On Fri, Jul 9, 2010 at 6:45 AM, Timothy Murphy <gayleard@xxxxxxxxxx> wrote:
Robert Myers wrote:As the OP, I didn't "demand" anything.
> What the OP asked for
> (demanded, really) is not possible even in theory. Enormous amounts of
> time
> could be expended, and programs *with* status indicators will still have
> long, inexplicable pauses if something unexpected happens.
Also I didn't say the program hung, as you seem to think.
I said it _seemed_ to hang, ie there was no response for a long time.
Have you actually used preupgrade yourself?
If so, what was the length (approximately) of the longest pause
you experienced?
I don't think it takes an "enormous amount of time"
to tell the user what is happening.
Most programs do this.
In fact, one of the long pauses seemed to be
at the point where yum says "Updating <package abc>",
going through the packages it has downloaded, one by one.
I have used preupgrade multiple times on multiple boxes. I also just installed a bunch of other software that involves long downloads and installs. I'm so accustomed to inexplicable stutters and stalls and things going wrong with long installs that I just ignore them. I find other things to do. Progress indicators can stall for no apparent reason, and there is no surefire way to tell if the install is dead or just taking a very long time. Firefox has this wonderful indicator that someone else aptly referred to as the pizza wheel of death. If at first you don't succeed, try, try again.
I have a computer that's so much faster than what the software designer of one package anticipated that informational messages appear and disappear too fast for me to read or to act on.
The range of hardware that's out there is so enormous that there is just no way to anticipate what all users will experience, unless you are Apple and keep control of everything.
How about progress indicators that race up to 98 or 99% finished and stay there seemingly forever? What good does that do? Maybe I could have found a better way to say it, but, while it may be possible in some limited circumstances to reassure the user that something reasonable is happening, it's a much harder problem than you apparently think, and, as I said, it's not even in theory possible to address in general.
Robert.
-- users mailing list users@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines