2009/1/16 Andreas Jaeger email@example.com:
What use cases does Novell want to reach? What should the user be able to do with SUSE?
- Portability/Laptop/Nettbooks and so on
- Connectivity of devices
- Support for syncronizing devices
- Home server setup, mediaserver, DNLA, UpnP, file , print and so on.
I think our problem is that we want to reach all of these at the same time. ;) openSUSE is a desktop distribution with server functionality that can be used by developers. In practice the primary focus is on desktop but we distribute packages for the rest as well.
There is a mission statement, which is "create best Linux distro, foster innovation, work together with upstream and make users happy" from the project guidelines, working in an open manner.
openSUSE has different needs:
- we use it as basis for our enterprise server and desktop products
- we want to grow the openSUSE community
openSUSE is used by the community nowadays in all the different scenarios you mention above - and you'll surely find one vocal user for each of them. ;)
Should we focus the distribution more? And what do to with the other use cases?
Removing or just de-emphasising areas of traditional support, would likely be very divisive. Folk get upset when only the "Other" desktops or window managers are available.
Perhaps it's useful to think of 3 pillars of user base :
1 - distributing and stabilising, upstream software, developing inhouse projects 2 - stabilisation and development towards SLE (including ISV projects) 3 - growing Linux market, with a reliable, easy to manage full featured Desktop
At the moment, OBS is really aimed at addressing 1, though reading Factory list archive suggests, difficulties on 1
The release engineering, distribution of openSUSE and post-installation support and fixing, addresses 2 though I'm sure SLE engineers would love having more problems solved, up front by competent enthusiast contributions
But 3 is IMO let down at present, by releasing prematurely, and long term benefit decisions with short term costs
If you spend some time, trying to field Installation questions, from new users, it becomes obvious that error notifications deemed in Bugzilla "cosmetic" erode confidence. Large numbers of known bugs in heavily used systems, are then hit over and over. The Internal developments are often let down by useability nits, sometimes just labelling of buttons, or font sizes, which the open process ought to benefit from early feedback, without expensive "focus group" testing.
There's a real conflict of requirement between the "Bleeding Edge" and "It Just Works".
Are there ways, of having the cake and eating it? Perhaps using stable fall backs, and having "Probationary" upgrades. Or do you really need a watefall model of improvements, trickling down, and finally having very widely tested good stuff frozen for SLE?