Winged Monkey

Matt Wagner matt.wagner at redhat.com
Mon Oct 29 19:00:10 UTC 2012


On Thu, Oct 25, 2012 at 08:51:28AM -0400, Greg Blomquist wrote:
[lots of snipping below]
> I'm not, beyond a shadow of a doubt, convinced that the right answer
> isn't to drop this into Conductor.  OK, one more time in English: I
> can understand wanting to drop this into Conductor, or spend the time
> revamping the "user" side of Conductor.  I'm also not convinced that
> won't eventually happen.

This is an interesting point -- even if what you do ends up being
wholesale duplication of what Conductor provides, but in a cleaner form,
there's value there. (And I don't think what you're doing will end up
being wholesale duplication anyway.)

What I don't want to see happen is for us to either (a) neglect
Conductor's UI because Winged Monkey will make it better, not (b) put
substantial effort into the Conductor UI that turns out to either be
replaced by Winged Monkey, or to make Winged Monkey unneeded. Put more
succinctly, I think we should make sure that our UI/UX work and Winged
Monkey work are complementary, rather than duplicates.

> But, I see this as an opportunity to explore some possibly very 
> complicated user experience ideas geared specifically to non-technical
> users in a more free-form environment.

Yes, I think that would be fantastic.

> Re: the admin stuff.  You're right, we're studiously brushing that 
> under the carpet in these early days of exploration.  However, I'm 
> definitely keeping the concept of how to provide good admin 
> capabilities as a main topic.  The main idea I keep circling around
> is how to tie in Conductor's cloud administration.  For all the of
> seemingly superfluous pieces in Conductor, I think that it has
> managed to correctly capture the difficulty in administrating
> connections to multiple cloud connections.  A feat I don't think
> anyone wants to attempt to repeat.

I guess part of the problem is that building/pushing images requires a
lot more setup in some cases than just using a provider. Building for
RHEV, for example, requires us to set up NFS mounts.

> Re: all the great ideas swirling around in your head.  Yes!  This is
> precisely what we want to talk about.  We want to mine all the ideas
> of end user cloud interaction with the primary focus being on the 
> non-technical user.

Yes, definitely. But what I find particularly remarkable is that some of
these ideas are things multiple people seem to have described
independently, which is what makes me think that there's something
really compelling in here.

> VerySoonNow(tm), we want to start setting up G+ hangouts, and 
> getting all the ideas out in the open.  Jeremy has done an amazing
> job capturing the concepts we've thrown around, and synthesizing 
> them into some fantastic wireframes.  But, I'm sure there are other
> ideas out there.  I feel like we've only just begun exploring this
> concept, and there's still quite a bit left to figure out before we
> can really start diving into producing a piece of software that fits
> the bill.
> 
> But, that's not gonna stop me from coding. :)

I guess what concerns me is that it's not clear what the scope of this
is. As a way of testing out ideas, I think this is fantastic. My
question is really just whether it makes sense as a stand-alone project.

-- Matt



More information about the aeolus-devel mailing list