Slashdot Mirror
PHP5: Could PHP Soon Be Owned by Sun?
At first glance, the obvious changes to PHP are a result of the success of the Java platform and the weaknesses of PHP revealed in comparison. With the release of PHP 5, it's apparent that the developers of PHP and the Zend Engine (essentially a single group) felt compelled to make PHP much more like Java with respect to object-oriented programming. Zend, the Israeli company backing the development of PHP, promises on their web site that "full integration with Java will be closer than ever before." Hmmm, full integration with Java, huh?
On November 4th, 2003, Zend announced a strategic partnership with Sun. This deal also included advisors from Borland, Macromedia, MySQL and others. The purported purpose of this deal was to make PHP part of Sun's web server and bring it to the corporate world of development that previously had been dominated by ASP and ColdFusion. Now with the release of PHP 5, it's far more apparent which path PHP is taking.
PHP's object model was re-written from the ground up and mimics the abstract properties of Java's object method. There are private and protected members and methods, abstract classes and interfaces, in practice, identical to Java's object model. The extent of the influence that Sun has on PHP today is clear. If you have experience with Java and PHP, reading the details of the object model reveals the absolute cloning of the Java object model within PHP 5. From throwing exceptions to static variables, PHP 5's object model mimics Java in concept all the way to the syntactical level.
This is great for enterprise developers using Sun products, but with the release of PHP 5, what does this mean for the half-million PHP developers worldwide who have depended on PHP for open-source development, or for the developers whose ideas and efforts have brought PHP up through the ranks from its inception in 1995? When PHP goodies were bundled with Sun's web server on November 4th, 2003, with a $775 price tag, PHP began down the path of corporate ownership. For years developers have eagerly contributed their ideas and efforts to be a part of the success of PHP. Now that all the hard work and volunteering has paid off and PHP is a worldwide success, it appears that PHP could soon be another corporate shill owned by Sun, MySQL, Borland and Macromedia, if not on paper, then by direct influence on the people at Zend. Of course it will remain open source so that those half-million developers can continue to contribute their time and genius to its success, but if those thousands of contributions lead to direct financial gain for companies whose coffers are already overflowing and are simultaneously using those contributions to manufacture software with price tags out of reach to anyone but corporations, is PHP still the language developers should be focusing on for use in the open-source community?
On the positive side, this edition of PHP does bring improved performance and a new suite of MySQL functions. Backward incompatibility is limited to a list of ten issues. Additionally, there are only minor configuration file changes that need to be made to the web server. Several directives have been introduced for configuring php.ini files, mainly dealing with hashes for encryption.
Some very useful functions have been added to PHP5. It's been nine years in the making, but PHP5 now includes two functions to uuencode and uudecode. Combining those functions with the new socket and stream functions, developers can create a lots of "kewl" applications. An application to automatically encode and decode files to and from news servers comes to mind as an example of how to incorporate these new functions. At that point of course, a developer could use any of PHP's existing functions to continue to manipulate the files, store the contents in databases, and so on.
An addition to error reporting aids developers in keeping their code up-to-date. The E_STRICT message tells developers when their code is using deprecated functions or is in danger of not being forward compatible. However, don't assume that E_STRICT will be output if using E_ALL, because it won't. E_STRICT must be explicitly declared to output its suggestions to PHP 5 code.
While the rewriting of PHP's object model to essentially that of the Java object model does raise flags about the direction of PHP, it is still a powerful addition to the PHP5 release. Java became successful for a reason: it's intelligently designed and facilitates code reuse. By borrowing the best features of Java's object model, PHP has leveraged itself with far more credibility as a programming language that can stand on its own two feet (even if Sun, Borland and Macromedia are holding it by its arms).
Some vital re-workings in the PHP object model lie in how objects are treated in low-level fashion. Instead of passing the actual object itself, PHP's object model passes by reference. Now when operating on objects, developers can pass around multiple handles to the actual object allowing for more powerful and efficient applications. Existing PHP objects do not need to be re-written to take advantage of this change in PHP 5.
In general, developers who have experience with Java will easily adapt to PHP 5's object model. On the downside, if PHP is a developer's primary language and he or she hasn't been introduced to the world of static variables, public and private methods and the host of aspects included with this new model, they may have a bit of a learning curve adopting the higher-level format of object-oriented programming in this release. Overall, though, this change will be a plus for creating large-scale, object-oriented applications with PHP.
Keeping pace with the developments in MySQL and PHP's tight relationship, PHP5 has added a new suite of MySQL functions relating to the new features added since MySQL 4.1. Denoted as Improved MySQL Extension, its purpose is to allow developers to take advantage of prepared statements and the other additions to MySQL 4.1 and above.
Something very interesting to note with the addition on the Improved MySQL Extension is the absence of bundled MySQL client libraries with PHP5. There are numerous reasons given for this, including the different licenses that PHP and MySQL are under (PHP is under a BSD/Apache type license and MySQL is under a GPL license). The PHP5 documentation also assures developers that "there will always be MySQL support in PHP of one kind or another," but doesn't go into details as to the future of MySQL support. This perhaps is further evidence that the long-lasting popularity of LAMP environments (Linux, Apache, MySQL, PHP) will soon be replaced by SLOP environments (Sun, Linux, Oracle, PHP). If Zend continues to shy away from MySQL and completely joins forces with Sun, MySQL may soon no longer be part of the picture, and cheap, fast development may no longer be possible for PHP developers in the same capacity as it is today.
