Re: priorities for PEAR project as a whole

On 2006 09 27 15:52, Greg Beaver wrote:

Hello all, I want to re-iterate the priorities for PEAR as a project that I brought up last spring, and see if we might be able to pull a few of you off of other projects for a short while to expedite this process. Summary: #1: make the PEAR website PEAR-installable, instead of syncing from CVS #2: add package groups for the website and for installer REST #3: any other changes, political or technical Long version: What problems do we have in PEAR at the moment? First, there is an awful lot of bickering over how we're going to do the PHP5 thing in PEAR. Second, almost all recent political decisions have not been clearly documented on the website. Third, we have lots of packages and very little oversight and cooperation between developers to do QA and assist each other with details. Most of these problems go back to the website, and the way that it directs the political and technical discourse in PEAR. Right now, it is hard - and also dangerous - to update the website, because it is synchronized directly from CVS. Hence, priority #1: #1: make the PEAR website PEAR-installable, instead of syncing from CVS. This will allow us to try out some experimental things, and also make it much easier to QA the website prior to a public release, spurring innovation and stability. Part of that innovation involves a technical change to the website that will enable a better political dialogue in PEAR: #2: add package groups for the website and for installer REST this will allow us to do a lot, especially with separating older packages from newer, and even grouping related packages in intuitive ways beyond static categories. #3: anything else deciding on the exceptions usage, re-organizing PEAR group - all of this can wait until we have a more dynamic front page and presence. Comments? Thanks, Greg

Man, go away for a couple of days and look what happens! As a relative outsider, I have to say that the web site has been a huge disappointment. Sure, it's my fault for not digging deeper, but it wasn't until I used the web based installer that I was forced to look at all the capabilities in PEAR. Granted it's grown a lot over the past two or three years, but during that time I've developed implementations of a number of packages that duplicate PEAR functionality that wasn't there before. Duplicate effort really irks me. Instead of inventing my own wheels, maybe I could have been working in a team to help build PEAR packages. There's a few things I'd love to see from the web site: * multiple views of packages, by title, by keyword (or tag, if you prefer), an overview list that includes package status, a big page with a short abstract of every package. * Something that lists what people are working on or want to work on, no matter how informal. Maybe some process where an idea can incubate from idea to concept, to development team, and finally to released packages. * User comments in the documentation, like the main PHP site. There are hundreds of developers out there with something to contribute. They should be able to have a voice without mastering docbook. * Some way of allocating effort for large scope projects. I'd love to help transition a package from PHP4 to "native" PHP5 (which in my book means far more than just E_STRICT), but I don't want to spend weeks on something just to discover that there are three others doing the same job -- I'd rather work closely with those folks to make the job easier and quicker.