There are a few reasons. The main reason is that we wanted Zend to be an independent scripting language engine, that could be used in other products as well. After implementing most of Zend, we can say that we've been successful at that - the Zend engine isn't tied to PHP, and it'll be possible to use it in other programs that need to embed a scripting language engine. Another reason is that we expected Zend to have related commercial aspects (even though, we can't stress it enough, Zend itself will be freeware and open-source), which would not be related to other developers of PHP. Giving the engine a name of its own makes the distinction between PHP and its underlying scripting technology very clear.

We expect it to be the only engine, for several reasons. First, as the authors of the PHP 3.0 scripting engine, we feel that it would be very difficult for others to write a scripting engine that is completely compatible with the PHP behavior, in each and every case. This is exceptionally more difficult when you consider the fact that PHP 3.0 is such a 'strongly-untyped' language, which behaves in several different ways according to the types of the variables and expressions used.

Also, and just as important, we believe that the Zend engine is a very fast implementation of the PHP language. It's about as quick as it can get, without adding platform-specific compilation. To better understand this, you can compare it to Java. Java binaries consist of platform independent bytecode, that requires a platform-specific implementation of a virtual machine, in order to run properly and efficiently. In contrast, Zend is, by itself, platform independent, and would run on just about any machine that has an ANSI C compiler. Zend can't be made much quicker without implementing something similar to Java's platform-specific virtual machines, which isn't our intention at all. Also, compiling to Java-like bytecode requires much more compilation overhead, which doesn't suit a web scripting engine such as Zend.

We're not planning to base any scripting languages on Zend, other than PHP. However, Zend can be used as a scripting language engine in any program. The 'scoop' is that we're looking into integrating Zend as a stored-procedure language within MySQL, which is another very popular and impressive piece of open-source software. In fact, since PHP and MySQL are a very popular combination, we're confident people would love MySQL's stored-procedure language to have the same syntax as their web scripting language.

We feel that PHP 3.0 is already superior to ASP (in its VBScript form, at least) and Cold Fusion, in the amount of power and functionality that it extends to its users. Zend, and the future PHP version that will be based on it, will supplement this by boosting performance drastically, and by adding even more functionality to end users. An example of the new functionality that Zend introduces, which was lacking in PHP 3.0, was native COM support for Win32 users. The new Zend object interfaces allows COM objects to be manipulated and referred to as regular PHP objects, similarly to ASP. CORBA enthusiasts should not be discouraged by this statement at all - the Zend object interface allows for easy adding of CORBA support, or any other object functionality.

The timeline for Zend hasn't changed. We expect to finish the implementation of the engine within two months. After that, we would hold a limited beta test cycle that will include the major PHP developers and Zend sponsors. Assuming there won't be major surprises during that initial beta test, a beta version of PHP 4.0 would be released to the public, and stabilized towards a final release in a few months. We expect Zend to be finalized in the second quarter of '99.

As for PHP 3.1 and 4.0, there hasn't been any official announcement from the PHP development team; However, at this point it's accurate to say that PHP 3.1 is cancelled, and that the next major release of PHP would be version 4.0. Most of the major ideas behind PHP 3.1 will be incorporated in PHP 4.0.

The commercial product that we're planning is an optimizing compiler for PHP scripts. That compiler will boost the performance of Zend-based scripts even more, and in addition, allow to distribute compiled versions of scripts, without revealing their source code. As Zend itself, without the optimizing compiler, is already extremely fast, we believe that regular users, and small-medium web sites will not need these performance gains. Only large, high profile sites will find that optimizing compiler useful. In addition, companies or individuals that sell PHP scripts will benefit from the fact that they can now distribute their programs without revealing its source code. The bottom line is that it won't be difficult for those that need the optimizing script to file compiler, to pay for it. All other users, which are the vast majority, will be able to use the new powerful and high performance implementation for free, something that is extremely important to us. We may also have other commercial products in the future.

Marketing Zend-related commercial products will be separate from the PHP efforts.

They'll require the Zend engine, they won't be standalone binary executables. A machine that would want to run a precompiled PHP 4.0 script, for example, would have to have PHP 4.0 installed in order to do that.

The main reason we've asked for sponsorship, is because we invest most of our time in Zend development, which doesn't leave much room for anything else. For that reason, we're interested in monetary sponsorship.

Beyond our sincere gratitude, and an honorable mention on our sponsors page, we 'reward' sponsors in several ways. Sponsors of Zend will be included in the small-scale closed beta test that we plan, which means they'd be among the first to receive the new technology. Major sponsors will also be given a copy of the Zend-related commercial products as they become available. Companies interested in making exceptionally large donations will also receive technical support from us (in a reasonable amount - as after all, we're only two people). In such cases, of course, the details would be worked out with them.