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

Greg Beaver wrote: > Sergio Carvalho wrote: > >> However, I resent the -1. XML_RPC2 is driver-based. Nothing stops you >> from writing a new backend. In fact, I'd appreciate the help in writing >> an xmlrpc-extension-based backend. Logically, this would be the best >> backend for performance conscious users. > > > You are missing the point. Your code is unnecessarily complex. That is > always a problem. You also don't seem to be aware of the xmlrpi > extension and the fact that API compatibility is impossible if you use > all these XML_RPC2_Value_* objects. I'll bet you a lunch in the best restaurant here that API compatibility is possible. As for complexity, I can and did justify the existance of every class you questioned. XML_RPC2_Value classes are type classes. While you can do away with them, you'll then have to refer to types using enumerations as a linear classification system, unable to deal with hierarchies (like the one rooted at XML_RPC2_Value_Scalar) > >> I'm also starting to feel you have an unfounded ill stance against this >> package. I've lost many hours writing and documenting this. The least >> I'd expect is that people don't suggest using SimpleXML to *write* XML. > > > You seem to forget that I've spent many (if not more) writing the > original code this was based on plus the now quite old Chiara_XML_RPC5, > which is being used by all Chiara_PEAR_Server implementations with > absolutely no reported problems whatsoever. You continue to refer to Chiara_XML_RPC5 as if it were in PEAR. Is it? Will it ever be? Will you force PEAR to do without XML_RPC support for PHP5 because you personally favour a solution found elsewhere? Isn't there a conflict of interest? > In fact, all I see is that you've ignored all of the negative comments > from very experienced developers from last November when you originally > proposed it. No I did not. The original discussion is here: http://marc.theaimsgroup.com/?l=pear-dev&m=109940004317627&w=2 http://marc.theaimsgroup.com/?l=pear-dev&m=109970648507437&w=2 I discussed and reached for a consensus, found here: http://marc.theaimsgroup.com/?l=pear-dev&m=110001310605761&w=2 http://marc.theaimsgroup.com/?l=pear-dev&m=110002379218044&w=2 I implemented what I designed then in response to problems you posed. > I have serious doubts that you will ever listen to them, You'll hardly find more package developers here doing what I did, redesigning a package so that it could be of use to you. > and that is why I registered a -1. I note that you registered an active vote against it, not a +0 or a -0. You actively think XML_RPC2 is not interesting to PEAR users. > In my opinion XML_RPC2 is not yet > ready for acceptance into PEAR. But fail yet to explain why. You defend a procedural design, while I defend an OO design. We can do religious battles until the end of times. I propose we stick to these evaluation principles: 1) Does the package do what it announces? 2) Does it do it efficiently? (numbers please, no gut-feelings) 3) Can it accomodate extension for predictible further uses? > > I don't sense any spirit of collaboration here. Geez. I rewrote a package to fit your proposed API!!! > You took code I wrote, > re-wrote it in a way that doesn't make any sense to me, made a proposal > without consulting me, and called for votes without consulting me. If > this package is accepted, please remove all references to my name so > people don't get confused about who wrote and is maintaining it. I don't mind removing your references. I am quite troubled by the fact that you took the ideas that we launched last November, put them to use in a project incompatible with PEAR (license-wise) and now based on that project results, refuse to accept the package into PEAR. Oh, and by the way you didn't consult me either (not that I think you should have). > > Greg Sérgio Carvalho