Do recommendations !! use your influence to get more organisatio

Hi all hi manuel

I think you being even too careful. It's good to tell people not to follow blindly every influencers, but in the same time, as you mention, php needs some organisation.

In this view, may I make a critic to phpclasses.org ? Have you noticed that loads of database wrappers, templating engines, and cache engines have been published here ?

Its good for people to do such scripts when they improve their skills, but I am convinced it is really misleading to some newbies that will use poor packages because they are not really able to judge and choose them. Also it increase the fragmentation of the php community : If if everybody uses a different template, cache, db wrapper, then sharing code becomes almost impossible. Is that a good thing ?

Of course you do by yourself what you now you can do well, but what about people with lower skills ? the meaning of phpclasses.org is sharing code, isn't it ?

Why not go into a improved quality and sharing process ? I think you are well qualified and recognized now and you are in position to coordinate the definition of some recommendations for the publication of new classes. This way, the different frameworks would have a way to know how to integrate well the packages that are published here, and the publishers of new classes will have a way to write code that is compatible with most framework.

Wouldn't that be a nice php world ?

I'm pretty sure the guys developing the different framworks would be happy to discuss with you and everybody would gain from that.

thx for reading.

olivvv

Your comment is a bit odd. Let me explain.

The PHPClasses site was not meant to find only the best classes for each purpose, but rather to give every author the opportunity to get their work exposed and eventually get some feedback or recognition.

As I mentioned in the post, PHP developers are not droids, they have feelings. If I excluded anybody for eventual lack of "quality", I am sure it would break the hearts of many authors.

This is a democratic site. Everybody has an equal chance to expose their work, regardless whether your work is very sophisticated or very simple. I prefer to let the users decide which are the best classes for whatever purposes.

On the other hand, there was a lot of concern and some initiatives were taken to help the users to find what can be the most recommended classes, according to most of site the users, thus in a democratic way.

What is odd in your comment is that these initiatives do not seem to have been noticed by yourself and probably by other users.

Maybe it is my fault and there is some problem in making such initiatives visible enough, so the users can understand that they were meant to help them to find the most recommended classes.

Let me be more explicit and tell you about those initiatives:


- Top rated classes by group

For each class an user downloads, he can rate it in several aspects: utility, consistency, documentation, examples and unit tests.

The class ratings are processed on a daily basis to compute a ranking that takes in account the average overall rating of all classes.

On the page of each group of classes there is a link named "Top rated classes". This link leads to a page that displays all classes of that group sorted by their position in ranking.

There are also links to these top rated classes pages in the listing of groups on package pages.

This top rated pages should help the users to pick which are the most recommended classes.

I wonder if you have not noticed these top rated pages. If you have not noticed, maybe there is something wrong in the site presentation or in the layout of the links that point to these pages.

In that case, something needs to be done to improve the visibility of the top rated class links.


- Innovation award

Every month the most innovating classes published in the site are nominated to the innovation award.

The idea is to encourage authors to focus on contributing classes for innovating purposes or at least implement innovating approaches to problems also solved by other classes in the site.

Basically, it should award authors that contribute classes which are not just more of the same.


Despite these initiatives do not stop authors to contribute classes for the same purposes, at least it should help users noticing classes that deserve being noticed.

Maybe these initiatives are not having that effect, as much as I hoped. In any case, I am open to any suggestions to make these and other initiatives more effective and helpful to the site users. Such suggestions should not excluded any author and stop them from sending whatever classes they want to share.

You're right, I was a bit excited while writing my message, and it may sound a bit abusive.

My point is not to forbid anybody to publish their work, indeed. My point is that in php, code sharing became difficult just because everybody choose template engines, db wrappers, cache engines, etc... that have just petty differences in their performances, but that are enough different to require code to be rewritten when ported to another application.

Now, between the super-organised frameworks and the democratic beautifull mess, I think there is a room for code following simple recommandations that will make sharing and community effort possible.

