Criterion revision proposal: KDE default applications

[Kamil Paral]

> Given that, I propose re-wording as follows:
> 
> "All applications that can be launched using the standard graphical
> mechanism of a release-blocking live image after an installation of that
> image must start successfully and withstand a basic functionality test."

I have read the whole thread, multiple things are discussed here. My take on this:

1. I agree with the proposed criterion. Our team is inherently best-effort driven and we need to very carefully and cleverly prioritize. This helps us guide us in the right direction.

2. Some people claim that DVD and Live install result should be the same. Adam's proposal doesn't contradict that, just gives us a working solution before such state is reached. Once relevant people implement the changes in DVD or Lives, the "live image" part of the proposal becomes redundant and can be removed.

3. As long as DVD and Live install differ, I agree that the Live package set contains the most important packages from the Spin's view and those should get the most testing. Our website heavily promotes Live installs, and most of the general users do Live installs, because they use our website (if you read this, you're not a general user). The other packages, only available on DVD, are clearly not that important. That doesn't mean we can't accept a proposed blocker in such a package, if it is especially important (eat-your-computer type), just the "basic functionality" criterion won't apply automatically to them.

4. As for DVD existence - yes, I also consider it outdated. It tries to solve bandwidth problems in a wrong way. I imagine a world with Lives for general audience, netinst for expert audience, and multi-Lives for marketing purposes. The bandwidth problem should be solved by a simple program:
a. you run it on computer A (lacking bandwidth) - it gathers the list of installed packages and exports a file
b. then you run it on computer B (good bandwidth) - feed it the file and tell what programs you want to download, it fetches all the RPMs and makes a repository
c. then you bring the files back to computer A, and either using that file again or GNOME Software it mounts the repository and allows you to install the programs available
The whole program is trivial (even its GUI), and it solves everything DVD attempts to solve. You can use it any time, not just at the release date (it downloads fresh versions of programs, DVD only contains stale ones).
With this approach, we can also make a special-purpose media for special occasions - want to ship apps for kids in Africa? Just download the RPMs based on the package set of a default installation, copy/burn the repo to the media and ship it together with Lives. Since there's nothing else than a yum repo, it's trivial to check for correctness. No QA needed.
We can make a different set of this extra offline-access media for Africa kids, for Graphics Designers conference, for Gamers festival, for Scientific classes in universities etc. Do you see the large possibilities that generic DVD can never help with? And hey, we can even put the Live installer and this additional repo to the same medium as a second partition, no need to have two media. So you stick it in, boot Live and install, and then you reboot to a working system, stick it in and have additional software available.