On 07/19/2016 04:17 PM, Arnt Gulbrandsen wrote:
> Simon Walter writes:
>> Oh the insolence. Amazing. "You're holding it wrong" comes to mind.
>> There is this guy named Lennart who might agree with you.
>
> Quite likely he might, he's not stupid after all. And I agree too:
> Multiseat is unimportant, barely significant. The price of computers has
> dropped enough that the ones with UIs are now personal devices. The
> architecture of backends has changed such that UIDs aren't used for
> customer IDs.
>
> A few exceptions remain, increasingly rare. I'm sure there are people
> who have used that blue icon near the top right corner of the android
> lock screen.

Disclaiemr:
I don't mind being schooled and I hope I don't offend anyone.

I can remember being told be many seasoned engineers and reading in
several books:
One should not assume how their software will be used.

"Lehman states that the gap between a system and its operational domain
is bridged by assumptions, explicit and implicit [Lehman 00]."

"Usually these assumptions are not documented and often they are not
validated by the people with the knowledge to verify their
appropriateness. Additionally, the real-world domain and the software
itself are always changing. While the initial assumption set was valid,
individual assumptions will, as time goes on, become invalid with
unpredictable results or, at best, lead to operation that is not totally
satisfactory [Parnas 94]."
http://www.dtic.mil/dtic/tr/fulltext/u2/a443152.pdf

It should be obvious to the seasoned developer that s/he cannot know all
the use cases. Therefore by limiting the purpose of their software (do
one thing) and not making any assumptions about how it will be used (do
that one thing well) s/he can make useful software.

I am taking task with the comments about how mulitseat is not important
because it displays a disregard for this wisdom.

One can program a login manager that is only suited for single user
machines. One can program a login manager that allows shutdown and
restart without a password. I think that those already exist. I am not
defending the choice of SLiM.

What I am trying to express is that a narrow minded attitude towards use
cases will make your software brittle.

Since this is Devuan (something about veteran *unix* admins, and coming
from Debian - the *universal* OS), I would not have expected Devuan's
fans and users to be so close minded.

I have seen so many use cases for software that I couldn't imagine had I
not seen them first hand by working in various industries. It has
humbled me.

If individual users are Devuan's main focus, it will fail to attract a
very important and often neglected segment of computer users - business
and industry. Not selling out to business and industry is just as
important. That will also make one's creations brittle.

Kind regards,

Simon