Hi,

On Thu, Oct 1, 2015 at 4:32 AM, Eric Auer <[email protected]> wrote:
>
> as mentioned earlier in this thread, generic operating systems
> are allowed to implement VFAT (FAT32 and LFN) without fees, so
> I would not be too worried about those two more years.

I seriously think you're wrong here. But I'm not a lawyer (and all
this empty talk solves nothing).

> Regarding LTOOLS quality: This thing is ancient and of course it supports
> only ext2, not ext3 or ext4. It is probably better to use the
> source of VMSMOUNT as basis for TSR drivers for exFAT access :-)

ext2 might be harder to support, but at least it wouldn't go near the
stupid proprietary hell-fire.

> Note that HPFS is owned by Microsoft (written while OS/2 still
> was by MS and IBM together) and HFS+ is owned by Apple, so if
> you want to be totally safe license-wise, you should use ext2.

HPFS is from 25 years ago, long since unpatented. Plus, it's
compatible with OS/2. Like I've said before, Linux and FreeBSD both
(allegedly) have support for it as well. Even that is "better" than
FAT.

I don't know if I really recommend HPFS over ext2, but literally
anything is better than yet-another-proprietary garbage.

> I would not use more modern Linux filesystems, as you probably
> need a TSR of several 100 kB to support those ;-)

So?? Let them write a pmode TSR that requires SSE 4.2 for all I care.
If it works, it works. Granted, I don't recommend anything beyond 486
DX, but how is "something" worse than "nothing"??

> I agree that OpenWatcom would be nicer than Borland C++ here.
> Not sure if Pascal would be nice for TSR drivers: You can do
> it (see keyboard driver [FD KEYB]) but the extra effort is that you have
> to translate existing filesystem code which probably is in C.

I've never tried, but I'm pretty sure that Turbo Pascal can link to
Turbo C code.

At least FreePascal (32-bit host) uses "register" calling convention
(like Delphi) by default, but it can mix and link with "cdecl" code
too (and others). The 16-bit DOS target is still unfinished, only
supports "pascal" calling convention (AFAIK), for now. The v3.1
snapshots have internal tools, so no external NASM nor WLINK needed.
Not sure how to (easily) mix C with that (yet), but it can't be
impossible. I assume manual linking is still possible, among many
other things, but the very few people working on it are doing so in
their spare time, so (understandably) it's not moving very fast.

> Remember that the Samsung code carefully got GPL-ed together
> with some licensing experts after the initial accidental code
> publication. You can have open source drivers for things that
> need licenses to use, those things do not exclude each other.

It's very naive to want to play with fire here. I would not recommend
going near it. Even without knowing the details, I assume that there
are dozens of other (free) file systems that would work every bit as
well. At minimum, my wimpy suggestion would be to investigate ext2 or
HPFS first. (Or get it in writing from the Microsoft legal team that
they won't sue you and yours if you all use, study, modify, and/or
redistribute the aforementioned "GPL" code.)

> In any case: Do not panic. At the moment, you are insisting to
> deprive DOS of a feature that Linux has had for years. In spite
> of having a much larger market share, nobody sued Linux either.

Linux is very popular but also sloppy, so that doesn't mean they are
in the clear. The FSF does not recommend most popular Linux
distributions because they are "non-free". It goes well beyond just
simple software licensing (and leaks into firmware and other
complications).

> If anything, you may worry about having to get a license if the
> DOS driver gets used in some embedded / appliance device, such
> as a camera. Even then, the camera manufacturer would have to
> get the license. The blame would still not be on DOS itself.

Offloading all potential problems to someone else is anathema to "free
software". If you can't even guarantee the freedom of what you're
redistributing, then it's not free at all.

I'm not trying to suggest that all proprietary software is evil. But
it's very destructive and tiring having to deal with thousands of
licenses (and arguments, wars) just because people are egomaniacs who
can't get along. From a purely pragmatic standpoint, the less to worry
about the better. (Of course, "do without" isn't a good solution to
everything either.)

------------------------------------------------------------------------------
_______________________________________________
Freedos-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/freedos-devel

Reply via email to