IBM stopped their support for LTFS, unfortunately. Which I kinda want to use despite all the drawbacks. And well, the RAM disk workaround does the trick, so there’s not enough pain to compel me into investing more time into the issue.
Oh, there can be all kinds of uses.
For example, I own an LTO tape drive. Unfortunately, I couldn’t get the IBM drivers to compile on my particular flavor of Linux, but I do have a dual boot system.
Now, Windows is a shitty, shitty OS that I only boot up when I really need to. It interferes with all kinds of stuff in ways I hate, for example - copying files. It just refuses to read large files in a continuous, reliable data stream without any interruptions from my SSD when backing them up to tape. This causes the LTO drive to slow down, speed up, rewind, which is not a good thing because it causes additional wear on the tape.
Fix: Create a large RAM drive, copy files to RAM drive, run backup with RAM drive as source.
Built my new PC in late 2023… so glad I put 96 GB of RAM into it, despite several people asking me why the fuck I need so much RAM…
Well… Kingda Ka is said to have cost $25m. Using that as a ballpark figure.
I want wooden coasters, steel coasters, hyper coasters. Other attractions, too. Gotta buy the land, build the infrastructure around the park, employ lots of people, build hotels, gastronomy… a billion might not even be enough for everything.
Build my own theme park with all the roller coasters I want.
Lossy audio compression algorithms work based on psychoacoustic effects. The average human ear will not detect all the “parts” in a lossless signal - there are things you can drop from the signal because:
So in order to determine exactly which parts of an audio signal could be dropped because we don’t hear them anyway, they measured a couple of thousand people’s listening profiles.
And they used that “average human profile” to create their algorithm.
This, of course, has a consequence which most people, including you apparently, do not understand:
The better your personal “ear” matches the average psychoacoustic model used by lossy algorithms, the better the signal will sound to you.
In other words, older people, or people with certain deficiencies in their hearing capabilities, will need higher bitrates not to notice the difference. In the 90s, I used to be happy with 192 kbps CBR MP3. But now, being an old fuck, boy, can I hear the difference.
Ironically, I can detect the difference not because my ears are “trained” or “better”, I can detect it because my ears are worse than yours!
So the whole bottom line is this: While it may be true that you, personally, do not require lossless to enjoy music to the fullest, other people do. Claiming that lossless isn’t needed by 99.9% of the population is horseshit and only demonstrates that you have no clue about how lossy compression works in the first place.
I’m a software architect. I work with large amounts of data. And a sizeable RAM disk is generally useful for many purposes. And yes, I also run AI locally, though that’s what my RTX 4090 is for.