Go PHP 5

[Eric]

I, as well as many open source developers, have joined the Go PHP 5 initiative. Starting February 5th, 2008 - we will be dropping support for PHP version 4, and will be using PHP 5 instead. My Domain Name Portfolio project is one of many going in this direction.

Making this post here, to encourage any of you that have open source projects, to join this initiative.

You can find the official press release here (posted on my blog - cause the original is a PDF which a lot may not be able to view).



:kickass:

 

[Peter]

This could be somewhat of a hindrance for non free softwares to be honest at present statistics for php 5 show less than 20% take up as of yet. This means that you will alienate 80% of potential users.

Free open source applications will also suffer as there may end up being a drop in user take up which may result in less people developing for the software.

http://www.nexen.net/chiffres_cles/phpversion/17283-php_statistics_for_june_2007.php#majeure

 

[Eric]

PHP 4 is now 7 years old, PHP 5 has been in use for 3. It's time people move on to PHP 5, free software or not. That's what this initiative is for; to encourage hosts AND developers (not just open source developers) to ditch PHP4 and move on to PHP 5.

 

[Marck]

PHP 4 is now 7 years old, PHP 5 has been in use for 3. It's time people move on to PHP 5, free software or not. That's what this initiative is for; to encourage hosts AND developers (not just open source developers) to ditch PHP4 and move on to PHP 5.

If I have an app that uses PHP4 and works fine, why should I even spend a moment to consider changing it? That would be nuts.

 

[Eric]

If I have an app that uses PHP4 and works fine, why should I even spend a moment to consider changing it? That would be nuts.

http://buytaert.net/php-is-dead-long-live-php
http://www.php.net/ChangeLog-5.php
http://www.php.net/manual/en/migration5.php

 

[Marck]

http://buytaert.net/php-is-dead-long-live-php
http://www.php.net/ChangeLog-5.php
http://www.php.net/manual/en/migration5.php

So, in other words, no good reason? We do not develop on a whim.

 

[Eric]

So, in other words, no good reason? We do not develop on a whim.

After reading those (which I doubt you even did) how can you say there's no good reason?

The new features, and OOP are good enough alone. There have been many performance enhancements, not to mention the bug and security fixes. PHP5 has taken care of nearly ALL of PHP 4's shortcomings.

Even if you don't like this initiative to drop PHP 4 altogether, making code PHP 5 dependent, atleast switch to PHP5 on your server(s) You'd be surprised of how well your scripts will run on a PHP5 server, compared to running on a PHP4 server.

 

[Marck]

After reading those (which I doubt you even did) how can you say there's no good reason?

The new features, and OOP are good enough alone. There have been many performance enhancements, not to mention the bug and security fixes. PHP5 has taken care of nearly ALL of PHP 4's shortcomings.

Even if you don't like this initiative to drop PHP 4 altogether, making code PHP 5 dependent, atleast switch to PHP5 on your server(s) You'd be surprised of how well your scripts will run on a PHP5 server, compared to running on a PHP4 server.

If a script is in production already, it does not need additional language features. Can I switch to PHP5 with no loss of service? (register_globals = on, etc?)

 

[Eric]

If a script is in production already, it does not need additional language features. Can I switch to PHP5 with no loss of service? (register_globals = on, etc?)

If you need register_globals you shouldn't be developing anyway :imho:

register_globals has been disabled by default since PHP 4.2.0

 

[Marck]

If you need register_globals you shouldn't be developing anyway :imho:

register_globals has been disabled by default since PHP 4.2.0

If you need hype maybe you're the one that shouldn't be developing?

I have apps that predate PHP 4.2.0. If they are written well, why wouldn't they last 6+ years? Why not 20?

 

[Eric]

If you need hype maybe you're the one that shouldn't be developing?

I have apps that predate PHP 4.2.0. If they are written well, why wouldn't they last 6+ years? Why not 20?

I know plenty, and I do not need the hype; I'm where I need to be.

No application that relies on register_globals can be considered "written well". And, 20 years? PHP 4 will probably hold it's marketshare for 2 more years at most. And in that 20 years, PHP 5 will be obsolete, in favor of PHP6 - where register_globals, magic_quotes, etc are removed from the codebase and disabled completely.

 

[Marck]

I know plenty, and I do not need the hype; I'm where I need to be.

