It also allows us to initiate a dialogue with commercial companies who mostly work with open-source versions of Qt.
This sounds creepily like Oracle’s requirement to have an account to download certain free java things and the ensuing dialogues they like to engage in with companies and the Oracle legal/sales teams.
To be clear, I am not asserting that I believe that is how the Qt company plans to behave. But given what we know about how the most prominent example of an actor who did this behaved, it’s an uncomfortable look for me.
From my interpretation it means you need to login to download the qt-*.runbinary installer, but not the source code packages. As the distros probably compile from source this shouldn’t be a problem, either.
This is an extremely concerning thing, and seemingly nobody in libre software noticed this.
Qt is turning the libre community into beta testers and guinea pigs, and they even specifically admit that they are doing this to make their commercial offering “more stable”. How will this affect distributions that want a stable ABI for more than 6 months (Debian, RHEL/CentOS, Adélie come to mind)? What about Telegram?
Qt 5.14 has been a disasterous release, it completely broke high-DPI setups with KDE. The KDE team has had to fight tooth and nail to get Qt to even acknowledge the issue, and it took until 5.14.2 to get the issue fixed. Can you imagine if all libre OSes HAD to ship 5.14 because 5.12 wasn’t LTS? This is just plain irresponsible.
According to an article from today this should make no difference. Apparently the distributions are already rolling their own versions of QT and are not using the official release cycle.
Unless I misunderstand the issue here this has been noticed and handled by KDE and Trolltech (the company that originally created Qt) over a decade ago.
IIRC there’s an agreement in place that ensures that the last version of Qt will become available under a BSD license in case the owner of Qt tries to make it unavailable for KDE.
When I was working for Quby, I was unpleasantly surprised at how expensive Qt is, how little flexibility the Qt company shows when dealing with prospective customers
Having witnessed similar licensing discussions with Qt in the past, I think that part of Qt-the-company’s problem is that they are incompetent when it comes to selling what they have.
The user in question was already complying with the LGPL on existing platforms, but was considering deploying Qt application on a new platform and wondered whether it would be simpler to get a commercial license to avoid some of the hoop-jumping. The Qt company could have simply pocketed a decent fee to license Qt commercially for that platform. Instead, they tried to push for all-or-nothing (i.e. commercial license on all existing platforms), at an astronomical price. Whenever talks seemed to be going in a reasonable direction, Qt would recall their salesperson and send another in their place, effectively resetting the discussion. Given that the user was already complying with the LGPL, they walked away, used Qt under the LGPL on the new platform, and Qt got nothing from what should have been an easy sale.
Talking of potential future KDE/Qt splits, I’d be interested to see an updated version of the graph in this blog post, if anyone knows how to get the data: https://pusling.com/blog/?p=362
I’m glad they mention funding for startups. Thing is, they should probably say first or seed round instead of a specific number. For instance, YC startups get more than $100k, having SV startups use Qt would be big for ecosystem, and successful ones could pay more later. I don’t really see a losing side to expanding it.
You know, for a long time I scratched my head, thinking “WHY would the big distros pick Gnome as their desktop of choice when KDE has some really clear advantages?”
This is why. It’s not even about Qt taking away open source projects ability to build with Qt, but about the never ending struggle between the immovable demands by shareholders or owners to make profit against the desire to give back to the community.
It’s as much about psychology and perception as reality, and while I don’t fault Qt at all for doing this because they are a commercial venture and this doesn’t meaningfully impact any open source users of the toolkit, I am now fully on board with the choice of Gnome as the mainstream Linux desktop as decided by Ubuntu and Redhat.
I’m running for a tender to modernise a SCADA system written in MFC. One of the requirements was to support it for 10 years: basically apply modification to keep up with future manufacturing line changes (nothing big is expected, but sometimes changes are necessary).
I was seriously contemplating using Qt for the HMI stuff, but the pricing for a small shop is pretty wild, unless you do lot of stuff with it, but that is only one small part of the equation, because that can be built in the quote, and might be amortizable over different projects. The real problem is this constant insecurity about the future of the project. I begin to understand why people choose GTK or some MS technology instead.
I hate when I need to register for downloading some crap, like the STMicro free tools, or nowadays the Atmel stuff.
edit: the new startup pricing is pretty nice, so that is a good point btw. Not getting closer to the decision.
The problem is that Qt has become, over the years, bigger and buggier, and I doubt whether releases made fresh off the main development branch will be stable enough to provide, say, a stable version of Krita to our millions of users.
As a semi-regular user of Qt:
I think Qt fell for the mobile hype + scope creep. I am not sure how popular Qt Quick is on mobile, but Qt was actually a widget toolkit (what is now basically QtWidgets). But with QML/Quick, QtWidgets went into maintenance mode. Last time I made a Qt app, there were quite some glitches and bugs in QtWidgets that had also been in the bug tracker for a while. And now their Qt 6 vision states that they also want to branch out to machine learning and AR/VR.
I can see why they are chasing trends as a company. But it would have been nice if Qt would just have focused on traditional desktop widgets, iterating to come closer and closer to native widgets, so that one can write applications seamlessly for Linux, macOS, and Windows. Plus improvement of the binding story outside Python (where Gtk+ clearly does better, since it is a C library).
Who knew .io would take off like it did… Don’t kick yourself. It’s at least a cool, alternative universe where they might have written you a check for it. :)
It’s at least a cool, alternative universe where they might have written you a check for it.
I have the unusual honor to say as a 16-year-old I was able to sell a domain to a company for a four-figure amount. The alternative universe is a reality for me, albeit in a different context.
