Al Sutton wrote:
> My concern is that the the Camera is a "built-in" app, Weather channel
> app is *the* number #1 app in terms of popularity in Market, and snake
> is the number #5 game by popularity, and all three throw up errors like
> this which will make the *average user* feel that Android is flaky by
> association.

That assumes the average user is seeing these errors. I feel like a
broken record, but one or two data points, out of a set of a
million-plus devices, does not make a pattern. They aren't a good sign,
but they aren't evidence of Armageddon, either.

Ideally, apps should never fail. By the same token, ideally, I should
have hair.

> Do we name and shame
> apps to try and get them improved? 

One entertaining-yet-probably-controversial way to do this would be to
bake into Android a top-level exception handler that, in addition to
providing better error messages, attempts to push the identity of the
app and the exception stack trace to a Web service. Said Web service,
perhaps running on App Engine, would make the information available to
the application author...and anyone else.

Average users would not find the information meaningful. Those
interested in a name-and-shame approach, though, could create their own
site to roll up the data to more average-user-friendly details.

This won't work for all situations (e.g., crash while in airplane mode),
but it might work for enough.

> Do we look at improving the OS to
> deal with situations like this in a better manner that doesn't require
> users to be shown error messages they may not understand?

The "deal with situations like this in a better manner" side is beyond
my pay grade.

Better error messages might be useful. Even if the automatic
post-the-exception stuff I list above is deemed too intrusive, it would
be helpful if Android had a send-the-crash-info button to do it more
manually. Developers can put that in their own apps, but if you want it
across the board, it's better to have it in the OS, IMHO.

> or do we
> carry on with the current system where we expect users to educate
> themselves to deal with problems like this?
>  
> To me the last one isn't part of the path to success.

Indubitably.

Here are some other not-mutually-exclusive options:

-- Better educate developers on proper patterns for this sort of thing

-- Better frameworks to help lead developers in the right direction for
this sort of thing

-- If people have figure out a way to get at the error log from
application code on the device (and I forget if someone has a trick for
that), it should be possible to create an AlarmManager-triggered app
that scans for errors and publishes them to a Web service without having
to modify the OS or instrument every app. We'd need to encourage users
to install it, and power users might, giving us access to some level of
crash details.

-- 
Mark Murphy (a Commons Guy)
http://commonsware.com | http://twitter.com/commonsguy

Android App Developer Books: http://commonsware.com/books.html

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"Android Discuss" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/android-discuss?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to