I am not too worried about anything pre-2000 as emulation has got really good over the last few years, just in time as the surviving legacy hardware became prohibitively expensive to acquire. The internet archive has also made it easier than ever to access old software.
What concerns me are early 2000s stuff with DRM and server-side content to make preservation difficult if not impossible.
Cloud anything is also a huge red flag for preservation.
On the other hand there is a game I played on PowerPC macs but the developer released it on github and now it's been updated and compiled for ARM and x86. As long as source is available it's usually not very hard to port to a different architecture.
I always privately called it a "Legacy Tail". It was a factor in the selection of a lightweight markup language to support old-fashioned industry (aerospace), but I think it should be a factor in any software selection, to be honest.
Short version? With a pile of XML docs, Asciidoc's DocBook XML backend represented the "Legacy Tail" or Lenny Effect (Lenny Tail?) that gave us some insurance we'd have a stable format. We didn't see that in the Markdown ecosystem - we would have had to pick and choose Markdown variants and extensions, anyway - and the ReStructuredText ecosystem had no such tail and too many Python/Sphinx dependencies (corporate infosec had a real chip on its shoulder about Python).
There was a lot of nights and weekends spent prototyping and wrestling with various old aero/milspec formats, but at the end of the day we came out with a publications format that could do more than the original could, on common tooling, where everyone who wanted to could contribute. I put it in the "win" column.
Emacs and Firefox are not popular right now, and the author mentioned using even more niche apps with long history, and a lot of his challenges are with things that have been around breaking other things that have been around, so how does he ignore the effect?
Definition nitpicking: They are only popular among a specific crowd. When less than 5% of all users use the software, I wouldn't call it popular globally.
FF was at 6% in 2021 apparently. And emacs, because it is not spyware, unlike most competitors of it, has an unknown user base. The definition (from the Oxford dictionary) is sufficiently vague though; you put a % on it, they don’t.
Yes, and they are also notoriously unreliable. With KYCd users (which won’t be the case in voting ‘best coding editor’) it is even unreliable, let alone things like reddit polls.
They aren't as unreliable as to miss an elephant in the room of eternal ignorance, and you don't need to shift to "the best", a basic "have you even heard of this" is sufficient
Firefox is definitely popular. If you stop a random person on the street and ask them "what is firefox?", they'll probably know that it's a web browser, and a good number will have used it at some point in their lives. Emacs can probably not be considered popular except maybe among the niche of programmers who use Linux or UNIX derivatives.
6% of worldwide users doesn't sound unpopular to me unless you compare to the Chrome monopoly. I'm just giving you an alternate viewpoint to show that you shouldn't get caught up on how much more popular the most popular program is when trying to determine whether something is popular. There are too many browsers to list, but most people are using Chrome/Edge/Safari or Firefox.
6% wasn't your argument, and is also wrong - that's only counting desktop, so it's missing a huge part of the market, which brings the number down to a mere 2.6%, try to spin that as popular unless "monopoly"
And your last statement is true with just Chrome, so you can add any other browser there for the same effect
Hah, you'll get down voted to oblivion on HN for recommending Windows over Mac.
But broadly speaking you are right. I run software today on Windows 11 that I first acquired in the 90s. Some of it console based that run under the command prompt.
While Windows is justly famous for backward compatibility though, there are some exceptions.
Firstly 16 bit programs don't run on 64 bit Windows. (Well, unless I fire up a 32 bit Windows in a VM.) This is not a train smash for me, although I did have to replace some -really- old programs.
Secondly, games. MS removed a cdrom driver which a lot of my old games used for copy protection. I'll need to dual boot those into win 7 if I wanted to play them. But the urge to even bother only strikes about once a year.
Regarding word processors, I used Word Perfect in the DOS days and Word since then, and everything still opens fine.
Obviously everyone's path is different, and their experience is different, but if long-term commitment to running existing, and past, software matters to you the maybe select an OS that prioritizes that over one that very much does not.
> if long-term commitment to running existing, and past, software matters to you the maybe select an OS that prioritizes that over one that very much does not.
I guess that's the downside of vertical integration where the OS is basically just a means to sell hardware.
Truth. The only other Big Tech company better at consistent breakage is Google. MS is known for its legendary backwards-compatibility, although sadly is also going in the same direction as the others.
You mean a 64bit architecture literally cannot handle an 8bit instruction as its smallest register is 16bit without swapping modes?
Where did you get that piece of misinformation from? 64-bit x86 can handle 64, 32, 16, and 8-bit data just fine, and its instructions are still byte-aligned.
Also, the decision to remove 16-bit support from 64-bit Windows was entirely political. NTVDMx64 adds it back.
This genuinely just reads like a bucket list of Mac user problems.
The author would genuinely benefit from redhats decade long support models provided his desired tooling works on Linux.
The real concerns are in the SaaS / Remote Resource / Online Activation Space.
eg software that won't wake up if it can't call home, software that has reduced or no functionality if it can't stream remote data, "apps" that are just webkits loading a pre-set url for a completely online service.
As others mention; cloud is a big problem; anything you depend on in the cloud will probably not exist in 10 years in the way you are using it now. It will be switched off, changed (pivoted they call it) or out priced so you cannot use it anymore.
