Hacker News Reader

Comment by richardfey | Hacker News Reader

> What I have observed is that those persistent individuals are Puppy old-timers, very old-timers. A couple of them are in their 80s, and they have a collection of vintage PCs, all with optical drives. Optical media, CD/DVD, ISO files, that's what they know. And they ain't gonna change.

Hold on a sec...so the conclusion is: screw those old timers?

View on HN

I don't understand it either. Other distributions offer ISO files too, and that's very handy as it's still one of the source formats for virtualization platforms. I created my Ubuntu and Windows VMs from ISO images. I don't care how old that format is. Why change something that works?

And that is an absolutely insane state of affairs as well. I remember when my Packer playbooks stopped working because Linux kernel dropped floppy disk support and that was the method that was being used to inject the config on vmware.

99% of Linux installs are done on some hypervisor... ISO files are absolutely awful for this use case. Yet everyone keeps using them because the tooling is already jumping through thousands of hoops to make it work.

Why are they absolutely awful for this case, and what would be a better way?

I simply like the fact that I can take an ISO file and easily mount it as a drive in my OS. It might be a breakfast of a format, but all that complexity is nicely hidden away from me and it "just works".

The next paragraph about her non-tech 88 year old stepmother seems worse, he's throwing all of those people under the bus by comparing them to her.

Vintage PCs don't really have USB, but well, in his defense, people working with vintage PCs are already used to extra steps (e.g. if the PC has token ring and they want to connect it to Ethernet), so they'll need to add the extra step of "Write contents of USB image to ISO"...

I didn't read that as the conclusion at all. He didn't say screw anybody. He pointed out that he has provided tools for migration, and that about two of the people who are nevertheless still concerned by the situation won't be changing.

So to me it sounded more like admitting defeat for now with two people give or take, and letting the rest enjoy the benefits of the upgrade.

> screw those old timers?

Isn't it a volunteer project? If he doesn't want to support it, then he doesn't want to support it, end of conversation. If the old timers want support for a legacy format they can do it themselves.

I see no problem with that: setup a FAQ page and redirect people there. Link it from the downloads page and generally avoid receiving personal emails for your volunteer project.

I don’t see why anyone’s obligated to support Really Old Thing just because some people don’t want any change.

The value of backwards compatibility is directly proportional to the ratio of "value of user time" / "value of dev time". Depending on whether you are the dev, the user or a third party, you may attach different values to that ratio.

In the extreme, a dev might think that ratio is 0 and a user might that it's a large number. (People arguing that ISOs are already widely supported and useful are implicitly attaching a large value to that ratio)

I don't think there's an objectively correct value that can be assigned to that ratio, hence disagreements about whether Really Old Thing should be supported.

Supporting ISO takes basically zero effort. By dropping support for it the developer is just wasting both users' and his own time. It would be far better to say "you can download an ISO but I don't recommend it, here's why a plain disk image is a better way".

Obligated: no. Document how things are and instructions for those who want to make an ISO by themselves: it would be nice.

But there's no obligation, just expectations and communities.

More like "there is no change that will satisfy these users".

Reply on news.ycombinator.com