Re: Re: [PEPr] -1 for XML::XML_RPC2

Hans Lellelid wrote:

Lukas Smith wrote:

Err, all he said is that some of the comments from people he respects sounded like valid criticism. As we want to ensure BC as much as possible its not necessarily wise to go +1 when in doubt.

BC? I don't think this package is aiming to provide BC for XML_RPC ... is it?

once its in then we have "maintainers freedom" which states that its Sergio's decision of what to do technically within the PEAR CS and PEAR QA requirements. And once he makes a stable release there is no way to change certain things either.

I know that Daniel said that other people's opinions sounded valid, but those other people have their own votes. I'm suggesting that a -1 is a pretty strong statement to make without having even looked at the the sourcecode. Do we really disagree on that topic?

We disagree. PEAR is about high quality. While we do appreciate people making the effort to bring things in (as you can see there is a discussion to fix things that pear-dev's think is broken), the default rule is to rather not allowing something in.

Err, he doesnt have time to check the packages. Therefore a +1 would be very wrong. However he did have time to check out the criticsm made on this list.

Right ... my point exactly. Those people expressing their opinions on the list had their own votes. I personally think that people should vote based on actual package review, not second-hand information & opinion. Perhaps PEAR Group doesn't feel that this is important.

Opensource is alot about trust. Daniel trusts certain people to not come up with statements that are based on false data. He looked at the statements and found them to be valid. Since he trusted that the data that made people come to these statements is not invented he voted -1. You are turning a mouse into an elephant. Actually I dont even see the mouse.

Ok ... I know that you disagree with me by default :) but I'm really just trying to say that I think PEAR is better than this, that this seems like a popularity contest & personal problem rather than an objective evaluation of sourcecode. I believe packages should be

Its not a popularity contest. In that case Alexey would never have made it into PEAR (sorry Alex .. you are just the perfect example here). Its about code and to some extend about trust in developers you have come to know for quite some time. You havent used the "e" word .. yet .. but it sounds like you are driving the PEAR elite argument again. With 500 developer accounts eligible for vote, there is plenty of votes to go around with.

accepted based on quality, usefulness, appropriateness, etc. I am simply suggesting that there should be some expectation that people voting on packages have looked at the code -- and, sure, have also followed discussions about them. I also would have thought that a guideline about this would be unnecessary; it seems rather obvious.

No we dont expect people to always look at the code. Its just too time consuming. We have a select box in the voting form to indicate the level of review. You could be basing your decision on the summary, the API, the code, the examples, other peoples comments you trust etc. Note again: Daniel was not just multiplying previous votes either. He said he read the discussions and just felt to vote -1 until proven to be soundly implement. regards, Lukas