PEAR Midgard wrapup

First of all, I have to thank everyone for the kind advise and apologize for not being there on at the end of the thread. The Midgard dev community has probably to reflect a bit on what to do. I am very keen on PHPizing or PEARizing Midgard, but I have to confess some confusion. Now, to narrow things down a bit, I would like to restate the purpose: the Midgard community is developing many libraries and subsystems, and feels PEAR (repository + tools) are a good and nice wheel we don't want to reinvent. We also want to PHPize ourselves, and integrate better with the community. Installing the Midgard framework itself is not a high priority today. According to PEAR-dev, Midgard does not belong in PEAR as it is a complete application, and PEAR is focusing on libraries or components. Apparently the Horde project have tried 1st, and this thinking in PEAR-dev emerges from that discussion. However, reading back on PEAR history, I learned that PEAR means "PHP Extension and Application Repository" <http://beta.usephp.net/article.php3?id_article=17> . Maybe "extensions" are welcome at this stage and "applications" will be welcome... at some later date? Apologies if the last paragraphs sound a bit unhappy. But I must say that I am confused by the grounds for the refusal. Some of Midgard's code is so ugly I was expecting to be rejected on the grounds of messy code, not inappropriateness. Bah! guys like me, we are are _never_ happy. It was suggested that Midgard belongs to PECL, but that seems to be targetted for C modules. I have not been able to find the PECL guidelines either, so I am not sure if this is accurate. At any rate, our target is to be installable using PEAR tools, so that begs the question: are PECL modules part of the same repository and easiliy installable using _pear_ (the CLI tool)? Of course, we can serve our packages from our servers -- we don't even need a repository as the pear tools will take a valid URL, get the package, and install it. However, this negates many advantages, including dependance resolution, publicity (just by being alongside in the same repository), search functionality, etc. I still think using the PEAR tools is a great advantage when compared to reinventing them. Certainly a lot better than storing code -- libraries in particular -- in database. So I would like to invite opinions from fellow Midgardians on the path forward for our libraries, like NemeinAuth and others. I would like to thank the PEAR-dev list again. I have to guess that the decision not to allow complete applications is related to the early phase PEAR is in, and that complete applications and frameworks will be welcome at some stage. The Midgard community will be keen, and everyone will benefit. cheers, martin