On Fri, 2016-05-13 at 09:25 -0400, Matthew Miller wrote:
On Thu, May 12, 2016 at 05:49:03PM -0600, Stefan Nuxoll wrote:
> Regarding the 4th option, would GNOME Builder be a good starting
> point for this? It's still a ways from feature complete, but it's
> best option for an IDE written to the GNOME HIG to follow the
> requirements for inclusion as a default application.
Maybe. I think there's a balance between having something compelling
and new vs. providing a comfortable environment that people can slide
For right now, the development target for Builder is quite narrow - to
be a compelling environment for developing native applications for the
GNOME desktop. And there's a lot of work just for that target - xdg-app
support, UI editor, debugging, and so forth and so on. Making it a
universal entry point for all development on Fedora is completely
For case #4 (large organization developer), Eclipse is probably the
way to go. There's a large community around it already, so we
wouldn't be starting from scratch. It's used by 23% of developers
according to the Stack Overflow developer survey¹.
I think Eclipse should be easily acessible, easy to install, and work
well on Fedora. On the other hand, I'd be pretty hestitant to make it a
focus of development efforts... it's very possible to write plugins
that add user interface to Eclipse. It's much harder to create great
experiences around Eclipse.
Once the produce the full data dump from that survey, we could even
cross-reference the Workstation target audiences to preferred
environment, to make data-driven decisions here.
We need to cater to the diversity that you see in the initial survey
numbers - make sure that whether your preferred environment is Eclipse,
Atom, or Vim, it's easy to get it running on Fedora. That's more
important than figuring out whether we should install Eclipse or Vim by
default (I don't know if you were suggesting that :-)
I think that we can do some things that will make life better for wide
classes of developers - at least any developer who doesn't entirely
live within an IDE.
For example, we can look at doing better integration of git into the
terminal experience, we can make vagrant work better out of the box,
and so forth.
Installing git by default seems to be a simple thing of this class and
I'm in favor of it.
There are also things we should be doing to develop our ecosystem that
aren't necessarily of interest to a wide number of existing developers.
Making GNOME application development better with GNOME Builder is one
of these. Another I think we should be doing is making it really slick
and easy to create applications for deployment on Fedora Atomic Host.