On Thu, 2009-04-09 at 08:38 +0100, Mark Shuttleworth wrote:
> Peter Whittaker wrote:
> > The OS is different - when the computer needs
> > to tell me something, I probably shouldn't ignore it. Think firealarm.
> >
> Yes, that's true. But there are some apps that have fire alarms too -
> imagine, for example, an app which monitors your RAID array and
> alerts you to failures and issues.
Ah! Now you are thinking like a technologist and not a user,
differentiating between the OS and applications based on whether they
are kernel or user space, as opposed to what they actually do!
>From the perspective of many - I would think most - users, something
that monitors RAID - or monitors anything about the state of the
computer for that matter - is not "an application", it's part of the
computer. In that view, "applications" are things users start to get the
computer to do what they want to do: email, IM, edit video, etc.
> My point is that we should have clear guidelines about what
> constitutes each mode of operation, and we should follow those as
> rigorously with the OS as we would hope apps do, as well.
If you read my comments on the wiki page, you'll see I make clear
distinctions between system and user notifications: Let's apply that
distinction consistently. A user notification is a doorbell, it appears
in the user notification area, near where the user keeps their stuff. A
system notification is a firealarm and appears in the system
notification area, where other "systemy" things are.
Both use the same underlying technologies - API calls that bring up
bubbles - but the presentation is different: firealarms and doorbells
are presented differently and both are different from normal apps.
> As soon as we make exceptions for ourself we are weakening our
> argument
And foolish consistency makes a poorer system. Think back to the lack of
PageUp/PageDown under OpenStep on the Next boxes, because Steve Jobs
held the view it was a screen, not a page. So arrow keys giving you a
line at a time were the only way to scroll.
Jobs was being 100% consistent in his view, and missing the point that
ScreenUp/ScreenDown would have been handy.
> I *think* we're in agreement here :-)
Actually, not so much. Hence all of the on-going debate on this. It's
one thing to see a design, another to see the implementation. Some of us
saw the design and said "cool". Many of us saw the implementation and
said "Whoa, not cool".
On Thu, 2009-04-09 at 08:38 +0100, Mark Shuttleworth wrote:
> Peter Whittaker wrote:
> > The OS is different - when the computer needs
> > to tell me something, I probably shouldn't ignore it. Think firealarm.
> >
> Yes, that's true. But there are some apps that have fire alarms too -
> imagine, for example, an app which monitors your RAID array and
> alerts you to failures and issues.
Ah! Now you are thinking like a technologist and not a user,
differentiating between the OS and applications based on whether they
are kernel or user space, as opposed to what they actually do!
>From the perspective of many - I would think most - users, something
that monitors RAID - or monitors anything about the state of the
computer for that matter - is not "an application", it's part of the
computer. In that view, "applications" are things users start to get the
computer to do what they want to do: email, IM, edit video, etc.
> My point is that we should have clear guidelines about what
> constitutes each mode of operation, and we should follow those as
> rigorously with the OS as we would hope apps do, as well.
If you read my comments on the wiki page, you'll see I make clear
distinctions between system and user notifications: Let's apply that
distinction consistently. A user notification is a doorbell, it appears
in the user notification area, near where the user keeps their stuff. A
system notification is a firealarm and appears in the system
notification area, where other "systemy" things are.
Both use the same underlying technologies - API calls that bring up
bubbles - but the presentation is different: firealarms and doorbells
are presented differently and both are different from normal apps.
> As soon as we make exceptions for ourself we are weakening our
> argument
And foolish consistency makes a poorer system. Think back to the lack of
PageUp/PageDown under OpenStep on the Next boxes, because Steve Jobs
held the view it was a screen, not a page. So arrow keys giving you a
line at a time were the only way to scroll.
Jobs was being 100% consistent in his view, and missing the point that
ScreenUp/ScreenDown would have been handy.
> I *think* we're in agreement here :-)
Actually, not so much. Hence all of the on-going debate on this. It's
one thing to see a design, another to see the implementation. Some of us
saw the design and said "cool". Many of us saw the implementation and
said "Whoa, not cool".