Linus Torvalds (torvalds.delete@this.linux-foundation.org) on July 17, 2020 9:41 am wrote:
> gallier2 (gallier2.delete@this.gmx.de) on July 17, 2020 3:19 am wrote:
> >
> > And it was a deliberate choice, the Apple II had case sensitive file names, at least with DOS 3.2 and
> > DOS 3.3 (for ProDOS I don't know, never liked it, it was the start of the "sissi-fication" of Apple
> > software, i.e. taking its users for idiots to the point that only idiots will want to use it :-)
> > Ok, I checked. It was ProDOS that introduced the case insensitivity.
>
> The thing is, not being case sensitive actually made more sense
> back in the 80's. So at the time, it was quite reasonable.
>
> On the one hand, it was more work, and CPU's were weak and more work was bad.
>
> On the other hand, it wasn't all that much more work. Back then, case sensitivity pretty
> much universally was limited to ASCII, and being case sensitive in ASCII is very simple.
> Depending on how you do it and your instruction set, it's anything between a single instruction
> for a trivial table lookup and 4-5 extra instructions in assembler.
>
> So it's only when you extend outside of the normal A-Z range that you start getting any real complexity
> at all. Even then, particularly in 8-bit encodings, you can certainly continue to do it badly (but
> well enough that most people won't notice) really easily with just a trivial lookup table.
>
> So even if you didn't limit yourself to ASCII, but had a single 8-bit codepage that you
> cared about (and didn't care deeply about odd cases), it just wasn't complicated.
>
> Filesystems were simpler back then too. A lot of them just upper-cased everything (eg FAT).
>
> And one big change since the 80's is that we now type in filenames less (and those of us that still
> type a lot of filenames probably rely a lot on auto-complete - or maybe that's just me).
>
> It simply made more sense back then to be case-insensitive. You didn't
> have any serious issues, and you had a relatively bigger payback.
>
> These days? The complexity cost has gone up enormously, and the advantages have all but gone away, while the
> disadvantages in other respects have grown. The only real reason to be case insensitive is legacy.
>
> But that is a very very compelling reason. Even to Apple, apparently.
>
> Linus
>

Isn't autocomplete in terminal actually a case where case insensitivity is useful? If I want autocomplete certain file I have in mind, I don't have to remember if I named it with uppercase or lowercase in the first letter.

Remembering names of files just as a string of characters and not as a precise string with exact information about case, diacritics, spaces VS underscores VS dashes and so on is much easier.
I'd say at least for me being case insensitive when operating with the files (but case preserving) matches my way of thinking/expectations more than being case sensitive.