Sure they can. If you pay 100-1200x more for your cloud services than you need, maybe not. And I guess that is a theme, if you have subscriptions you don't really need to care about expenses because you have a steady income. Absurd? Yes, but that is how our industry behaves. I'm appalled at what people are paying for the cloud that literally could be served from a raspberry pi. Yes, you need redundancy etc. and that takes time and thought. But have you ever factored in the time and expertise required to manage the cloud? It is incomprehensible.

So, how many subscriptions do you need to be able to maintain your software? Is 10 enough? No? Conclusion, subscriptions are not sustainable?

Similarly, how many purchases do you need to recoup the initial development? You will hardly break even after the first 10 sales. But once you do break even every single purchase goes 100% to maintenance and new features. That is a very good position to be in.

Of course it matters what kind of product you have and how big of an audience you can get. If it is very niche product and you only expect 100 sales then maybe a one off payment isn't particularly appealing. But then again, maybe the value lies in support instead. Or, of course, a subscription.

Paying for upgrades is also fine! But without shafting your current users. Compute is dirt cheap. If you sell someone a piece of software for $300 you can afford 20 cents per year for cloud costs to maintain your relationships.

Agree. However the core problem with most SAAS products is: the companies are beeing greedy and the products fail to deliver enough value or keeep missing a certain quality bar to make you happy.

You could easily fix that by a base price and a maintenence fee (per year, since first buy) which would make it easy to opt in/out of a subscription model. The fact that companies are beeing greedy and fail to deliver prevents this from happening.

One of the problems of no-one being liable for security, is that no-one has the incentive to thoroughly separate software you really need to trust, and that you care whether it's vulnerable to hacking, from that which you don't. "everything needs a revenue for security reasons" would make sense if companies were giving an effective security guarantee and investing in fulfilling it, which they aren't. If software liability existed, most providers would recuse themselves from the most security sensitive features and we'd have a few big providers doing stuff like syncing, handling long term data storage, sandboxing, etc. rather like now with credit card data. At which point a lot of software becomes unimportant to update.

Conversely, it isn’t my problem to make a business model work.

Personally, I’ve decide if I can’t buy something and just have it I will do without. So be it.

This is why I sell my products with a one-time payment (lifetime access/usage), but with an optional support/updates payments.

It seems to work quite well for both my customers and my products.

I think the subscription model needs to be refined.

500$/year (invented amount) for a Photoshop subscription that I’ll use 3 times in a year? But when I want it, I need photoshop, not $SOMETHINGELSE. Sell me a $5/day license and I’ll be happy to pay every time.

Subscription model is ok for heavily used software, but doesn’t properly address whatever I use every now and then.

The campfire model specifically is even worse; it becomes a liability to maintain that I can’t request of the vendor, but have to take care of myself. The burden of actively looking for vulnerabilities, and fixing them, is on me.

The idea may be venerable, but I don’t quite understand who would think you could consider a networked online chat server as finished…

Many SaaS disappoint after a while. They suggest you are paying monthly and benefit from ongoing development in return. Instead prices get increased, and essential new features are locked behind additional pricing tiers. Premium, professional, enterprise, what’s next? The user interface becomes an advertising app for the upsell. It’s an abuse of trust. So the problem is not so much subscriptions and SaaS themselves, but the business practices they enable.

Yes. That is why I want a Hardware Product with Software Model like iPhone. I want my NAS / HomeServer with all the OS and Software included. With Security update for at least 8 years. ( Synology is the closest thing we have got )

Apple could have done this, but they were so focused on their Services Revenue they want everyone to subscribe to their iCloud. They could have allow the NAS / HomeServer to backup to their cloud.

Fully agree. I am a software developer myself. I can't imagine making a living by charging my customers just once, ever. It just doesn't make any business sense.

Software needs to be maintained and developed, and just following the evolving technology (operating systems, libraries, environments) is a lot of work. On top of that, you need to fix bugs, provide support, and yes, develop new features. This is not feasible with a one-time purchase.

For some reason many people like to play a game of pretending. The person buying pretends that it's a one-time purchase. The person selling pretends that it's a one-time payment. But then it turns out that a new version comes out every two years or so and you need to upgrade, so you end up paying. It's a subscription, but everybody pretends it isn't.

