Re: Re: Publicly listed channels (was [PEAR-DEV] Considering package proposal:

The points you have listed below point to a slightly fuzzy logic, in terms of registering a channel. PEAR as a group of people with lots of varying opinions, may not be the perfect enviroment for deciding who should be added, and who should not. The current voting solved a similar issue about deciding which packages should be added, One idea is to perhaps use it to propose a channels, in that it would be clearer and the comments about the channel at proposal stage may give people an idea how the channel is regarded... However, I guess some would see it as if pear was 'recommending' these channels by listing them on pear.php.net. (although our current list has sort of crept in over time, without much thought) Another idea is that either channels.php.net or www.php.net/channels.html would, like the links page on www.php.net, be a neutral place to list them, there is no implied recommendation. It would also free up the decision making for the process.... and also enable the listing of projects that most of the pear developers would never go near, yet benefit PHP in the bigger picture... The reality is that you, or someone is going to have to commit for a while to maintaining the list as well. (but I guess that was assumed.) Regards Alan On Mon, 2005-04-11 at 23:13 -0400, Greg Beaver wrote: > Alan Knowles wrote: > > >I guess it's a double edge sword, > >+ channels = promote other locations with quality php code > >- channels = promote other locations with crap php code. > > > >Do we promote phpNuke, phpBB channels.. ? > > > > > Perhaps, each channel that has a huge following should be seriously > considered. > > >Do we promote myFirstPhpProject type channels ? > > > > > Absolutely not - PEAR should only list channels that have an established > userbase. Paul Jones has demonstrated that it is possible to create a > userbase out of thin air in 6 months to a year, so I don't think this is > too much to ask. Compromising on this point would be suicide for PEAR > and lead to your "- channels" point. > > >How do we decide if a channel is good, bad? > > > > > Each channel should be considered independent of any other. I would say > that pear.php.net probably should not list channels that are gray areas, > but instead wait until they get out of the gray either way. > > >Does the channel page end up looking like hotscripts/phpclasses? > > > > > No. If used properly, there will be a few important, large sites that > also have channels. These sites should be listed through pear.php.net. > I don't think it will be hard to tell which these sites are, as they > will be blogged about by independent parties, have high rankings on > sites that track popularity, and active user forums of some sort. > > >I think the idea about pointing to a single (or maybe later two or > >three) external site(s): phpcodechannels.net or something? which can > >decide it's own listing rules seems more sensible. > > > >PEAR is more about a repository of quality code, which is peer reviewed, > >and maintained.. - not really the place to do a hotscripts type search > >for projects... > > > Channels are independent of pear.php.net. I agree - pear.php.net (PEAR) > is a repository of quality code. That will not change at all. > > Let me be more clear. The first year or two of channels, I can't > imagine that more than 10-20 channels will fit the bill of being an > active channel. The only channel that fits the bill today is Stephan > Schmidt's channel. Even my own channel at pear.chiaraquartet.net is not > a mature channel. It works, but does not have a community base behind > it, and only provides 1 package. > > Channels listed at pear.php.net should be established as important > centers of PHP activity prior to being listed. Niche groups can easily > publicize their channels to their niche followings. > > My point is that PEAR (as in pear.php.net) people do not need to judge > the quality of code, only the quality of their channel support. > > 1) can you successfully install/list/download packages from the channel? > 2) is it obvious that the code has a large following amongst PHP users, > and is here to stay (think Horde, for example)? > 3) do they take care of their own QA (have documentation, bug > reporting/feature request centers/forums for users?) > > These two questions are really enough. > > Greg >