Published 2004-06-10 01:04:53

Having been involved with PEAR for quite a while, and a member of the much critisised PEAR group. I have to admit I was really impressed with Hans Lellieds comments today. The pear-dev mailing list has evolved over time to occasionally become a very intellectual conversation location.

We are all very invovled in the disucussion, is adding another template engine to PEAR a good idea?... All of the current engines are very much branded, IT, ITX, PHPlib, Sigma, Xipe, Flexy. To a certain degree, many of them provide the same functionality.

An ideal goal would be the provision of a core engine, that provided the features of all of them, yet had the overhead of none of them. Of course I have a biased viewpoint, having written one of them, but I would be extremely happy to depreciate Flexy, in favour of a neutral core API .

As an asside to this, the question should the PEAR group veto Savant?, two of the members have voted -1 on it, and Hans's excelently written email infers that there is something wonderfully great about a benovolant dictatorship. I've still mixed feelings on this. This overwhelming veto power should not be used without serious consideration of the concequences, and adding yet another template engine to PEAR, while not exactly desirable, may not be up to the barrier of justifying the use of the veto.


A
Mentioned By:
google.com : december (64 referals)
google.com : april (52 referals)
blog.akbkhome.com : AKBK home - Smoking toooooo much PHP - A Vision for PEAR (26 referals)
google.com : pear flexy (12 referals)
google.com : compare pear template flexy it sigma xipe (8 referals)
tools.search.yahoo.com : Translated Version of http://www.akbkhome.com/blog.php/View/39/A+Vision+for+PEAR.html (5 referals)
google.com : pear teens (5 referals)
google.com : Flexy-Tenns (4 referals)
www.procata.com : Professional PHP (3 referals)
www.procata.com : Professional PHP » 2004 » June (3 referals)
www.dogpile.com : Pear Sigma Template - Dogpile Web Search (3 referals)
google.com : Hans Lellelid home page (3 referals)
google.com : november (3 referals)
google.com : pear template (3 referals)
google.com : pear Template Sigma (3 referals)
google.com : " Hans Lellelid" (2 referals)
google.com : best pear template engine (2 referals)
google.com : flexy tenns (2 referals)
google.com : pear flexy inclusion (2 referals)
google.com : Sigma ITX template pear (2 referals)

Comments

The thing about PEAR that we don't have a lot of but i think is its greatest power is that it has the infrastructure and the visibility to support multiple maintainers for a single package. Anything that leads to more of that and less of 10 similar packages where 5 are moving towards deprecation because people don't have time forever is a good thing.

Not being in the PEAR group i would just say don't use a Veto unless your willing to deal with the enourmous flame it will start.
#0 - Joshua Eichorn ( Link) on 2004-06-10 02:09:01 Delete Comment
(benevolent dictatorship)++

I don't think the flame is a problem. The PEAR group has made itself ultimately responsible for the quality of PEAR, and as such it is paramount that they do what they feel is best for what they have defined as PEAR.

However, having said that I have to agree that on the larger scope, if Savant gets accepted the world will continue to turn. I don't feel Savant is important enough to warrent using your Veto power and have to cope with a lot of unhappy developers.
#1 - Aaron Wormus ( Link) on 2004-06-10 03:10:54 Delete Comment
Yes, I'm the troublemaker. Didn't know you were in PEAR Group, Alan, I guess I should pay more attention. :-)

We (well, I :-) did not sign up for dictatorship, benevolent or otherwise. We signed up for peer-review according to published coding standards under a voting system. If PEAR Group wants to dictate, then voting means nothing, and peer-review is reduced to "does PEAR Group like it?".

What PEAR Group needs to do is open their hands, not clench fists. The community is speaking, and the Group should listen.
#2 - Paul M. Jones ( Link) on 2004-06-10 03:32:32 Delete Comment
Paul, I'm very happy that you stirred the pot :) the questions that Savant brought up certianly needed to be answered.

As far as the dictatorship goes, with projects that I take responsibility of I consider myself the (benevolent?) dictator. It is the only way that I can ensure that my vision is completed. I don't know that it's better for the community, but it's certainly better for seeing a vision fulfilled.

I agree about the vote though, it's gotta be one way or the other.
#3 - Aaron Wormus ( Link) on 2004-06-10 03:58:40 Delete Comment
I'm glad Alan thought my email was well-written (thanks, Alan). I don't know to what extent I am a dictator in the projects I manage, but I think from experience that the only way to have a really coherent set of classes with the range of functionality provided by PEAR is to have someone (or Group) ultimately committed to maintaining that vision.

Otherwise it's sourceforge, or better: CPAN. I don't think there's anything wrong with that model myself, but it is not a model with vision. So if PEAR wants vision I think it needs to make the Group more powerful & remove some of the voting practices that just frustrate people (like Paul) who then expect the open CPAN model.

I like the quality of the code on PEAR. I disagree with things, but you'd be hard-pressed to find such high-quality code anywhere for PHP. I think the vision would be great. More like the Foundation Classes; done right, PEAR could really be a component library that underlies every big PHP app.

Hans
#4 - Hans Lellelid ( Link) on 2004-06-10 04:34:43 Delete Comment
Another approach to handle this might be the re-introduction of the "PEAR Foundation Classes" concept. I haven't heard from it for a long time now, so I guess it was dropped?

The concept of having foundation classes with approved quality that follow even higher standards than the rest, or, as Hans puts it, "fit into a bigger picture," would allow the PEAR Group to act out kind of a dictatorship just for the inclusion in this elite set of packages.

The rest would just be open to all who are willing to follow the PEAR guidelines as they already exist today.
#5 - Markus Wolff ( Link) on 2004-06-10 06:43:29 Delete Comment

Add Your Comment

Follow us on