This sounds creepily like Oracle’s requirement to have an account to download certain free java things and the ensuing dialogues they like to engage in with companies and the Oracle legal/sales teams.
To be clear, I am not asserting that I believe that is how the Qt company plans to behave. But given what we know about how the most prominent example of an actor who did this behaved, it’s an uncomfortable look for me.
When did Oracle buy Qt, or am I missing something.
From my interpretation it means you need to login to download the
qt-*.runbinary installer, but not the source code packages. As the distros probably compile from source this shouldn’t be a problem, either.I might be totally wrong though :)
I’m with you on this one. I’m not a native speaker, but the article’s not entirely clear.
How will this affect KDE? For example KUbuntu LTS
This is an extremely concerning thing, and seemingly nobody in libre software noticed this.
Qt is turning the libre community into beta testers and guinea pigs, and they even specifically admit that they are doing this to make their commercial offering “more stable”. How will this affect distributions that want a stable ABI for more than 6 months (Debian, RHEL/CentOS, Adélie come to mind)? What about Telegram?
Qt 5.14 has been a disasterous release, it completely broke high-DPI setups with KDE. The KDE team has had to fight tooth and nail to get Qt to even acknowledge the issue, and it took until 5.14.2 to get the issue fixed. Can you imagine if all libre OSes HAD to ship 5.14 because 5.12 wasn’t LTS? This is just plain irresponsible.
According to an article from today this should make no difference. Apparently the distributions are already rolling their own versions of QT and are not using the official release cycle.
Unless I misunderstand the issue here this has been noticed and handled by KDE and Trolltech (the company that originally created Qt) over a decade ago.
IIRC there’s an agreement in place that ensures that the last version of Qt will become available under a BSD license in case the owner of Qt tries to make it unavailable for KDE.
Having witnessed similar licensing discussions with Qt in the past, I think that part of Qt-the-company’s problem is that they are incompetent when it comes to selling what they have.
The user in question was already complying with the LGPL on existing platforms, but was considering deploying Qt application on a new platform and wondered whether it would be simpler to get a commercial license to avoid some of the hoop-jumping. The Qt company could have simply pocketed a decent fee to license Qt commercially for that platform. Instead, they tried to push for all-or-nothing (i.e. commercial license on all existing platforms), at an astronomical price. Whenever talks seemed to be going in a reasonable direction, Qt would recall their salesperson and send another in their place, effectively resetting the discussion. Given that the user was already complying with the LGPL, they walked away, used Qt under the LGPL on the new platform, and Qt got nothing from what should have been an easy sale.
Talking of potential future KDE/Qt splits, I’d be interested to see an updated version of the graph in this blog post, if anyone knows how to get the data: https://pusling.com/blog/?p=362
I’m glad they mention funding for startups. Thing is, they should probably say first or seed round instead of a specific number. For instance, YC startups get more than $100k, having SV startups use Qt would be big for ecosystem, and successful ones could pay more later. I don’t really see a losing side to expanding it.
You know, for a long time I scratched my head, thinking “WHY would the big distros pick Gnome as their desktop of choice when KDE has some really clear advantages?”
This is why. It’s not even about Qt taking away open source projects ability to build with Qt, but about the never ending struggle between the immovable demands by shareholders or owners to make profit against the desire to give back to the community.
It’s as much about psychology and perception as reality, and while I don’t fault Qt at all for doing this because they are a commercial venture and this doesn’t meaningfully impact any open source users of the toolkit, I am now fully on board with the choice of Gnome as the mainstream Linux desktop as decided by Ubuntu and Redhat.
I’m running for a tender to modernise a SCADA system written in MFC. One of the requirements was to support it for 10 years: basically apply modification to keep up with future manufacturing line changes (nothing big is expected, but sometimes changes are necessary).
I was seriously contemplating using Qt for the HMI stuff, but the pricing for a small shop is pretty wild, unless you do lot of stuff with it, but that is only one small part of the equation, because that can be built in the quote, and might be amortizable over different projects. The real problem is this constant insecurity about the future of the project. I begin to understand why people choose GTK or some MS technology instead.
I hate when I need to register for downloading some crap, like the STMicro free tools, or nowadays the Atmel stuff.
edit: the new startup pricing is pretty nice, so that is a good point btw. Not getting closer to the decision.
I honestly don’t get it. Does that mean that the QT binaries won’t be available through the Linux package repositories?
No, because they’ll build their packages from source, as they currently do.
As a semi-regular user of Qt:
I think Qt fell for the mobile hype + scope creep. I am not sure how popular Qt Quick is on mobile, but Qt was actually a widget toolkit (what is now basically QtWidgets). But with QML/Quick, QtWidgets went into maintenance mode. Last time I made a Qt app, there were quite some glitches and bugs in QtWidgets that had also been in the bug tracker for a while. And now their Qt 6 vision states that they also want to branch out to machine learning and AR/VR.
I can see why they are chasing trends as a company. But it would have been nice if Qt would just have focused on traditional desktop widgets, iterating to come closer and closer to native widgets, so that one can write applications seamlessly for Linux, macOS, and Windows. Plus improvement of the binding story outside Python (where Gtk+ clearly does better, since it is a C library).
Off-topic: I used to own this domain.
I should have never gotten rid of it. Oh well.
Who knew .io would take off like it did… Don’t kick yourself. It’s at least a cool, alternative universe where they might have written you a check for it. :)
I have the unusual honor to say as a 16-year-old I was able to sell a domain to a company for a four-figure amount. The alternative universe is a reality for me, albeit in a different context.
P.S. I’m holding out on w-1.net and w-3.net. ;-)
That’s cool!