> a rugpull is certainly a shitty experience for them
It would certainly be a shitty experience, if there actually was a rugpull, which there was not. People who were using the version of Datastar that had all those features are still free to keep using that version. No one is taking it away. No rug was pulled.
> a contract is established between developers and potential users
Sorry, but no. The license makes this quite clear–every open source license in the world very explicitly says 'NO WARRANTY' in very big letters. 'No warranty' means 'no expectations'. Please, don't be one of those people who try to peer-pressure open source developers into providing free software support. Don't be one of the people who says that 'exposure' is a kind of payment. I can't put food on my table with 'exposure'. If you think 'exposure' by itself can be monetized, I'm sorry but you are not being realistic. Go and actually work on monetizing an open source project before you make these kinds of claims.
> why even bother releasing software?
Much research and study is not useful for many people. Why even bother doing research and development? Because there are some who might find it useful and convert it into something that works for themselves. Open source software is a gift. The giving of the gift does not place obligations on the giver. If you give someone a sweater, are you expected to keep patching it whenever it develops holes?
> If a project is not maintained anymore, it can be archived, or maintenance picked up by someone else.
Then why can't it be maintained by someone else in the case of using the old free version?
> A web framework, no matter how simple, will break eventually, most likely in a matter of months.
Sure, the ones that depend on a huge npm transitive dependency cone can. But libraries or frameworks like htmx and Datastar are not like that, they are single <script> files that you include directly in your HTML. There is no endless treadmill of npm packages that get obsoleted or have security advisories all the time.
> You expect people who want to build a web site and move on with their lives to dig into a foreign code base, and fix the web framework?
Well...ultimately, if I use some open source software, I am actually responsible for it. Especially if it's for a commercial use case. I can't just leech off the free work of others to fix or maintain the software to my needs. I need to either fix my own issues or pay someone to do it. If the upstream project happens to do it for me, I'm in luck. But that's all it is. There is ultimately no expectation that open source maintainers will support me for free, perpetually, when I use their software.
> A few people (that we know of) were directly impacted by this
What impact? One guy blogged that just because there are some paid features, it automatically kills the whole project for him. There's no clear articulation of why exactly he needs those exact paid features. Everything else we've seen in this thread is pile-ons.
> Doubling down on this, ignoring and dismissing such feedback as "performative"
Aren't you doing the same thing? You have been ignoring and dismissing the feedback that this is actually not that big of a deal. Why do you think that your opinion carries more weight than that of the actual maintainers and users of the project?
The open core part of the project was removed from NPM. Available only on GitHub.
There are no published plugins from the community, nor is there a repo where the community could have collaborated on OSS adding/plugins.
Are people being entitled expecting it ? Yes.
Is there something stopping people from taking up this work and creating a repo ? No.
But it is illustrative of the attitude of the owners.
The point is not to accuse of rug pull but how confident is the community in taking a dependency on such a project. The fact that the lead dev had to write an article responding to misunderstandings is in response to what the community feels about this.
The argument on their discord for licensing for professional teams 'contact us for pricing' goes like it depends on the number of employees in the company including non-tech folks.
Our community is fine, all this is coming from people so far that haven't actually used Datastar on project at any kind of scale. If this is not the case please show your code and how it effected you directly. Otherwise, it's false outrage that I and the core team care zero about.
> People who were using the version of Datastar that had all those features are still free to keep using that version.
Why are you ignoring my previous comment that contradicts this opinion?
> No one is taking it away. No rug was pulled.
When Redis changed licenses to SSPL/RSAL, users were also free to continue using the BSD-licensed version. Was that not a rug pull?
In practice, it doesn't matter whether the entire project was relicensed, or if parts of it were paywalled. Users were depending on a piece of software one day, and the next they were forced to abide by new terms if they want to continue receiving updates to it. That's the very definition of a rug pull. Of course nobody is claiming that developers physically took the software people were using away—that's ridiculous.
> Sorry, but no. The license makes this quite clear
My argument was beyond any legal licensing terms. It's about not being an asshole to your users.
> I can't put food on my table with 'exposure'.
That wasn't the core of my argument, but you sure can. Any public deed builds a brand and reputation, which in turn can lead to financial opportunities. I'm not saying the act of publishing OSS is enough to "put food on your table", but it can be monetized in many ways.
> Open source software is a gift. The giving of the gift does not place obligations on the giver. If you give someone a sweater, are you expected to keep patching it whenever it develops holes?
Jesus. There's so many things wrong with these statements, that I don't know where to start...
OSS is most certainly not a "gift". What a ridiculous thing to say. It's a philosophy and approach of making computers accessible and friendly to use for everyone. It's about building meaningful relationships between people in ways that we can all collectivelly build a better future for everyone.
Seeing OSS as a plain transaction, where users should have absolutely no expectations beyond arbitrary license terms, is no better than publishing proprietary software. Using it to promote your brand while ignoring your users is a corruption of this philosophy.
> Then why can't it be maintained by someone else in the case of using the old free version?
I addressed this in my previous comment.
> Sure, the ones that depend on a huge npm transitive dependency cone can. But libraries or frameworks like htmx and Datastar are not like that
Eh, no. Libraries with less dependencies will naturally require less maintenance, but are not maintenance-free. Browsers frequently change. SDK language ecosystems frequently change. Software doesn't exist in a vacuum, and it is incredibly difficult to maintain backwards compatibility over time. Ask Microsoft. In the web world, it's practically impossible.