Let's be friends.

No application that relies on register_globals can be considered "written well".

Any application that performs it's function for 6 years with no problems or bugs is written well.

And, 20 years? PHP 4 will probably hold it's marketshare for 2 more years at most. And in that 20 years, PHP 5 will be obsolete, in favor of PHP6 - where register_globals, magic_quotes, etc are removed from the codebase and disabled completely.

I have misgivings about a platform that achieves market share at the expense of longevity and backwards compatibility.

In 20 years my PHP4 app will probably still be functioning and doing it's work.

 

[psalzmann]

In 20 years my PHP4 app will probably still be functioning and doing it's work.

No. In 20 years you'll be laughing at how much of a learning curve developing a php script actually was and would be thankful how things in 20 years have changed.

I doubt php will be around that long (but we'll all still be able to use it), there will be something else that will take it's place. Programming languages change. Perl was the best, now it's seeing the face of PHP and what it offers people like you and I. For every 20 lines of perl, PHP can perform the same, if not faster reduced in 5 to 10 lines.

Ruby is a newcomer as well. I guess for every 10 lines of PHP code to perform an action, ruby can do it in 3 to 5.

20 Years is a long time. According to google, that is precisely:
7 304.84398 days

Things change. Period.

If you need register_globals you shouldn't be developing anyway :imho:

register_globals has been disabled by default since PHP 4.2.0

Then why did the developers of PHP include it in? What is your take on register_globals SecondVersion?

 

[e39m5]

How do I go about dropping support?

 

[Eric]

No. In 20 years you'll be laughing at how much of a learning curve developing a php script actually was and would be thankful how things in 20 years have changed.

I doubt php will be around that long (but we'll all still be able to use it), there will be something else that will take it's place. Programming languages change. Perl was the best, now it's seeing the face of PHP and what it offers people like you and I. For every 20 lines of perl, PHP can perform the same, if not faster reduced in 5 to 10 lines.

PHP will remain, and I have no problem in saying it will still be the best in years to come.

Ruby is a newcomer as well. I guess for every 10 lines of PHP code to perform an action, ruby can do it in 3 to 5.

Eh, don't think so. I've taken ruby code (20-30 lines+) and done the same in 10 or less with PHP - guess it depends. I hate ruby. :P

20 Years is a long time. According to google, that is precisely:
7 304.84398 days

Things change. Period.



Then why did the developers of PHP include it in? What is your take on register_globals SecondVersion?

A terrible mistake, is why they included it. That's why it was disabled in 4.2.0, and is being completely removed in 6.0.0-dev and later.

I can't stand anyone / any script that uses register_globals - poor coding habit, not only security-wise, but also shows laziness - if you depend on it. (to be blunt) :imho:

How do I go about dropping support?

If you have a read over at www.gophp5.org, it will tell you how

 

[Marck]

No. In 20 years you'll be laughing at how much of a learning curve developing a php script actually was and would be thankful how things in 20 years have changed.

I doubt php will be around that long (but we'll all still be able to use it), there will be something else that will take it's place. Programming languages change. Perl was the best, now it's seeing the face of PHP and what it offers people like you and I. For every 20 lines of perl, PHP can perform the same, if not faster reduced in 5 to 10 lines.

Ruby is a newcomer as well. I guess for every 10 lines of PHP code to perform an action, ruby can do it in 3 to 5.

20 Years is a long time. According to google, that is precisely:
7 304.84398 days

Things change. Period.



Then why did the developers of PHP include it in? What is your take on register_globals SecondVersion?

In 20 years my PHP interpreter will work just as well as it does now. There are still perfectly good FORTRAN binaries today as well.

For every 20 lines of perl, PHP can perform the same, if not faster reduced in 5 to 10 lines.

That is not factual. But I don't wish to get into a religious war. Let's just say that Perl lives on doing what it is best at. IMO data processing, unix administration, and networking. (Would you write a webserver in PHP or Ruby? You'd write one in Perl. (Webserver, not website))

I think a trend we will be seeing in the next decade is platforms that support any language you like, so these arguments become a thing of the past.

 

[neroux]

Its interesting how this thread turned out.

