big nono.
We are pulling from Docker Hub, we're nobody special or big, but still. This shenanigan made us look for alternatives and consider the value the service provides - which didn't end in Docker Hubs favor.
When the community becomes frustrated, users start to leave the platform. Among these departing users are influential decision-makers who may lose trust in the platform and refrain from using it in the future. Additionally, some individuals publicly express their dissatisfaction and reasons for leaving. This creates a negative feedback loop, ultimately leading to the platform's decline.
To be honest I'm not sure which move was worse: They announcing they would be sunsetting, or they going back on the decision after the backlash. I feel like the trust was already lost after the initial decision; there's no going back after that.
You can not be the main player and betray the trust of your users (free or not).
It's easy to forget the value of being an important part of a community that makes decisions on millions of $$$ in software spending, but be sure that there are plenty of companies who would love to take your place.
Reverting this incredibly dumb decision is the right course of action but there is a long way ahead to restore trust.
I'm not saying they have the obligation to keep the free tier, but don't expect people to stick around.
Also free plans can definitely last, best example being Github. If Github would cut off free open source repos it would be over, and I'm pretty sure they are very aware of that.
My point is, it's a mistake to assume people will pay for something that was free and it's a misunderstanding to think there is no symbiotic relationship. Hosting a giant community can be incredibly valuable, especially when it's tech people with big budgets in your sector.
It's imo a business mistake to monetize in a way that goes against that. The community will have no issue finding a new home that does appreciate them (e.g. Github).
Here is how you stop free service without betraying trust:
First, lock creation of new free accounts. Then announce that:
- After a long time (at least a year), newer images won't be accepted on free plan and after even longer time, existing images would be deleted.
- After a long time (at least a year), free plan images would be deleted, starting with least used first.
- Downloads from free accounts are heavily throttled. Speed becomes even slower the more stuff that IP downloads. The full speed is restored if either account owner or downloader have paid accounts
- In a few months, accounts get a quota, and you only get 10 downloads/hour for free. Everyone else gets an error message and have to try again in an hour.
... see? So many way to wind it down without having everyone freak out. Docker has really flubbed that decision.
This strikes me as quite an entitled attitude. Who are you to dictate how someone stops giving away services offered at their own expense?
Regardless of how long they've done it, nothing can be free forever. People who build their own companies or products based on the assumption things can be free forever set themselves up for disappointment. And it's not the fault of the providers.
Nice. The issue I have with some of my multi-arch stuff is that I began doing it before buildx stabilised, and I also launched multiple runners in parallel because some of the builds take... a long time.
I get that images costs money to host, but they could instead add a new feature that would spread out the costs (~bittorrent), while the hub would be the seed of last resort.
What if it changes tomorrow? I'm moving even personal projects to a private repo and gonna build the base image or two myself, just in case someone's investor wants to step on the gas and milk me some moneys.
Having some personal infra is not a big deal. It's actually getting easier, if not legshooting with overcomplexity.
You might want to add that that's the author of the previous post calling attention to this policy (https://news.ycombinator.com/item?id=35166317), because without knowing that information your post reads like a spam comment that has nothing to do with this story.
There's no way out of it. Investors want their investment back and docker at this point is just a registry that otherwise anyone can deploy backed with some object storage or use any of other dozens out there.
No other big player willing or interested to buy it either.
When this all went down I was looking to move to some type of proxy for their hub, like I do for other tools (aka: mirror). My mirrors for CPAN, NPM, Packagist, Go and Debian were all (mostly) trivial to configure and use. Docker not so much.
Still working to figure out how to route both search and pull to use our mirror.
That it's hard to mirror feels like it's even more important to mirror.
A rug pull never happens at a good time and makes it's occurrence more expensive than a planned event.
Can someone explain this to me since I don't really use docker so all this feels weird to me? It would appear the entity provides oft-used service. Did they overstep on price or the suddenness of the announcement or some other combination I am not seeing?
From my perspective a large part of the issue is that they were trying to charge the wrong people. What they needed to do was figure out some way to make the people using dockerhub to build things to pay. What they attempted to do was charge people making it possible for other people to build things off of dockerhub.
This meant that a lot of projects providing free services to docker were going to walk. For example why would the curl project upload curl images if it was going to cost them money? The people benefiting from those uploads were the people downloading the images, dockers customers (some already paying), not the curl project. (Curl is famous enough that docker wasn't going to charge them, but there are lots of less famous similar projects that they were trying to charge).
Moreover these were frequently open source projects, which are very cost adverse by nature, and who you want to be seen as supporting not exploiting.
