Re: PEAR mirror

"Tim Perdue" <tperdue@valinux.com> wrote.. > How far away is PEAR from reality? What would it take to > make a mirror of it on PHPBuilder? > > I think this makes more sense than re-building the library on > PHPBuilder. If you are considering re-instating a code library on PHPBuilder I would like to make a couple of comments, which don't necessarily relate to a PEAR mirror. Firstly, the old library on PHPBuilder was never very comprehensive, and some of the code was of dubious quality and utility. I think you are right to consider a new library with clear coding standards, however they are enforced (whether by dictate or peer feedback). Secondly, the old library on PHPBuilder was a repository for the code itself, rather than just an index. If I write a short script and go and submit it to several code repositories, then I find and fix a small bug or enhance the code in some way, am I going to go to the bother of going round all the places I submitted it and submit the new version, even if I can remember where I submitted it, and even if the repositories provide an update method? I don't think many people do. A comparison can be made with the very successful cgi-resources.com which acts as an index primarily for Perl scripts. The index has basic details of each script and provides the searching, cataloguing, feedback and ratings systems, but it does not store the scripts themselves. It just points to the author's own web site. That way the author can maintain the latest version of a script in one place and the index doesn't have to be concerned with keeping up-to-date with releases. Zend has the right idea with their "Applications" section (not the "Code Gallery") but they do not have a large enough library, or a sophisticated enough ratings system. The value of cgi-resources.com is that I can go to a topic of interest, see a dozen or more scripts described, and quickly see from the ratings and the detailed user comments which ones are the winners. In short, it is the searching, cataloguing, feedback and rating systems that represent the added value of the index, rather than the storage and distribution of the scripts themselves which is trivial. These are the features to concentrate on in a new library. Commiserations, by the way, on the lack-of-backup fiasco. I can tell you I made damn sure I had all my local and remote systems backed up immediately after reading your post.