Zend clearly has underplayed the extent of the shift that has taken place concerning the future of PHP. While this version of PHP does provide a much better object model and added features, is this the beginning of the end of PHP as the choice of web scripting language for the open-source community and developers not under the employ of corporations? Will the average developer still be using PHP five years from now, or will the usefulness of PHP be limited to companies who can afford to shell out thousands of dollars for all of the necessary software that may be required to make PHP a viable option for development (along with the purchase of products from Sun, Macromedia, Oracle, Borland and others)?
While today this is still speculation, the evidence and tone lends credence to the thought that with the success of PHP, built on the backs of developers worldwide, the near future of it may include an alienation of it from those who energized it at its genesis, propelling it to the corporate enterprise status that those in control of PHP are seeking today. No matter what actually happens, developers should be aware of the major developments with PHP beyond the surface level function additions and new object model. Companies and developers who are employing PHP 5 for large-scale applications today at a reasonably low price may be in for a surprise in the next few years, if operating PHP at full capacity involves the purchase of additional, expensive software.
All of these comments are irrelevant. These predictions of PHP's future may come true or may not. But if you don't like the direction PHP takes, fork the project, take the source code, remove the parts you don't like, grow it in a direction you do like.
See that's why Open Source is different than proprietary software. It's not just another choice, it's fundamentally DIFFERENT. Nobody can take the software and force it down a direction you don't like because you and like-minded individuals can take it in the direction you like.
- For the complete works of Shakespeare: cat
What drives me crazy about PHP is it just keeps growing. Need another function? Ah let's just add it! The problem is the language grows from something quite simple to something that's impossible to master.
I like the approach Python has taken. Everything is kept clean and simple, and the complexity is added through importing modules. Need another function? Import it! I guess that's why Python is said to "fit in your head".
I'd better stop before I start a flame-war. The point I wanted to make is PHP and Java will both probably collapse under their own weight, and another simpler language will take their place. If the plan is the grow PHP into Java, then there will be tonnes of books needed to reference everything, which is good if you want to sell books, but bad if you want to write programs without having to constantly look something up.
It seems to me that a programming language needs to plan for growth before it starts, otherwise it grows and gobbles up the mental resources of the programmers using it. Once it's too big, people will just fall back on simpler tools.
Ruby on Rails Screencast
Keep using PHP4 or PHP5 if you dont like where Zend/Sun takes PHP6. Its not like features will start to magically disappear from the old versions.
The way Java handles protected variables (due to packages) is starkly different from the way C++ handles protected variables. Fortunately, it looks like PHP picked the (less broken, IMHO) C++ way to do it.
As for your final comments--all too many PHP developers don't know "CS 101 stuff", serious or no. Also, I know that when I first learned about the OO methodology, it was quite confusing. Now that I know more about it, I'm convinced that there's a lot there to be avoided, and all of it should be carefully considered.
Fortunately, (like the crippled "object system" in PHP 4) if you don't want to use it, you still don't have to use it.
pb Reply or e-mail; don't vaguely moderate.
This article is packed with FUD of the worst kind. Bleeeech!
A Multiplayer Strategy Game for Mac OS X, Windows, and Linux
now don't get me wrong, i'm not bashing php. i use php all the time and it is a pretty straightforward tool and quite easy to pick up. the inevitable problem with trying to reform a language is that you need to "break" it in order to fix it
In the movie "City Slickers" Jack Palance's character quips that the secret to life is just one thing, and once you know what that one thing is, everything else makes sense. I'm beginning to think that programming languages are the same way. The "one thing" about Visual Basic was introducing components. Perl's one best thing is powerful reporting capabilities. Python's contribution is namespace (just type 'import this' into the interpreter for an easter egg's explanation).
PHP's "one great thing" seems to be initial ease of use. It's dead simple to install, the php website's documentation is second-to-none, and it's relatively painless to cut-and-paste code inside HTML to make stuff work. My problem, however, is the same complaint I have with the Windows operating system: PHP is impossible to master, because it's becoming too broad with too many functions and too many special cases.
According to http://tnx.nl/php there are 3079 core functions in PHP4 (as of november 2003), compared to 206 in perl.
3079? That's just seems insane to me.
Ruby on Rails Screencast
Java became successful for a reason: it's intelligently designed and facilitates code reuse. By borrowing the best features of Java's object model, PHP has leveraged itself with...
For years I have been asking for concrete examples of OO producing code reuse in the business domain, but have yet to see a convincing example. OO does NOT have objectively demonstratable magic properties (except in a few narrow conditions that I don't encounter very often).
Some people pick OO because they personally like it, NOT because it is objectively better. Enough with the OO hype.
Table-ized A.I.
True. Forking PHP and building a successful community around the fork would send two powerful messages to corporations. Message 1: You can profit from OSS but you can't (successfully) dictate its politics. Message 2: Normal people are in control of OSS, not corporations. If you ask me, Message #2 will upset them the most.