What docker really needs to figure out is a way to charge companies downloading images, not people uploading them. Given how their service works, I have no clue how they manage that.
My issue with it, for open source projects, was that it was not really clear what exactly would happen, and then my understanding of what would happen did not please me at all.
My understanding was that I had 30 days to migrate my projects away (or pay, but for many open source projects there is no money). If I did nothing, I would lose access to my images on their registry, but the image would stay (and rot) there, which is even worse than the image being removed.
Docker announced they would end the free "teams" tier and that people would need to migrate to a paid plan to retain their teams. The communication was rocky and non-specific as to the details of the impact and only gave a 30-day notice.
They quickly (next day or two?) followed up to give more specifics as to what that means (and doesn't mean).
Now a couple of days after that they are pulling back on it completely due to backlash.
Probably they could have gotten away with extending the time-frame... I'm not sure why anyone would think 30 days would be ok, but I suppose it is the free tier.
If you weren't planning on switching because of the for-cost team plan, you really should switch now. If Docker can't charge for their software at all without having so many people migrate away that they have to reverse the decision, then they are not a viable commercial entity and will randomly disappear in 1 or 3 or 5 years, probably when it is very inconvenient and you haven't prepared for it.
Haha! Victory! Suddenly they realize that people will go their own way rather then be strong armed, all their work writing themselves in as the center of the docker universe was about to go poof. I’m pretty sure it’s their destiny to be usurped but they’ve bought some time to become useful or find a way to lock people in to their platform.
Now is the time for everyone to find an alternate means to share their docker images.
If they want to make money, they should stop listening to the community and start asking themselves "What would Oracle do?"
I think a lot of companies would not really be capable to setup an alternative in time and would just have to start paying instead. Which would not be a bad thing. The opportunity cost of switching might be higher than just paying Docker for their services.
Oracle would've charged per core and per GB of ram and number of containers to license your docker install.
Free would allow 3 containers, on single core and 512MB ram.
Sales people would be calling everyday to upgrade to docker pro, which is $120 per month/per user that will allow you to self host your registry, on premises or Managed Oracle cloud (vms and bandwidth extra).
Ran docker run without a valid license? Prepare to be sued.
I do think this is a smart business move. You don't want to trade being "the place" for Docker for being "one, slightly more profitable place among many."
MBA CEO: We beat earnings estimates this quarter by more than 100 golden eggs, via our sale of the goose that lays golden eggs.
Makes me laugh! So many issues (not just this) because no one wants to pay for an FTP server and write some FTP commands, secure of course. And the VC companies make something say a tiny but more
convenient but then pull the rug.
Docker is awesome tech but a terrible company with a terrible business model. I don’t see them surviving long term. Especially with the way they operate and alienate people.
I'm curious, are they profitable? Docker is a super popular tool and I thought whoever has been there from the beginning should have 7 digits in their accounts.
maybe they can use the expenses maintaining docker images of projects that are beneficial to a broader FOSS user community to pay less taxes or something, or brag about their phylantropy on accounting reports.
I don't believe the revenue from these avenues would save or kill the company anyway, so better make a lemonade of these lemons, right?
Read: We won't do it today or next week, since we got a bunch of negative publicity, but don't actually address why they didn't catch that it was a terrible policy.
They describe that they have come to the light and now realize the policy was bad, as well as the communication around it, but don't describe how they are actually going to address this fundamental misalignment with their userbase and out-of touch decision making.
It's incredibly tiresome that the first post on HN whenever a company does a press release is going to be from someone who will never be happy no matter what was said.
"What does take responsibility mean?" "This is just PR fluff". "They are just saving face." Criticism of good things because the poster didn't like the motivation. Stuff like that.
Did you really think the company was going to bare their soul to you like you were their psychiatrist or their priest? Do you think you are even entitled to that?
Commenting about whether or not this is a good decision is valuable. Complaining about the content of PR fluff is not.
It’s no more or less tiresome than a press release is itself. Whatever the material contents, a press release has “positivity” set to 1.0 (or as close to 1.0 as permitted by law and professional standards). This upsets the balance of the world, and that balance can be restored by submitting a reply with “positivity” set to 0.0 (or as close to 0.0 as permitted by law and standards of etiquette).
Analogous to “someone who will never be happy no matter what was said”, the tone of the press release will always be happy no matter what they have to say.
It seems very natural - almost inevitable, even - that someone will reply to an artificially max-positive document with an artificially max-negative comment.
