Re: Trying to advance fabric ....


Brian Behlendorf <bbehlendorf@...>
 

Great!  Thanks for the attention.

As far as "without a clear decision, voting will help" - I'd suggest that the community avoid too many technical architectural debates for which there isn't overwhelming consensus on what the right thing to do is.  Anytime you have a close vote, you end up with winners and losers, and amongst the growing group of losers there'll be some sense of a project moving too quickly.  Moving fast is important now while the project is still young (and before we have too many production implementations to worry about migrating over architectural changes!) but over time, especially for maintainence branches, a lack of conviction amongst the developers for a change, or even a clear single-person veto, should be a sign to pause and  continue to improve the proposed change for later submission rather than committing it now.

MHO, of course.

Looking forward to Gerrit too,

Brian

On 07/06/2016 03:20 PM, Sheehan Anderson wrote:

I merged a number of PRs today. I am working to clear the backlog. Please note that two maintainers are out this week which will cause a slowdown.

For the earlier PRs

#1719 - Merged
#1722 - Waiting for review from Murali as indicated in the latest comment
#1754 - Ready to merge, but needs to be rebased on master due to conflicts after I merged #1719

I am looking forward to Gerrit. In PRs that are architectural changes, there's often quite a bit of discussion without a clear decision. Voting will help us to determine where everyone stands.

-Sheehan

Brian Behlendorf ---07/06/2016 12:11:38 PM---Thanks for raising the pull request notice here - I'd encourage folks to comment on the PR itself,

From: Brian Behlendorf <bbehlendorf@...>
To: Tamas Blummer <tamas@...>, hyperledger-fabric <hyperledger-fabric@...>
Date: 07/06/2016 12:11 PM
Subject: Re: [Hyperledger-fabric] Trying to advance fabric ....
Sent by: hyperledger-fabric-bounces@...





Thanks for raising the pull request notice here - I'd encourage folks to
comment on the PR itself, bringing it back here if there's anything
controversial.

I think that should also be done by developers for other pull requests
that propose additions or changes to the functionality or protocol, or
even bug fixes with ramifications on performance or tests.  That would
help avoid the issue of things "languishing" in the queue, since few
people are as crazy as I am to subscribe to the Github firehose, and
even with that firehose it's hard to see when something is going
unattended.  The Fabric weekly calls should be an opportunity to also
perform a sort of breadth-first traversal of unattended proposal PRs to
give some quick feedback, to make sure that contributions from new
participants (no matter who they work for) aren't being ignored.

Re the earlier PRs mentioned below - could other developers take a look
and give it some feedback?

Brian

On 07/06/2016 02:56 AM, Tamas Blummer wrote:
> Gabor Hosszu (Digital Asset) submitted a notable new pull request to fabric:
https://github.com/hyperledger/fabric/pull/2101
>
> Its motivation is to move code base toward the consensus architecture proposed by Marko Vukolic (IBM), here:
https://github.com/hyperledger/fabric/wiki/Next-Consensus-Architecture-Proposal
> The proposal was discussed and supported by the Architecture Working Group.
>
> The promise of the proposal is higher scalability through decoupling consensus on order and consistency from endorsement that includes computation of chain code.
>
> This PR introduces a consensus service  API and an endorser API which mark the separation line between these future services. The PR is a viable first step of a refactoring effort that has long way to go.
>
> I hope the PR will not share the fate of our (Digital Asset) other attempts to advance fabric, and starve in the PR queue:
>
> #1754 Add transaction rejection events:
https://github.com/hyperledger/fabric/pull/1754
> #1722 Add noop system chaincode
https://github.com/hyperledger/fabric/pull/1722
> #1719 Do not store failed transactions into the ledger
https://github.com/hyperledger/fabric/pull/1719
>
> The newest PR is #2103 therefore I conclude that 385 other PRs were processed by maintainers in the meanwhile, one of them was ours.
>
> TAMAS BLUMMER
> CHIEF LEDGER ARCHITECT
> Digital Asset
>
> T: +36 1 883 0300
> E: tamas@...
> W: digitalasset.com
>
>


--
Brian Behlendorf
Executive Director at the Hyperledger Project
bbehlendorf@...
Twitter: @brianbehlendorf

_______________________________________________
Hyperledger-fabric mailing list
Hyperledger-fabric@...
https://lists.hyperledger.org/mailman/listinfo/hyperledger-fabric






_______________________________________________
Hyperledger-fabric mailing list
Hyperledger-fabric@...
https://lists.hyperledger.org/mailman/listinfo/hyperledger-fabric


-- 
Brian Behlendorf
Executive Director at the Hyperledger Project
bbehlendorf@...
Twitter: @brianbehlendorf

Join fabric@lists.hyperledger.org to automatically receive all group messages.