On 08/15/2015 12:46 PM, Sriram Ramkrishna wrote:
> I'd love to see if someone could write an article on xdg-app for
> opensource.com <http://opensource.com>. The reason I'm not doing it is
> that I want to see if I can reach out such things to other people in
> order to get them invested. Yes, I can write it, but hey, let's try to
> take advantage of opportunities when they come along.
>
> I will personally would like to write an article on extensions and their
> breakage. We will need to be a little careful on this, but since I've
> been pretty much put in charge doing this, I think I can thread the
> needle and come out with a quality article.
(I assume we're talking about xdg-app, in regards to distro concerns)
Before writing an article, we probably want to spend some time taking
inventory of what the concerns are (especially from a distro standpoint)
to ensure we address them.
Yes, some research could be done. I didn't see anything specific from opensuse.
xdg-app was designed in a way to play well with distros, and we ne
ed to
make that clear. I see articles coming out (lwn.net for example) that
imply xdg-app uses Yocto for runtimes. While our implementation used for
testing xdg-app itself does, this is not a requirement or likely to be
how it is implemented in distros.
Sure, it's designed that way, but what is our intent? From a pure technology point of view, xdg-app provides a safe and consistent environment for your application to run regardless of what distro your app is running on.
But politically, we probably need to come up with the right messaging as you say. But the question is what kind of environment are we setting up for these apps, exactly?
So personally, I feel distros are the reason why application writers do not have a personal relationship with their market/users. I've pontificated on this issue a number of times in person already so I'm not going to be much further, but I'm really looking to xdg-app to kindle that personal relationship and help create more apps because there is an incentive to create more users.
My worry is that if we start having branded environments 'opensuse', 'fedora', 'rhel' etc, we're going to be right back where we are. I don't think we need to have distros package apps anymore. A typical flow for a developer should be:
1) design and code the app
2) write the app data
3) choose the runtime/devel for the app depending on platform rather than distro.
4) users install from app store
I think if we start doing branded, it will become again a confusion on deciding 'which runtime is the most popular' again. I don't want to see that. A runtime or sdk should be based on system requirements not what distro you picked.
Otherwise, have we improved anything? Ultimately, this methodology is a win for GNOME and to some extent a loss for distros. But it isn't completely so, as QA for distros will go substantially down as well as the number of bugs found. There can be more focus on hardware and distro engineering at that point which will overall bring quality higher.
sri
-- Christian
_______________________________________________
engagement-list mailing list
engagement-list gnome org
https://mail.gnome.org/mailman/listinfo/engagement-list