Re: Ajax PEAR package

Matthew Fonda wrote: > Nothing wrong with packages that are not stable in PEAR, and being > developed in PEAR, but we should only have the *best* implementations of > a package in PEAR, i.e. we do not want to allow any bad code (I have not > looked at this AJAX code, so I am not saying it is bad). > > I agree with Joshua here, I can see that it would be good to have an > AJAX package in PEAR, but we only want the best implementation. As AJAX > is rapidly growing popular trend, it is important that PEAR has the very > best implementation. We do not want any half-ass implementation in PEAR. > (Again, I have not looked at the proposed package, just saying in > general...) This is a problem without solution. Because of the no-overlap rule in package acceptance, there is no way PEAR can guarantee it contains the best implementations for any given package. Packages are evaluated once and only once, at acceptance time. This is so clearly wrong I can't understand how the no-overlap rule is so widely accepted. While I do understand the reasoning behind disallowing two packages for the same purpose, I am of the opinion this is a serious PEAR flaw. This effectively removes competition, which is one of the primary forces behind OSS quality. > I think you should propose your idea in PEPr, but to not rush to call > for votes, allow to be very tested, and agreed upon by everyone. Also, > personally, I do not like the name AJAX (ajax is for cleaning bathroom, > not web), I think a name for this package such as XML_HttpRequest might > be better, or HTML_XMLHttpRequest (as we are js/html here). Some may > disagree on name here though, to some AJAX is fine. Waiting is not much of a solution. Ajax based development toolkits are a dime a dozen today, and it will take some time before the best may rise above the rest. Interestingly, this proves my point. A toolkit accepted into PEAR will probably grow stale out of turf holding, instead of being propelled by PEAR popularity as it should be. Under ideal circumstances, PEAR should be competing to be the cradle of *the* Ajax toolkit, not waiting for a toolkit to become leader and then come into PEAR. Cheers, -- Sérgio Carvalho