Stephan Assmus wrote:
> Hi,
>
> from my personal impression, the problems are clear. We have some likely
> solutions. I'd say we force *both* category and importance level to be
> given by the application (the more information, the more appropriate the
> server can handle it). Anything beyond that can be easily defined after a
> first implementation is there (like which importance levels, how to call
> them, which categories...). I don't anticipate a very complicated
> implementation with regards to the actual service, so changes should be
> relatively easy. The most complicated part of the thing is likely the
> settings panel.
I agree the thing should not be too hard - the settings panel is
definitely important.
Still, InfoPopper is BSD/MIT according to BeBits - that would provide a
good starting point I'd imagine.
> So, IMHO (of course), we should pick up discussion after Ryan has got
> something to show, unless someone has an idea which has really not come up
> yet at all. I'm just saying it, because *anything* can be misused, and I
> think we have found a reasonable approach. At a certain point, you have to
> let go and trust some responsibility with the user. If a developer misuses
> this service badly, you can always just deinstall the app or write an email
> to the developer. I mean there are ways, but we shouldn't let ourselves
> cripple our creative power by trying to forsee *too* much.
Just to make clear - the category approach I was thinking of would not
have any fixed categories, the application would be free to categorise
as the developer wishes. That gives a lot more flexibility than, say,
Windows Event Viewer.
Importance is useful on Windows Event Viewer as it is really a
replacement for a log file - you have a lot of noise ("information"
messages such as "Service x started successfully"). The importance field
is there to make errors and warnings stick out.
The thing I wish to avoid is having a low-importance category that may
encourage developers to send messages like "New window opened!".
However, I have just thought that importance could be useful for certain
situations - "Meeting in 5 minutes" being the same category yet more
important than "Meeting in 30 minutes". I feel the range of useful
importance values is pretty small (I don't think a discrete pop-up is
sensible for critical things, and is annoying if overused by many
unimportant things) but perhaps there still is value in having it.
All of that rambling leads me to agree with you completely Stephan -
let's wait for the first implementation! That wasn't going to be my
conclusion until I started writing this, but I'll send it anyway...
> Best regards,
> -Stephan
Simon