For the software I use and rely on, I would much rather see a subscription model, which is sustainable.

> So OSS aside (which has its own complicated economics)

Assuming you're actually referring to free software and not open source software, it really doesn't, though. It's straight up better in every way than service-oriented software and commercial software in every way... except actually compensating developers. I'd work for a pittance writing free software if there were any institutional support for it. But who wants to kill the golden goose, even if it means our lives would all be greatly improved?

That's not "complicated", this is the opportunity to make a shit ton of money by charging people for software despite insignificant marginal costs. Even if it means humanity writes the same goddamn software over and over and over and over and over and over and over again, mostly shittier than the last iteration.

You're buying once for $300.

It's not a "lifetime" payment unless you can run it off-line. If the software needs to talk to vendor licensing/authentication services, then any "lifetime" license is a lie for the reasons you mention.

Ok, but think what would happen if you made a subscription and the provider went bankruptcy. Or was acquired by someone who shuts down the service? You would have payed for something that all of a sudden disappears and you have nothing, not even an old unsupported version fir the time you need to find alternatives.

I'd say there is a fourth category too - things that would be perfectly fine as a simple, local program purchased once that grow over-complicated cloud features to justify a subscription model.

Examples of this would be Lens, Postman and now Insomnia. This sort of behaviour is why I use k9s and Bruno instead.

Jetbrains does this well. Pay for a year and then you fallback to that license/version forever

That’s the model of Nova from Panic. Stop paying, and you stop getting updates after the moment your subscription lapses. That’s it. There’s no other complex system. It works great!

I saw this adopted by Beekeeper DB manager - https://www.beekeeperstudio.io/pricing/

>Pay for a year and get a lifetime usage license.

Looking at a lot of this stuff, "lifetime" is a bit misleading. It's for the life of the version. Buy v1.x and you can use v1.x forever, but if you want v2.x, pay again.

Of course, "forever" is only until everything around it leaves to the software not working, such as not supporting a new version of the OS it runs on.

There is no need for a client/server model for most pieces of software, integration to existing cloud storage is far more useful than your server. Updates have existed for many years although less common as software was often more fully featured to begin with. There is still plenty of non subscription software that is out there, fully functioning, and of a high standard. Affinity is a key example.

Not all software has to be SaaS.

Self-hosting is easier than ever nowadays, devices a lot more capable.

Soon, smartphones will be able to run powerful LLMs locally, so no need to use a paid GPT app subscription.

Take Photoshop for example, it's a product that runs locally on your computer. Why would you pay a yearly subscrption for it instead of a pay-once, use forever? I understand (optionally) paying for updates, but not just to keep using a product that is already on your computer

> People easily forget about the hidden costs and why the software distribution model from ancient times doesn't cover the costs nowadays anymore.

If this was true then there would be zero solvent software businesses that don't rely on subscription revenue, which is clearly not the case.

At least in a B2B context a yearly maintenance fee was covering the costs, without cloud subscriptions. I think the move to SAAS in many areas is just to establish a stronger lock-in.

There's still plenty of software where using a very old version is perfectly fine. Matlab, Solidworks, Photoshop, etc. If you bought a 10 year old version it will still work fine and you probably aren't even missing much.

If you are using Mac OS, you can find the highest quality programs in any category for sale for between $30 and $100. The old model still works great. I have purchased some fantastic software and these prices are cheaper compared to the 80s and 90s.

This. A subscription can work if a service is actually provided, as in regular updates, however I would say that's rarely the case and it's simply to make cash. Goodnotes, Pro Create, Affinity, all show you can make a fantastic product without a subscription and still bring regular updates.

And according to libertarian thought, ownership is the basis of ethics. I'm a bit surprised there isn't more of a libertarian backlash against the "you will own nothing and be happy" trend in our culture.

For every software that is not abandonware, there is always a recurring cost (besides cloud, there is support, security updates, aligning with OS updates, etc.).

There is something to be said for both sides. The counterpart to your argument is that, especially in niches or other markets where there is a dominant player, there isn't much incentive anymore to improve software when all your users are subscribers. The money comes in anyway. Besides that, you often see regular price increases in such markets because the user lost all power. If you end your subscription, there is no good alternative. In the pas you could just not update if the new pricing wasn't worth it.

