On Thu, Oct 24, 2013 at 8:52 PM, Adam Williamson <awilliam@xxxxxxxxxx> wrote: > On Thu, 2013-10-24 at 19:56 -0400, Josh Boyer wrote: > >> It is up to each WG to determine their product requirements. That >> includes which architectures and target users they are trying to >> produce a product for. >> >> > We've done a lot of work over the last few cycles to really bump ARM up >> > to 'first class citizen' status, and a lot of that is coming together - >> > I think reasonably successfully - in F19 and F20. It would be rather odd >> > to go with a change for F21 or F22 which goes in the opposite direction. >> >> ARM is important long term, yes. I don't necessarily think that ARM >> is equally important across all of the existing products. I find it >> more likely that ARM is important enough to have it's own WG and it's >> own product, which may or may not have commonality with the other >> products. > > I'm not entirely sure that makes sense; it seems to be a conceptual > error. ARM is an architecture. In practice, at present, the > ARM-architecture based hardware we support mostly falls into a certain > category that kind of naturally lends itself to a particular kind of > product, but that seems a transient scenario, not a permanent one. > Looked at conceptually, it doesn't make any more sense for there to be > an 'ARM working group' and an 'ARM product' than it does for there to be > an 'x86_64 working group' and an 'x86_64 product', but those are, I > think, prima facie absurd. The concepts of 'working group' and 'product' > have been drawn up along broadly _functional_ lines, and a 'working > group' or 'product' for a specific system architecture doesn't really > line up with that design. > > I think the approach I implied in my email - making sure the functional > WGs and products we are inventing do not neglect any of our primary > architectures and use cases - is the correct way to go. You make a good argument but we probably do need to account for the case that a particular product might not be suitable for every architecture. I was thinking more along the lines of every primary architecture has to have one or more of the core products but should not be required to have to them all. How to accomplish that is rather fuzzy. John -- devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