Stop using closed software. Anything open will be around forever and can be compiled using old or new compilers. Mostly natively but if not possible, inside emulators.
Also; Visual Age for Java ; I thought I was the only person in the world who thought it was much better that Eclipse. Seems there is 1 other person.
It gets abandoned but unlike closed software, if you depend on it enough, you can continue using it forever. For some reason if GitHub says ‘updated 6 years ago’ it is not viable for people, but for many things, especially desktop apps, this doesn’t matter at all; it works fine and if you need updates, you can do them.
The point is you can, while with closed software, you maybe can and sometimes with a lot of trouble.
If the source is open, the file format is also open.
I agree, provided the software isn't network reliant / accessible.
If the software were to explicitly depend on <vulnerable library> and moving to <patched library> broke it that would be the nail in the coffin for me / make me question whether I want to spend the effort maintaining the project.
Yes, but for desktop software this is usually less of an issue. And at least you could fix it. When people are running ancient closed source software on Windows (Win 95 software apparently runs on Win 11), they don’t even think about the vulnerabilities and if they do, they cannot fix them. If you are so married to a software like the author seems to be, open is your best bet. And the OP is retired (well, at least 2021), so tinker time!
What concerns me are early 2000s stuff with DRM and server-side content to make preservation difficult if not impossible.
On the other hand there is a game I played on PowerPC macs but the developer released it on github and now it's been updated and compiled for ARM and x86. As long as source is available it's usually not very hard to port to a different architecture.
For example, Unix, Emacs, or Firefox (Netscape) are likely to be quite popular 10 or 20 years down the line. The OS or the editor du jour is not.
This is a good rule of thumb for choosing what to invest your time in.
I always privately called it a "Legacy Tail". It was a factor in the selection of a lightweight markup language to support old-fashioned industry (aerospace), but I think it should be a factor in any software selection, to be honest.
Thanks for teaching me a thing!
There was a lot of nights and weekends spent prototyping and wrestling with various old aero/milspec formats, but at the end of the day we came out with a publications format that could do more than the original could, on common tooling, where everyone who wanted to could contribute. I put it in the "win" column.
They are a lot smaller % than vs code or chrome, but compared to most other software they have a huge user base. They are popular software.
And if the definition is vague enough to include both, you can't use it as an argument to exclude one
Yes, and they are also notoriously unreliable. With KYCd users (which won’t be the case in voting ‘best coding editor’) it is even unreliable, let alone things like reddit polls.
And your last statement is true with just Chrome, so you can add any other browser there for the same effect
But broadly speaking you are right. I run software today on Windows 11 that I first acquired in the 90s. Some of it console based that run under the command prompt.
While Windows is justly famous for backward compatibility though, there are some exceptions.
Firstly 16 bit programs don't run on 64 bit Windows. (Well, unless I fire up a 32 bit Windows in a VM.) This is not a train smash for me, although I did have to replace some -really- old programs.
Secondly, games. MS removed a cdrom driver which a lot of my old games used for copy protection. I'll need to dual boot those into win 7 if I wanted to play them. But the urge to even bother only strikes about once a year.
Regarding word processors, I used Word Perfect in the DOS days and Word since then, and everything still opens fine.
Obviously everyone's path is different, and their experience is different, but if long-term commitment to running existing, and past, software matters to you the maybe select an OS that prioritizes that over one that very much does not.
And DOSBox iirc can run windows 3.x if necessary
winevdm: https://mendelson.org/otvdm.html
I guess that's the downside of vertical integration where the OS is basically just a means to sell hardware.
This is an architectural problem not an OS constraint.
And they're completely correct to abandon this software in my honest opinion as the developers already have.
If you need this run a 32bit VM under the 64bit OS so the virtualised CPUs can handle the 16bit application.
Where did you get that piece of misinformation from? 64-bit x86 can handle 64, 32, 16, and 8-bit data just fine, and its instructions are still byte-aligned.
Also, the decision to remove 16-bit support from 64-bit Windows was entirely political. NTVDMx64 adds it back.
https://learn.microsoft.com/en-us/windows/compatibility/ntvd...
SCNR https://xkcd.com/386/
Break: interrupt (a sequence, course, or continuous state). "this broke the pattern of generations remaining in the place where they were born"
Innovative: (of a product, idea, etc.) featuring new methods; advanced and original. "innovative designs
Ever modified a product to implement a new original advanced method that didn't interrupt a single existing mechanism or structure?
If so, I wish I had your codebase!
Refer to your own link.
The author would genuinely benefit from redhats decade long support models provided his desired tooling works on Linux.
The real concerns are in the SaaS / Remote Resource / Online Activation Space.
eg software that won't wake up if it can't call home, software that has reduced or no functionality if it can't stream remote data, "apps" that are just webkits loading a pre-set url for a completely online service.
Word and LibreOffice work fine on MacOS.
Also; Visual Age for Java ; I thought I was the only person in the world who thought it was much better that Eclipse. Seems there is 1 other person.
The point is you can, while with closed software, you maybe can and sometimes with a lot of trouble.
If the source is open, the file format is also open.
If the software were to explicitly depend on <vulnerable library> and moving to <patched library> broke it that would be the nail in the coffin for me / make me question whether I want to spend the effort maintaining the project.