It's amazing how many people in this thread are not only defending this, but talking about how great this licensing model is.

I'm guessing all these people have startups (or ideas for startups) that rely on gouging customers with never-ending subscription models.

I personally think subscription licensing models is the worst thing to happen to the software industry. I miss being able to purchase and own copies of software. I recently stopped using any Adobe apps and moved my entire workflow over to the Affinity suite. I prefer to support developers who don't rip-off their consumers

I'm an embedded Qt licensee. I don't get any of this at all and now I need to spend hours I don't have with my legal team understanding what the impact is. I suppose we're staying on 5.15 for the lifetime of the product I haven't even launched yet.

I mean, it's a given every time Qt pops up on HN a good 50% of the thread is sucked up with licensing questions and information/misinformation. Getting kind of tired of it all.

What I do recall is that if you let the development license lapse, you're locked into whatever version was latest at the time. So I can't upgrade Qt on any in-field device updates which is usually just fine, especially if you managed to get to a decent LTS version, but any Qt bugfixes after that are off limits.

That is correct. These distribution licenses for shipping Qt with devices were never perpetual in the sense that you can ship as many devices with Qt as you want. But this was a separate license, which AFAIK you also could buy in bulk, which is totally OK. As far as I understand it, you now need a full subscription for distribution with the danger that prices may change any time.

I gave up ongoing fees for software components I use to develop with Borland Turbo C++. I am not going back now.

My interpretation is that I could switch to LGPL Qt for client software and keep my source code if I dynamically link to Qt.

I may be too strident as a commercial developer in my anti-LGPL 3 rhetoric, but my cloud-apps have been avoiding it assiduously. Client side, I should look for this kind of flexibility.

A PWA? Sure. But a native app with seamless integration? No way.

Instead of Windows, macOS, and various Linux flavors, now you've got Windows, iOS, still macOS (for now), various Android versions, and various Linux flavors (including Chrome OS). Good luck!

(Depending on the type of app, you might also want or need to support various embedded devices or game consoles. I guess QT doesn't help there though.)

https://developer.apple.com/documentation/apple_silicon/runn...

I don't understand your point. In your opinion what's a mobile app that's "backwards compatible with desktop", and what makes that app not a desktop app?

The desktop OS would expose the same platform SDK as the mobile OS. (Needless to say, apps depending on the accelerometer, selfie cam, etc. for major functionality would not work.)

This is already the case with the latest generation of Macs. No speculation necessary. It's possible and has been done. Also Chromebooks can run Android apps.

What remains to be seen is when (not if) Windows would mimic what Apple did, as they have been doing for the past decades. I suppose the easiest way would be to run Android (and Google Play) on WSL. Android and Google Play can run on x86–no need for emulation. How clunky this turns out is how willing MS and Google are to work together.

In the end there will just be two platforms you would have to support as a GUI applications developer: iOS and Android.

Windows still owns the desktop but they mimic Apple. Also since ca. 2008 the population under 30 has been mostly on Apple. Those people will get older and replace the generation that was brought up on Windows.

On my 8 GB computer, I used to run in the same time :

- VsCode (700 MB)

- Teams (600 MB),

- PostMan (350 MB !? WTF)

- WhatsApp (consumes almost 600 MB after 1 hour !)

I had to switch to the web version for the last 3...

In contrast, running 25 torrents with qBitTorrent: ~25 mb.

I live in France and there are 4GB here too: https://www.ldlc.com/informatique/ordinateur-portable/pc-por...

hell, in one of the biggest online retailers, the second biggest category is 2 gigabytes (number in parentheses is the number of laptop models available with that ram amount)

https://imgur.com/a/LL8H7aP

I would be super happy if I could run several similar engines under one main process, I don't care as much about security concerns when I'm developing my own software, and would not mind some "intraprocess security holes" if I could reduce my memory footprint like that and then turn it off when I want the "security holes" to be closed again.

It is just to illustrate the fact that Electron, after all, might be a real problem for native app.

https://github.com/microsoft/vscode/issues/22900

Running with 32 and looking for 64Gb laptop

Having 2Gb is pretty limited

And If I need 32GB and 16-CPU to run 5 or 6 apps...

Asimov was right after all :)

The mantra was : "Yes, pretty slow but don't panic, in a couple of months better hardware will fix that" !!

Hardware matters more than people

Do you really wants to limit your folks in such way for something like $150?

My personal laptop has 8 GB, is 10 years old by now, I am not limited in anything I care about.

Seems reasonable? As long as you're getting benefit from Qt by distributing software using it you need to keep paying for it?