(Personally, I made peace with the tiresome nature of both press releases and critical replies by viewing it as an informal application of Laplace’s Rule - add one success and one failure to the total count to more accurately estimate the real underlying probability. The press release presents one observation guaranteed to be as positive as possible, and the inevitable critical reply presents another observation guaranteed to be as negative as possible, so with those two in hand you can look at other observations and get a better final estimation.)
> Analogous to “someone who will never be happy no matter what was said”, the tone of the press release will always be happy no matter what they have to say.
So grow your own filter. But as another reply said, this press release still has valuable information remaining after having been filtered. Complaints about their press release reading like a press release do not.
All this being said, this was actually a pretty insightful conversation for me, not having considered the complexity around press releases like this, but still being uneasy about them but not sure why.
Without making any commitments that they won't at some point in the future undo that decision, that they see the importance of why there are free accounts, or what they are doing to ensure that community feedback is incorporated into their plans before they make decisions.
Sure, they changed course on this single decision, but they haven't addressed the situation that allowed this to even take place, nor does it look like they are planning on it.
Let's be clear here. Even if a company commits to something in a public statement, it's no guarantee.
Even if the writer of the post explains lessons learned and policy changes they're making, it's no guarantee. The people who learned the lessons can leave the company. Policies can be changed at a moment's notice or simply never be enforced.
The best anyone will get is a contract that can be legally enforced, and that won't ever happen for free users.
(I edited my comment as you were replying, sorry about that.)
If the reflexive maximally-negative complaints are contradicting the actual information presented, they had better come with strong evidence or I’m dismissing them (and I do find those comments more tiresome than press releases or the other negative replies; they come across as conspiratorial crank-cases with an axe to grind, even bordering on spam to my mind).
The comments I had in mind are “the most negative interpretation of the information that comports with the facts”, much like the non-information components of the press release are the most positive interpretation of the information that comports with the facts.
A press release contains info and fluff. The opposite of soft fluff is hard edges, so perhaps a cute and pithy summary of my view is that no fluffy PR statement is complete without an edgy reply.
We have been conditioned to think it's acceptable, because it's now normal, to live in a society where we're constantly deceived and lied to. Everything is marketing, PR, spin.
From all the corporations that control so much of our lives, the government, science, medicine, religion. Every major institution in our society. Every message has to be carefully parsed and squinted at, one always has to deduce what might have been the (truthful, real, honest) conversation behind closed doors that produced what we're hearing or reading.
From every message it is possible to know only one thing: what the author wants us to think. What is real is almost always another story.
PR exists because honesty and frankness are punished in our society and people will jump at any opportunity to prosecute or sue. The result is PR fluff, endless disclaimers, and announcements with as little detail of the internal workings of a decision as possible which need to pass through review by legal and finance teams.
But we all know that, right? We don't need to pretend like we don't, or act like we are offended by it, as if it's the first press release we've ever seen.
You’re not wrong. But I’d add: benefits are always motivations. It’s easier to deceive, which is what PR is, but it’s also profitable. We can’t blame people for complaining that they’re being deceived just because it’s common.
thats not true, honesty is valued, but what PR does is attempt to lie and spin things in a way that misleads a stupid audience into thinking that something else is happening, and this can be beneficial.
this is why lying exists, because the ones doing the lying thinks it benefits them over telling the truth.
These decisions should never have seen the light of day, let alone a policy that was communicated in the way they did.
Why should we cheer on PR and product folks trying to break their fucking arms patting themselves on the back for walking back an incredibly disruptive, short sighted and user-hostile decision?
Because A) of all the permutations of follow up decisions that could have been made, this falls on the "good choice" side, and B) this isn't "pat on the back" situation, this is a save face situation.
The CEO could take a 100% paycut, relinquish all of his shares, take 20 lashes on the back, and fire himself yet the top comments will always be about "oh so this is taking responsibility huh, how about not screwing it up in the first place???"
> The CEO could take a 100% paycut, relinquish all of his shares, take 20 lashes on the back, and fire himself yet the top comments will always be about "oh so this is taking responsibility huh, how about not screwing it up in the first place???"
Interesting theory, but it doesn't match my experience - the only time I've ever seen a thread about a (japanese) CEO doing those things the comments were closer to "why can't our (USA) business leaders behave like this".
The original announcement was such that no followup would make people happy. It's not fair that you can mess up badly enough that even with perfect play, total recovery is not possible. Life is not fair.
If you want to hear the positives of something hang out with business people. If you want to hear a bunch of bitching about how literally everything is not good enough, hang out with engineers. We don’t get paid to shoot rainbows up everyone’s ass, we get paid to point out what’s broken and fix it.
