|Re: [config-dev] [External] : Re: Tree structure vs flat structure discussion thread|
This ballot has been started to unblock us. We went in circles the last 3-4 meetings with ZERO progress. I am fine to start (actually restart) the use cases discussion. To start the conversation, can you prepare the use cases and drive the next meeting? It would be even better if you write it down and post it to this mailing list or in GitHub discussions (https://github.com/eclipse-ee4j/config/discussions) before that.
My impression is that this was a change calculated politically rather than technically, but I may have misinterpreted it. It doesn't really matter though: we're going to continue to go around and around without accomplishing anything until we can finally stop arguing about implementation and start discussing the actual use cases we want the APIs and SPIs to support.
I for one do not have infinite time; in fact I have many responsibilities that are not related to this specification. Spending even just an hour every week getting nowhere is not a good use of what little time I have available. We should have been much farther along by now, but every week we're almost starting over again, because every time someone new joins the group we have to re-justify what little consensus exists. Why? Because we don't have a solid foundation of documented use cases. All we have is arguments over including this code base or that code base, and "my implementation already does this" and so on. I for one don't have time to spend on this kind of discussion, nor to work with others who think that this is what specification authorship is.
We need to stop treating this like a competition between N different APIs and start treating it like what it is: a specification. You can't have a cogent specification without starting from use cases. It is simply not possible. And it is very clear that we do not have consensus on use cases. But rather than working towards consensus on use cases, we are for some reason using code dumps as a proxy. My +1 vote was not a vote for the code or for anything the code implies. It was a vote to stop this pointless time-wasting argument so that there is at least a *chance* that we could start discussing use cases.
I will be using my vote and the votes of anyone I can possibly influence to block further progress beyond this point until we discuss, document, and agree upon use cases. If this does not happen, then I will probably withdraw and recommend to Red Hat to make it company-wide, and recommend others to do the same.
On Mon, Dec 13, 2021 at 5:55 AM Roberto Cortez <radcortez@xxxxxxxxx> wrote:
- DML • he/him
Back to the top