I'm just imagining a scenario where I build something with a subscription library and then 5 years down the line they get acquired by some private equity firm and increase the subscription cost by a factor of ten. Now I'm faced with a choice of either paying license ransom or losing the ability to use all of my own original software and creative effort.

I'm not in love with "everything is SaaS," but I recognize that a lot of software businesses really want recurring revenue rather than a rush of orders -- many of which are lower-priced upgrades -- when new releases come out, then a trickle until the next for-pay release (which a subset of your existing customers won't upgrade to). And, in an era where a lot of people have come to see upfront prices of $99+ for applications as outrageously high -- and worse, have been trained by app stores to expect free upgrades indefinitely -- the old model may just not be sustainable.

This is just cynical!

> I for one appreciate products that evolve and improve.

And I do not. Software updates are a huge pain in the ass for almost no benefit to me

Both sides have their benefits, so to say one is worse than the other seems not right. It’d be nice if there was an option on which model you chose (for example, $5k/seat for perpetual single version or $300/month/seat for subscription), but alas, that’s not available.

> And I do not. Software updates are a huge pain in the ass for almost no benefit to me

Ah, but there is a benefit that you just don’t see: bug and security fixes. Your boss isn’t going to be happy if your company is hacked because your IT department didn’t want to spend effort updating (assuming that’s in the budget).

This obviously ignores 0-days, but there’s not much a downstream user can do there.

Take photoshop... I'm not a professional graphic artist, and I have need for PS maybe once a year at most. Depending on the importance of that need and my desire to own the software it might justify a onetime purchase. But with their subscription model it just isn't worth it.

One of the things I liked about the old model was it put professional tools into the hands of amateurs if they were willing to save up their money. Now, instead, you pay a yearly tax that is not insignificant for no real gain

But the flipside is that $10/month is a lot easier to stomach for most of the general public than a one time cost of $300+.

It’s why there’s monthly payment plans for everything expensive. $40/month for a (24 month) lease-to-own phone is a lot easier to budget for than a one time $900+tax purchase. But when I go to BestBuy or wherever, I have the option to choose the payment plan or upfront. Not with software.

But yes, with subscriptions, it would be cheaper to buy it outright, but not everyone has that luxury. We can argue all day about whether people need Photoshop or whatever, but the best option (IMO) is to allow subscription and purchasing and let the user choose. Then those who can afford the $300+ up front can spend that while those with little disposable income can spend $10/month.

While the FAQ says 'Maintenance is included', they phrase that as 'access to the latest version' and 'support'. That doesn't place any legal obligation on Qt to fix bugs.

Qt can leave Qt6 to rot, never update, and still get money from subscriptions. Or Qt could be busy with Qt7, and in order to receive said fixes one has to significantly invest and port the application.

Access to support means even less - they could automatically close all tickets and still techinically comply.

No, but neither does the version of Qt in the already compiled and being distributed software, and yet even if no more development is done and no new version of Qt is used, it sounds like the licensing fee still applies.

You're welcome to argue QT saves more time than its license costs, but I didn't find that to be the case. Also cmake sucks, no bazel support, lame. I have found the QT value proposition questionable at best and have largely abandoned it. YMMV.

New features, sure, ask money for that, fair enough.

Which is why I would be OK with Qt 6 having a new license model, as long as all bugs fixed in 6 are backported to previous stable (at least) ie. 5.x, without 5.x getting relicensed.

Works for JetBrains.

Software is copyright literature under licence.

Why? Most businesses are not MS/etc. which have resilient multiple income streams. Many single-app companies either go under, or stop selling that app and so no longer update or support it.

A subscription model ensures that the developer always has a reliable app-specific income stream to justify investment & support.

My JB subscriptions expire at the end of this month, if I don't re-up, I have to downgrade to 2020.1 versions and forego all the improvements and bugfixes in 2020.2 and 2020.3 (2020.3 is the latest release, which I'm currently using).

It's a good thing their products are top notch.

“The little known KDE Free Qt Foundation makes sure that Qt stays free and open-source. It guarantees that all Qt modules currently licensed under LGPLv3 must continue to be available under LGPLv3 in the future. This covers all modules from Qt Essentials and many add-on modules. If The Qt Company discontinued the FOSS version of Qt, the Foundation would have the right to make Qt available under the BSD license. This is a very powerful protection of free and open-source Qt.“

https://www.embeddeduse.com/2019/12/21/safe-guarding-the-lgp...

https://dot.kde.org/2016/01/13/qt-guaranteed-stay-free-and-o...