> We don’t get paid to shoot rainbows up everyone’s ass, we get paid to point out what’s broken and fix it.
(I don’t disagree with your point at all but …) I’m reasonably sure many of us get paid to do both, sometimes simultaneously to satisfy various undue optimisms, and other times encapsulating our pessimisms from certain audiences.
Parent poster was expressing distrust. I get it. There's emotional turbulence from flip-flopping on a important decision within a week.
The tiresome aspect is also putting his own words into the CEO's mouth, to make his point. Even though parent poster is probably somewhat right and we are feeling similarly, the cynicism shines through.
Don't understand your parent comment as something insightful or with good information density. Its function is cultural. We need reminders from time to time that corporations default to evil if not left unchecked. We need rally points to assure us that we are not alone in condemning corporate greed and ruthlessness, in fighting against neo-liberalism. If you just accept the status quo without any form of pushback, you risk leaving the fine line of realism and stray into complicity.
And you are right, nothing that they would realistically have said would have satisfied me. But that is because they have set themselves up for failure long before. Quoting from GordonS above [1]:
> I get what you are saying, but I gave little sympathy for their situation - they used the VC tactic of "give it away to gain and monopolise the market, then do the old switcheroo once we're #1", so they had to know what was coming.
At least we know that we're on borrowed time and maybe the wider ecosystem can come up with a decent (distributed, even?) alternative instead of everybody panicking and just moving toward another rug-pull.
Yup, this should be addressed with the same level of detail and transparency as a postmortem for a technical failure.
What were the multiple points of failure that led to this decision and how are they altering their operations from preventing something like this from happening again?
As soon as one company gets away with leveraging namespaces to extort money the flood gates will open and everyone is going to have their hand out asking you to pay just to maintain a consistent identity across platforms.
It's time to make sure you've got trademarks on your namespaces IMO (if possible). At least that gives the upside of being able to make legal demands to companies that don't do enough to prevent impersonation and namespace squatting.
No company will ever say they will never do something in the future like that because the future is unknowable and who knows what circumstances will come up.
What we can hope for is they learned something from this so when a hard decision like this must be made they’ll use a different approach.
100% this. If this wasn’t Docker’s first time making a bad decision maybe I would think that this decision is on hold indefinitely.
But it isn’t and they don’t deserve the benefit of the doubt. I will be continuing with my original plan of communicating ideally not using DockerHub internally and making every effort remove our reliance off of it.
>After listening to feedback and consulting our community, it’s clear that we made the wrong decision in sunsetting our Free Team plan. Last week we felt our communications were terrible but our policy was sound. It’s now clear that both the communications and the policy were wrong, so we’re reversing course and no longer sunsetting the Free Team plan
Seriously, what else do you want? This kind of pointless whinging is just noise. They flat out said they were wrong in the first place.
No business should ever build their business on the free tier of some other business. Ever. The change in position in this blog post does not change that. The change that precipitated this blog post also does not change that.
I self-host 100% of my dev environment and tools. I don't care too much about getting services from Docker Hub, regardless of whether it's free or paid. However, I do care about my namespaces and I don't want someone acting as if they're me, even if it's innocent and unintentional.
Because of that I feel like I'm forced to deal with services like Docker Hub. It's a net negative for me to begin with, because I have to take the time to register an account, but it gets really awful if everyone starts asking for money to simply own a namespace.
On the Docker Hub side, there can't be much cost in letting me hold my namespace and, because I've got my preferred namespace there, it makes it low friction to check them out as the first option if I ever need reliable image hosting that scales. There's some reciprocal benefit IMO.
In principal I should probably be paying them something because all of my self-hosted stuff runs on Docker.
True, but I always use a team or an organization if it’s an option because it feels more correct to me if the namespace is transferable. If a dedicated user account is the only option I’ll take it. Ex: Docker offered to convert namespaces to normal accounts which I would have been happy with.
Docker didn't do any of this. They literally tried to take a free product and just start charging for it. It feels the same as a restaurant who gave free toothpicks away suddenly want to start charging for them. This speaks to how out of touch their senior leadership is, and the only possible thing they can do to recover is to part ways with the leaders who made that terrible business decision.
I've been in meetings like this before -- some analyst runs a SQL query and makes the case for charging for something that has always been free, and hamstrung leadership goes along with it because it sounds like such an easy way to "unlock value". What they really mean, is, easy way to make money off unsuspecting people via something that should probably be illegal, and would be in a different context (i.e. bait and switch).