Re: Quilt reboot, and "compound" projects
I think we need to clarify what, exactly, a compound project *is* before we can say whether or not they make sense as projects.
Does the compound project provide multiple different functionalities that reuse and share a lot of code? It probably makes sense for this kind of thing to be a single project.
On the other hand, what if the compound project is a collection of different code bases for a similar functionality that don’t have any interconnection? In this case, it probably makes sense to keep things separate.
I guess what I’m getting at is that I think it should be the level of interconnection between the components/modules/features/functions that should determine whether some collection of code is one project or multiple projects. We could technically state that, say, Indy is a compound project because it has two “functionally different” libraries Indy-SDK and Indy-Node. But this makes absolutely no sense due to the interconnection of the two code bases.
Does this make sense? It’s been a long day of travel so it’s entirely possible this makes no sense.
I am also super-excited for a new interop proposal.
From: tsc@... [mailto:tsc@...] On Behalf Of Middleton, Dan
Sent: Saturday, June 22, 2019 8:14 PM
Subject: Re: [Hyperledger TSC] Quilt reboot, and "compound" projects
Compound projects feel a bit unnatural to me.
One elephant in the room is whether we can let projects end the lifecycle. I think that’s a really healthy thing for any portfolio.
I’m not so sensitive to the potential loss of name recognition. In this particular case, I don’t [yet] see an interest from the Quilt maintainers in progressing the ILP code. To me that signals the project is at its natural conclusion.
My view of Ursa is that we are still in the storming stage. I think as a team we are getting closer to its definition, but we shouldn’t point to it as a template of a compound project right now. In fact, if ursa turned into a bag of disjoint parts I don’t think it would be successful, but that’s a different thread.
A lot of what defines a project to me is the maintainers and contributor community around it and their shared drive to create a coherent piece of software. If we sideloaded a second code-base with a different set of developers that doesn’t feel like a unified project to me. That feels more like a branding decision that I think is overthinking our incubation stage.
(and to be clear I am very interested to see a new interop proposal and evaluate it on its own merits)
Intel Principal Engineer
From: <tsc@...> on behalf of Vipin Bharathan <vipinsun@...>
If it is different from ILP, what is it? Will there be a doc describing it much like a project proposal. Any technical papers? Is there code you would like to bring in?