I think this room could be built on top of phpclasses. I think this recommandations/specifications could be made by you, the main framework authors, and some others.

olivvv

As I mentioned, I only use my own classes, so that is all I can honestly recommend. It may look biased but to be fair I cannot recommend classes that I have not tried.

Even if could recommend something that I did not write nor use, it would cause resent among all the authors of all similar packages that I would not recommend. I would rather avoid hurt the feelings of excluded authors.

Furthermore it could make some of the excluded authors to campaign against the PHPClasses.

It is already a bad that there are authors that excluded themselves from sharing their work in the PHPClasses for whatever reasons. For those I cannot do anything because they were left out upon their own choice.

As for the authors that participate in the site, I would better preserve them and avoid causing any sort of resent.

I know that, as an user, you would like my recommendations, but I hope you understand why I prefer to not take part in any selection.

I know the idea to let the users rate the classes and build top rated charts, is not exactly what you would like, but I hope it can be almost as useful as the recommendations that I or any restricted group of people could make. After all it is community effort.

Hello everybody


Manuel I've read the whole post you've made which I consider to be very helpful particularly because you share with the PHPClasses comunity some of your own experiencies which I believe is the spirit of this site. To share packages that are the result of each author's individual experiencies and to increment their own developer experience by taking constructive feedbak and inproving thair packages and class development.

I believe though, that you've missed the point of Oliver's last comment. I don't think he meant that you, and the rest of the most experienced and well known developers in this community should judge the classes and recommend one over the other. What I believe he meant is that, due to the enormous ammount of packages that relate to the same aspects of web developing, and the vast experience some of the most frantic testers have acquired, some guidelines should come in handy in order to "standarize" (please forgive my awful english) and make packages more compatible with one another.

I wouldn't ever consider one database wrapper to better than other, but I do realize that perhaps for one particular problem or requirement one would be more accurate than the rest.

The chaotic way in which packages are being publish today makes cmpatibility and flexibility almost imposible. More than once I've found myself switching from one downloaded class to another an, for example, building a "wrapper to wrap a wrapper" becase of the really different interfaces I found myself dealing with.

So, what I believe, Manuel, and, what I believe Oliver meant, is not that some developers should be discouraged or others should be more encouraged than the rest, not even that there should be any sort of "filters" for tha class that get published. What I believe is that some standard lines should be more or least mantained or at least considered when a developer sits down on his pc and starts a new package. I believe that, in that way, not only you wont discourage developers to publish, also, more classes will be downloaded and tested because it would be easier to apply a different implementation for a solution to a similar problem


Matías

Yes Matías you got my point

the vocabulary I used wasn't the right one. I think it should have been "specifications" instead of "recommendations".

Manuel you noticed that I'm not a publisher on phpclasses. Still I write and I'm able to write some code, but the code I would have published here would have been "Yet Another Bla blah" not really better than what existed before, with just some little customization for my own very particular needs. I didn't found the way in phpclasses to take part in a collective effort.

That is the reason why I went toward a framework (namely cakephp) where even with limited skills I feel able to contribute in useful way. Also I am more trusting the code I can found there and I have the feeling that I can get some help if there is any problem.

In no way I am claiming that frameworks are better than phpclasses.org they are just different and not comparable things. My little dream is the link between them.

olivier

ooh and yes I think that community rating has definitely its place in such process.

Yes, you are right, I was misunderstanding your suggestion.

Actually, I mentioned in the post that PHP needs specifications for different types of frameworks, pretty much like Sun specifications for several Java frameworks. It would be a sort of PHP Community Process, similar to Java Community Process.

I think the biggest challenge is to make the most important PHP players sit down and agree on something.

If there is no general agreement on specifications, there is not much point of just a few of us try to push any specifications at all.

There are several factors like developer egos or implicit commercial interests of certain companies that get in the way.

Trying to get all or most interested parties to agree on anything may be a long shot, but if there is some interest, depending on me I can cooperate if I see there is enough good will.

ooh and yes I think that community rating has definitely its place in such process.