As the numerous installations show, PHP 4 isnt bad at all and does its job very well. However its a fact that PHP 5 is now already available for three years. While a new version does not always necessarily mean better performance it usually however does. In this case the most obvious improvements were done in the OO part. If you write primarily procedural code you actually wont notice a big difference, from a developer's point of view. However even then an upgrade is recommended due to the many internal improvements of the core.


Concerning register_globals, they were a fundamental feature in the early versions of PHP and I would not agree that code relying on them cannot be called "well-written". However we now have for more than five years better ways to access this data and so I would actually expect code to use the super globals. If it doesnt it either has not been updated for 4 to 5 years or is actually "not well-written".

 

[Dave]

Wouldn't a better incentive be to guide users into the path of the wonderful Python and Django. :bingo:

Ok seriously this is fantastic news. I have been using PHP5 for like 2 years so its about time everyone else catched up. :hehe:

 

[nicholas]

Progress is progress (go PHP5). But there's always room (and in many cases, a need) to keep the old (Cobol and FORTRAN being prime examples, which were supposedly long dead languages 20 years ago, and still running in many data centers to this day).

Obviously those are different situations and PHP4 isn't exactly a legacy language, nor will it be anytime soon, but like CSS being all the rage of late as more and more designers and programmers celebrate its virtues, this initiative for PHP5 may take the imagination of the world by storm soon enough, as long as it isn't forced at the expense of PHP4. And that may be the reason it hadn't taken off stronger than it has yet.

There are many merits to championing PHP5, just as I personally do; but you must also know that like many other examples before (Intel being successful with its CISC chips due to backward compatibility even when RISC was thought to be superior; or Borland's fall after trying to rewrite dBase in the superiority of rewrtiting it in OOP at the expense of time and competitors catching ground; etc.), staying with the current popular standard may not really be that bad either.

 

[monaco]

I don't think anyone is talking about going through all your code and rewriting everything...but why not use new language features when coding new features in your applications?

As for an application working for 20 years...I doubt that. PHP will do something down the road to the language spec that will make it not work. Case in point, when Java 1.5 came out, enum became a reserved word, so 1.4 apps that used enum as a variable name wouldn't compile without the -source 1.4 parameter to javac. I don't doubt that PHP will end up doing something similar, and since nobody will really want to keep more than 2 versions of php running on a server, you'll have to go in and fix your code.

 

[JFS]

my opinion in this is that php 5 never really took off. when compared to other languages that when we see an update everyone goes rushing updating it, the same didn't happen to php. Many people still use php 4, and in many servers you don't have php 5, there are also the cases when you buy scripts, you buy a php script and it may not work on a php5 server without modifications.

 

[Eric]

Sorry PHP4 lovers, but PHP has officially announced PHP4's end of life.

PHP 4 end of life announcement
[12-Jul-2007]

Today it is exactly three years ago since PHP 5 has been released. In those three years it has seen many improvements over PHP 4. PHP 5 is fast, stable & production-ready and as PHP 6 is on the way, PHP 4 will be discontinued.

The PHP development team hereby announces that support for PHP 4 will continue until the end of this year only. After 2007-12-31 there will be no more releases of PHP 4.4. We will continue to make critical security fixes available on a case-by-case basis until 2008-08-08. Please use the rest of this year to make your application suitable to run on PHP 5.

For documentation on migration for PHP 4 to PHP 5, we would like to point you to our migration guide. There is additional information available in the PHP 5.0 to PHP 5.1 and PHP 5.1 to PHP 5.2 migration guides as well.

www.php.net

 

[JFS]

then answer me this... why not wait a few weeks/months more and migrate directly to php 6 ? that way at least you migrate to something that is new and not 3 years old and at the midle of it's life ( or near end ).

 

[dhscott]

then answer me this... why not wait a few weeks/months more and migrate directly to php 6 ? that way at least you migrate to something that is new and not 3 years old and at the midle of it's life ( or near end ).

Perhaps this is the reasoning for discontinuing support for PHP 4...

Windows won't officially support Windows 98 / ME anymore now that Vista has come out, but they did when XP was out because they knew that some don't like upgrading when everything is working fine. But there comes a time when they have to force them to do it, otherwise they lose business.

Ofcourse, that example doesn't apply to PHP .

 

[Eric]

PHP6 is still alpha, and although PHP5 is 3 years old, it is far ahead of PHP4 and just short of PHP6. Iit's not in the middle of it's life, it's just now beginning.