[ptico]

“We produce an extremely expensive hardware and bill millions for consulting and support but don’t want to spend a single dollar to update a firmware, so now, dear OSS developer, it’s your responsibility to keep my solutions functional for the half of your career years”

Did I read this correctly?

[crabbone]

> Did I read this correctly?

No.

Firmware upgrade is a risk that has no mitigations for certain kinds of systems. No amount of money will solve this problem. No sum of money will convince a sensible patient to upgrade firmware on their pacemaker or anything like that.

But even for less critical systems -- what's the problem with wanting to pay less? This is like one of the primary economical drivers...

You also for some reason think that software upgrades are some sort of a natural phenomena which others have to adjust to, and it just happens on predictable interval, and if you miss your cycle you have to pay. Which is obviously ridiculous. A result of industry conditioning you to expect this to work a certain way, w/o questioning the reason for it to work this way.

What should drive software upgrades is in the large part the longevity of hardware. The author claims that the longevity of hardware has improved, even though industry didn't particularly invest into it. It's upsetting to have to generate a lot of e-waste just because we (as an industry) set our sights on a particular release schedule designed to maximize profits for those who provide releases and minimize them for those who consume them.

[ptico]

It’s not about pacemakers at all here for the same exact reason you described: if you’re not able to update firmware, you’re able to update underlying OS as well

The problem here is when you sell LTS solutions for $10M thinking you can pay $1000 for keeping your solution afloat for two decades.

It doesn’t work this way. Operation systems as well as frameworks and runtimes, are constantly changing, because the industry is constantly moving forward. The only way to keep up is to constantly (and regularly) update YOUR software too (and plan the budgets accordingly). It’s your responsibility to fulfil your obligation.

After all, you can still run an up to date Linux (or better NetBSD) on a very old hardware, the problem is that you didn’t update your software regularly to just keep up with changing API/ABI, means didn’t invested too much on a longevity of your product

[crabbone]

> Operation systems as well as frameworks and runtimes, are constantly changing

No. They aren't. It's not a natural phenomena out of our control. We decide when to change them. Presently, we make bad decisions. We should learn to make better decisions.

And you are incorrect when you think that pacemaker and similar equipment isn't the problem -- it totally is. Imagine that after ten years the hospital that installed a pacemaker needs to do a checkup or some other maintenance work on it with external equipment. But they were forced to upgrade the external equipment because there weren't any with LTS long enough to allow them to use a certified and vetted copy of. And now they have no way to connect to the older equipment they distributed to patients, equipment they have no means of upgrading, but also no means of dealing with, because they had to upgrade their own system.