In discussions about archiving data I’ve seen some people say you might have your data, but you may not be able to open it in the future because that file type is no longer supported. I don’t think this is likely. JPG is more than 10 years old and shows no signs of going away. Even video with numerous codecs/containers still supports MPEG. Maybe the only format that has decreased in popularity is ATRAC.
So, have file types gone extinct before?
It might be hard to define “extinct” with regards to data types. Certainly file types can decrease in use, subsequently ease of reading, but a programmer can always write a routine to read or convert any kind of file as long as the structure is known.
That last bit is kind of important, though.
Even if the structure isn’t well defined, reverse engineering can fill in the gaps. Granted, as file types get more complex, that’s more difficult, but not impossible. I myself have dissected unknown structures in order to read data that was obscure. it depends on how valuable the data and/or resource time is.
I’ve written parsers and decoders for file formats long-since abandoned, and for which there are no commercial solutions available any more. Some of them required some pretty brutal reverse-engineering. These were commercial database formats at one time, not proprietary in-house stuff.
I also once had to write a device driver for a proprietary storage peripheral made by a company that had been out of business for 15 years. That was a pain in the ass.
File types aren’t the only problem; media may be a bigger one.
Remember 5 1/4 inch floppies? How about 8 inch? Even if you had one and you somehow knew the data on it was valid, could you find a way to read it? Even if you have an old drive stuck in closet somewhere, are there hardware drivers for it anymore?
And those are relatively recent and common formats. Go back further and you run into all kinds of weird tape cartridges, disk platters, reel to reel formats and who knows what else. If you’ve got actual bits you can manipulate you can probably get the data out of them, but you’ve got to get to the bits first.
Based on what I’ve read, it’s not so much file types, but rather the physical forms of storing the files (floppies, etc.)
Here’s an example that combines both concerns:
Let’s say that you have some sort of legal issue, and I walk up to you and tell you, “Your uncle wrote a letter that explains what happened 20 years ago. Here’s the floppy. And oh, by the way–it was typed on a pre-Windows 3.1 computer.”
How likely do you think it would be that you could access that information?
There are always drivers. I used a few in the day to find obscure files. But also I am assuming the OP is excluding very esoteric files. For example, I just noticed that Minecraft uses .mcr files. I am assuming that they made them up, and that old games/programs have their own personal files.
Some have obviously declined in status. I remember the late 90s, when everybody loved GIFs, and 256 colors was okay. Now they are occasionally used for animation (the best successor, .mng, seems to have not caught on). But certainly JPEG/PNG/BMP are the main formats now.
Are .PIC files still used?
Yes I thought so, but CDs still can be read by Blu-Ray drives, which are 2 generations ahead of it. And I expect USB mass storage to remain in use.
I haven’t seen a .bat file in a looooong time.
Pretty decent, but you’re not wrong there. 
I once had to recover a custom font lost in the bowels of a mainframe using an HP parallel port capture device with a proprietary floppy drive. That was a real PITA, but once it was on my desktop, reconstructing the original data from the raw output wasn’t such a big deal.
NASA has this problem with magtape - stuff gets safely “stored”, but those fridge-sized tape machines were only around so long, and what do you do with a tape reel nowadays?
:dubious:
.bat is alive and well..I see them at least a few times a month in use
They’re still useful if you need to automate, and suck at programming.
I would say that is pretty close to a definition of extinct. You are engaged in Jurassic Park like reconstruction of the basic genome. It’s dead Jim.
Something I have as bit of a collection is a set of old storage devices, and knowledge of people who have other old devices. I even have a paper card reader. Most live on the end of a SCSI bus or serial line, and are not too hard to get working. Sometimes it involves resurrecting a very old computer. I have a menagerie of old Sun hardware. Once you have a Unix box on Ethernet that will talk to the device you are set. But there is a point where the devices won’t work unless coupled to the end to end system, and that gets really hard. But tape was pretty universal, so although the formats might vary, getting the bits off isn’t too bad. (OK, a CDC 7 track tape is going to be a real pain. That might need some custom work.) Old PC hardware that lived on the end of the parallel port is not going to be pretty.
Yeah, .bat files are still around, but when’s the last time you’ve seen a .com executable file?
Look at DRAW. It’s a graphic format that should be deader than dirt. It was part of the RISC operating system that was supposed to compete with Windows back in the eighties. (Spoiler alert: Windows won.) Acorn, the company that designed RISC, broke up in 1998.
But a quick google will show you there are people still writing RISC software, including graphics programs using DRAW files. And acknowledging Microsoft’s temporary lead in the OS field, there are programs that will read DRAW files in a Windows system.
Well, I would call my buddy Christian who has a Sinclair that he had when he was 12 years old … if that doesn’t work, he has every computer he ever owned :eek: so I am sure he could figure out which one he can use to read it.
I have everything I owned from win 3.1 to current, including a compaq with win ME on it with a zip drive 
[which oddly still works. ]
Good god, I still write the buggers. I don’t always have the luxury of being able to install the script engine of my choice at a customer, so sometimes you just have to go native.
.bat files are plain text - they are safe for the long term. This is also why XML (for all its disadvantages) is a good solution for longterm archiving - it is plain text, and structural elements are explicit (ignoring binary blobs stored in XML).
For some media types (3" Amstrad floppies, anyone) the chances of being able to read the data are slim, and then the files need to be reverse engineered if the format is not known. For some file types (fixed record, or primarily textual data), this is not hard, but for others (particularly when compression or some sort of tree structure is used) it is not trivial, and harder if you cannot debug an existing reader.
Even worse are embedding file formats (OLE embedded objects in Word documents, for example). Without an appropriate handler, it is almost impossible to extract the embedded data to try to view/import it. I can imagine some document archives with OLE embedded objects that cannot be usefully opened because there is no combination of compatible tools available to convert/open/display.
Si
To read a pre-DOS 3.1 file, just load command.com (that .com file is still used) and list it, assuming it is a text file. I think even old word or wordperfect files can be read that way if you just omit all the formatting garbage. The point about the physical medium is more significant. My wife still has a computer with a 5 1/4" disk drive as well as a 3 1/2", but in the case of the latter, the disks were utterly unreliable and probably nothing useful is left. I have read that NASA has records of old missions on tape and there is no tape reader for them any more. I’ve not seen an 8" floppy in three decades.
In 1984 I was coauthor of a book that was perhaps the first ever done in latex. The publisher sent me a tape (since it was done on a unix machine and unix didn’t get along with floppies, at least in those days). Parts of the tape were just unreadable. So we took what we could from the tape and retyped the rest. (It is pointless to try to scan mathematics; the formulas and text are hopelessly mixed.)