I want options. Don't tell me what I want.

In any case, "options" sounds good, but it is precisely this short-term economic logic I am claiming many misperceive as being in their interest.

There are many instances of apps (, libraries, etc.) simply disappearing and being frozen in old versions because of no revenue stream against them.

You may "want" something quite different 2yr after you've invested a massive amount in using Qt, only for there to be no future versions and your app becomes uncompetitive with others which are using increasingly modern libraries.

The software subscription business is essentially a huge price hike masked by paying in installments.

>simply disappearing and being frozen in old versions because of no revenue stream against them.

It means that the new features in version X were not good enough to entice new buys or upgrades.

Subscription model is great - for the software companies. It's not a good model for consumers.

The current subscription model will end up being just like Fifa or Madden games, the same thing every year with minor changes because consumers are locked-in and have no alternative.

None of them will ever abuse that dependency to take all of your profit. No, sir.

Now you need a subscription to keep using your tool? You need permanent Internet access, to keep using your tool? You need t keep paying to access you files?

This is why Affinity Photo will probably end up taking more and more market share away from Adobe's products.

If you don't like it, keep using your old version under the old licence.

In most jurisdictions, buying a copy of Photoshop is the purchase of a good.

This may be how you wish things were, but not how they actually are.

Read any software licence and see what you are actually getting for your money.

You can try to find old licenses for CS6, sure, but at some point, but it's a very different purchase and licensing system from what they have now.

A typical proprietary software licence text is irrelevant when it comes to terms of ownership, transfer of ownership, use of the good etc. as the law applies and determines "how [things] actually are". A contract, let alone a unilaterally dictated licence text, is legally unable to take away certain rights the consumer enjoys.

No, the reasonable model (as someone how's shipped commercial libraries and researched the market) is that subscription gates support and future updates. So as long as the subscription is active you get updates, patches and support. Once it passes you're kept on old version. It's fair to the seller (keeps a steady revenue stream and most customers will want to keep the subscription for updates anyway) and fair to the customer (they can't be easily extorted with price hikes and their software doesn't get sabotaged by external subscription pricing).

Why expect good faith and dependability with the precedent of their decision to screw KDE?

This allows price increases with major changes.

"Free support for life" isn't sustainable, and striking the balance here is hard.

Image-Line (makers of FLStudio) have been pulling it off for a couple of decades now...

"Celebrating Lifetime Free Updates for over 20 years" -- right on their front page

Why would we have to pay multiple times for the exact same piece of software, lines of code?

Also given that software is by default sold with defects (as it is "impossible" to write perfect software), it makes sense that the selling price includes, at least, future minor patches. I would consider them as the vendor covering for its own manufacturing defects (because that's what bugs are).

I agree however with charging for new features. That's fine and makes sense.

Development, distribution and usage are different things. I don't have much of a problem with selling a subscription for the development environment, but for distribution and usage you should only need a (perpetual) runtime license, which you usually buy in bulk. AFAIK that's how QNX does it, for instance.

Maybe to you. To me it reads like: do not ever touch this tool to do anything that has even remote chance to go commercial.

Yes. I dont quite understand what's the fuss here. So someone please explain to me like I am five.

QT 6 decide to charge money, in this case a commercial license for commercial usage. i.e If your software is making some money using QT6, you will have to pay. Otherwise you can happy use the LGPL License. Seems fair. You might disagree with its pricing, depending on your business model, And may be that business model would not be a good match / fit for using QT6.

But so far all the comments seems to be against the company charging anything. Why?

Nope, you can make a commercial closed-source application that links against LGPL-licensed library. Also, the source code model has nothing to do with "making money". For example, Qt is certainly "making money" with their source code, but did you know that the WebView component in Qt is actually COPIED from Apple's Safari WebKit source code? Yes! So they are STEALING Apple LGPL source code and making money with it! And what more, Apple Safari certainly runs on iPhones and Apple if someone is certainly making some big $$$!

You seem to have misunderstood what changed. No, Qt did not change the licensing terms to charge for "making money" with Qt. They just made more of their previously free support functions exclusive to paying licensees.

So in this case, you'd have to package the software so it doesn't include Qt (in an RPM/deb) so it pulls in the system Qt? Or on Windows, provide them a link so they can download that component manually?

Also, businesses do pay for support as insurance; say tomorrow Windows patches something that screws up a bit of Qt your product depends on, what are you going to do? Waiting for a reply on a random mailing list and hoping someone has time to fix it, is not going to cut it.