Quality Control Suggestion

Richard Kleeman kleeman at cims.nyu.edu
Thu Apr 27 22:26:15 BST 2006

This is coming from a user who has been involved in 3 releases and 
lodged maybe 50 bug reports.

I understand that Canonical is hiring a QA  person at present so what I 
am about to say might be redundant.

I have found that in testing a release for my operational and productive 
use it is extremely useful to draw up a list of maybe 30 applications 
that are essential to my productivity. I then regularly run through each 
of these apps as the development progresses. The beauty of this is that 
it picks up any relevant bugs in a systematic fashion. I have lodged 
most of my bugreports this way. The developers have been very responsive 
to these bug reports in general.

So here's the suggestion:

Draw up several bigger lists based on extensive community feedback. 
Lists might include

1) Home Desktop User
2) Corporate Desktop User
3) Server Room User
4) Scientific Cluster User

I am sure there are enough Ubuntu users in each of these categories to 
allow for useful community feedback.

Ask for volunteers for each list who have access to a wide variety of 
different hardware types. eg i386, amd64, opteron, nvidia graphics, ati 
graphics etc etc. This is a bit of a crap shoot obviously but a 
reasonably broad set of hardware would pick up most issues and the 
others not picked up would eventually make it to malone anyway.

Ask that volunteers run through a series of very specific tests listed 
on the wiki for each application. Ask that this be done at a regular 
interval during development.

Bribe the volunteers with an appropriate reward mechanism in order to 
ensure a good response.

Anyway hope this might be useful. It is based on empirical evidence from 
a very satisfied Ubuntu user....

More information about the ubuntu-devel mailing list