> What impact? One guy [...]
Yeah, fuck that guy.
> Everything else we've seen in this thread is pile-ons.
Have you seen Reddit? But clearly, everyone who disagrees is "piling on".
> Aren't you doing the same thing? You have been ignoring and dismissing the feedback that this is actually not that big of a deal. Why do you think that your opinion carries more weight than that of the actual maintainers and users of the project?
Huh? I'm pointing out why I think this was a bad move, and why the negative feedback is expected. You can disagree with it, if you want, but at no point did I claim that my opinion carries more weight than anyone else's.
> Why are you ignoring my previous comment that contradicts this opinion?
Because it doesn't contradict it, it just disagrees with it. Because what actual argument did you have that people using an old version of the software can't keep using it? The one about things constantly breaking? On the web, the platform that's famously stable and backward-compatible? Sorry, I just don't find that believable for projects like htmx and Datastar which are very self-contained and use basic features of the web platform, not crazy things like WebSQL for example.
> When Redis changed licenses to SSPL/RSAL, users were also free to continue using the BSD-licensed version. Was that not a rug pull?
Firstly, there are tons of people on old versions of Redis who didn't even upgrade through all that and weren't even impacted. Secondly, Redis forks sprang up almost immediately, which is exactly what you yourself said was a viable path forward in an earlier comment–someone new could take over maintaining it. That's effectively what happened with Valkey.
> My argument was beyond any legal licensing terms.
And my argument is that there is no 'beyond' legal licensing terms, the terms are quite clear and you agree to them when you start using the software. In your opinion should it be standard practice for people to weasel their way out of agreed license terms after the fact?
> Any public deed builds a brand and reputation, which in turn can lead to financial opportunities.
Notice that you're missing quite a lot of steps there, and even then you can only end with 'can lead' to financial opportunities. Why? Because there's no guarantee that anyone will be able to monetize exposure. No serious person would claim that that uncertain outcome constitutes any kind of 'contract'. Anyone who does should be rightly called out.
> It's about building meaningful relationships between people in ways that we can all collectivelly build a better future for everyone.
Then by your own logic shouldn't everyone contribute to that effort? Why is it that only the one guy who creates the project must bear the burden of maintaining all of it in perpetuity?
> Seeing OSS as a plain transaction
Isn't that what you are doing by claiming that OSS is about providing software in exchange for exposure?
> Yeah, fuck that guy.
The guy who didn't even explain what exactly he lost by not being able to use the new paywalled features? The guy who likely was not impacted at all, and was just ranting on his blog because he didn't like someone monetizing their own project? You want us to take that guy seriously?
> everyone who disagrees is "piling on".
Everyone who disagrees? Yeah. Anyone who provides a coherent argument about exactly what they are missing out on by not being able to afford the paid version? I would take them seriously. I haven't seen anyone like that here.
It would certainly be a shitty experience, if there actually was a rugpull, which there was not. People who were using the version of Datastar that had all those features are still free to keep using that version. No one is taking it away. No rug was pulled.
> a contract is established between developers and potential users
Sorry, but no. The license makes this quite clear–every open source license in the world very explicitly says 'NO WARRANTY' in very big letters. 'No warranty' means 'no expectations'. Please, don't be one of those people who try to peer-pressure open source developers into providing free software support. Don't be one of the people who says that 'exposure' is a kind of payment. I can't put food on my table with 'exposure'. If you think 'exposure' by itself can be monetized, I'm sorry but you are not being realistic. Go and actually work on monetizing an open source project before you make these kinds of claims.
> why even bother releasing software?
Much research and study is not useful for many people. Why even bother doing research and development? Because there are some who might find it useful and convert it into something that works for themselves. Open source software is a gift. The giving of the gift does not place obligations on the giver. If you give someone a sweater, are you expected to keep patching it whenever it develops holes?
> If a project is not maintained anymore, it can be archived, or maintenance picked up by someone else.
Then why can't it be maintained by someone else in the case of using the old free version?
> A web framework, no matter how simple, will break eventually, most likely in a matter of months.
Sure, the ones that depend on a huge npm transitive dependency cone can. But libraries or frameworks like htmx and Datastar are not like that, they are single <script> files that you include directly in your HTML. There is no endless treadmill of npm packages that get obsoleted or have security advisories all the time.
> You expect people who want to build a web site and move on with their lives to dig into a foreign code base, and fix the web framework?
Well...ultimately, if I use some open source software, I am actually responsible for it. Especially if it's for a commercial use case. I can't just leech off the free work of others to fix or maintain the software to my needs. I need to either fix my own issues or pay someone to do it. If the upstream project happens to do it for me, I'm in luck. But that's all it is. There is ultimately no expectation that open source maintainers will support me for free, perpetually, when I use their software.
> A few people (that we know of) were directly impacted by this
What impact? One guy blogged that just because there are some paid features, it automatically kills the whole project for him. There's no clear articulation of why exactly he needs those exact paid features. Everything else we've seen in this thread is pile-ons.
> Doubling down on this, ignoring and dismissing such feedback as "performative"
Aren't you doing the same thing? You have been ignoring and dismissing the feedback that this is actually not that big of a deal. Why do you think that your opinion carries more weight than that of the actual maintainers and users of the project?