The best balance may be in the JetBrains-like model, where you pay a yearly amount, but you keep access to historical versions when you decide to cancel. In this way money flows in for updating and improving the software. But if you fail to provide value, users can always cancel their 'subscription' and continue using the last version.

It does make sense, though? If the developer has to work to change the product to work for new MacOS versions, because of API that Apple changed, the developer should be reimbursed for the development time, right?

In what sense?

Because most issues arrise in this thread are about future updates, not the ownership of the current product.

With FOSS you get no guarantee for future updates.Many products that started FOSS, did it just to use it as a lead magnet, and once they had enough leads, they switched to SaaS as their main model, stopping support and updates for the FOSS version.

> what would an Office suite from 2005 yield me now

An office suite.

I occasionally write documents in Office 97. Yes, in emulation. And despite the emulation, it loads quicker and has no lag.

Some software does "rot" in time. But a lot doesn't. And we shouldn't be afraid of keeping things running.

>Am I ever going to play the copy of Dungeon Siege again that I bought in ca. 2002? No. It looks unbearable today.

Haven't played Dungeon siege specifically, but there's plenty of games from that era that still hold up, some even graphically. Be careful, or you're going invoke the wrath of the site's retro gaming community :)

You know what is my favourite software license model? The one Bitwig uses:

I pay once and get $years of updates and the last version I am stuck with I can use as long as it runs on my system.

Updates then prolong/refresh the update period for another $years and cost something like a quarter to a third of the full price.

This way the manufacturer has incentives to deliver new things and I pay them when my live circumstances allow it without them locking in the work I made using their software.

I bought 14 games last year from the early 2000's, racing simulator games. They absolutely suck compared to today's racing simulators, except that they don't because they did stuff differently that they don't do today and I miss some of those features and games for what they were.

I also routinely run tons of software that I paid once for, it works for me, and I never have need to upgrade. It rarely stops working with a new OS version, as far as Windows and Linux goes at least. Office 2005 will still export all types of formats, or I could print to a PDF, or I could even just send out an old school .doc file and most people can still open it in whatever program they use today, even if it isn't optimal.

Additionally, even if something only works for a limited time, I still got my use and money's worth out of most of that software. I don't expect physical things to last forever, so I also don't expect software to last forever. I'd rather pay once, twice, or more for "pay once" software that I use over 20 years than pay every single month, probably adding up to a lot more than the several copies of different programs over the years would.

> Am I ever going to play the copy of Dungeon Siege again that I bought in ca. 2002? No. It looks unbearable today.

On the other hand, I fairly regularly fire up an emulator to play Space Invaders or Galaga. And there's no end of projects porting Doom to pretty much anything with a display and a microcontroller in it.

The graphics aren't always the point, sometimes it's the gameplay, or the nostalgia.

I guess you don't like the proverb "Old is gold".

I still play Virtua Cop 2, and many companies/individuals aren't tech savvy but still use windows XP/7 with Office 2007.

My point is that if we can't distinguish which things are our assets, liabilities, precious, interests and hobbies then we won't understand where we pay once or in subscriptions.

I'll choose to pay once and own software forever if those software are too good for the next 3 to 5 years (like how we buy a smartphone or other things in life) or are being updated regularly.

> Am I ever going to play the copy of Dungeon Siege again that I bought in ca. 2002?

On that note, the makers of the Eufloria series from back-in-the-day have started working on a sequel. :)

• https://store.steampowered.com/app/221180/Eufloria_HD/

They have a Discord channel too which people are welcome to join:

• https://discord.com/invite/unsbgkR23z

You would be surprised. When I still used Linux on the desktop, I stuck with Office 97 (IIRC) for a long time, because it was the only version that worked well with CrossOver Office (Wine-based). It worked well for a long time, opening other people's files etc., even as new versions of Office came out.

> Am I ever going to play the copy of Dungeon Siege again that I bought in ca. 2002? No. It looks unbearable today.

Funny, considering many games I play to this day are older than that.

In many ways, I think they look great still. Pixel art for example ages very gracefully.

Why would software lose value? It's not a mechanical item that physically wears and rusts. What you're talking about is planned obsolescence. Those companies benefit from the software seemingly becoming less valuable. TeX is no less valuable today than it was decades ago. Why would it be?