Re: [Hampshire] Application installers

Vic wrote:
>>> Yes, but you're also making assertions about the way things need to be
>>> done when in fact that need is only the result of your own (broken)
>>> rules.
>> The way I read it they aren't his rules they are stuff the copmany he
>> works for does.
>
> "Your" is used in the plural above, thus encompassing exactly that.

My mistake.

>> Maybe we can keep this about the technical stuff
>
> But this *is* about technical stuff.
>
> Creating world-writeable areas for system-wide files is a recipe for
> disaster. It means any one of your users can trash the system for
> everyone, and must therefore be considered a bad idea.

Preaching to the choir. But you could have a separate group that the
program is installed as in a place like /opt/.

For example lets say we have a small research project using the EGEE
toolkit. They specify a whole bunch of exact software versions you have
to have to make their stuff work (beyond your control). All the users
(developers in this case) need to be able to start/stop EGEE daemons and
mess with stuff (this is research!). In this case put what we need in
/opt/, make EGEE group and put the people that need to have access in
that group.

>> It is clear that there is more than one
>> way to skin this fish and we all know the purest technical ideals are
>> sometimes at odds with things like money and time.
>
> But less money and time would be spent using methods which have been
> proven to be correct over some decades, rather than re-inventing the wheel
> in a broken fashion and then having to fix it.

Now when other people are breaking the rules before you get a chance to
have a say. Sometimes we get stuck in legacy situations where the cost
of breaking out and doing it right might actually be more than hacking
bravely onward (especially when resources are limited).

> There is a reason we have tiered privilege on *nix systems. Defeating that
> model might get you immediate gratification, but it costs more in the long
> term.

I'm not sure that statement is true for all instances.


Mat