[Info-vax] Creating an open source version of VMS, was: Re: OpenVMS Hobbyist Notification

[Simon Clubley]

On 2020-03-11, Bob Eager <news0073 at eager.cx> wrote:
> On Wed, 11 Mar 2020 14:07:28 +0000, Simon Clubley wrote:
>
>> However, that doesn't change the fact that there is a clean isolation
>> between the architecture specific and none architecture specific code
>> and that Linux has a very nice module plugin architecture that allows
>> all kinds of device drivers and filesystem components to be added and
>> removed cleanly during normal system operation as required.
>> 
>> That same architecture also allows those same components to have cleanly
>> defined boundaries between them.
>
> Not unique to Linux, though. VMS itself, and BSD, come to mind.
>

The modern BSDs: yes, I certainly agree

VMS: no way

Any so-called modular interfaces within VMS have no real benefit
to the end-user. It's one of the things which could be massively
improved in a VMS built using modern concepts.

For example, you can't remove a driver after it's loaded, you can't
plug a new filesystem into the XQP and from what I can tell, anything
you can add via an ACP is an undocumented mess which needs access to
the VMS source code to even begin to understand how to do it.

There is serious potential for massive improvement in all those areas
in any modern VMS.

Simon.

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