[Info-vax] CRTL and RMS vs SSIO

[Simon Clubley]

On 2021-10-15, Arne Vajhøj <arne at vajhoej.dk> wrote:
> On 10/14/2021 10:10 PM, Lawrence D?Oliveiro wrote:
>> On Friday, October 15, 2021 at 2:36:27 PM UTC+13, Arne Vajhøj wrote:
>>> On 10/14/2021 9:11 PM, Lawrence D?Oliveiro wrote:
>>>> And VMS can do the same thing. All the compatibility stuff can live in userland.
>>>
>>> I am having a hard time seeing: VMS scheduling, VMS memory protection,
>>> VMS clustering, ...
>> 
>> None of which are relevant, because Linux already offers better functionality in all those areas.
>
> All of which are relevant, because without them it would not be VMS.
>

I wonder if Lawrence has compared VMS clustering functionality with
the clustering functionality offered elsewhere. The gap has got closer
in recent years but there's a reason why everyone looks to VMS as the
model to emulate when it comes to clustering.

>>> ... ODS-2 & ODS-5 ...
>> 
>> New filesystems don?t have to be implemented as kernel modules, they could be done via FUSE.
>
> Maybe.
>

FUSE filesystems come with performance limitations on Linux.

Besides, running a driver in user mode sounds like the kind of thing
that a microkernel would do. I thought Lawrence didn't like that kind
of thing ? :-)

On a more serious note, adding a new filesystem to Linux as a kernel
module is a much easier thing to do on Linux than on VMS. Having written
Linux kernel modules in the past, I don't see writing a filesystem as
a kernel module as being much more difficult than as a FUSE module.

Simon.

-- 
Simon Clubley, clubley at remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.