From ouster@tcl.eng.sun.com Fri Aug 8 17:39:45 CDT 1997 From: ouster@tcl.eng.sun.com (John Ousterhout) Newsgroups: comp.lang.tcl Subject: The SunScript story Date: 7 Aug 1997 16:32:49 GMT Organization: Sun Microsystems, Inc. Lines: 80
We've received lots of messages lately from people who weren't at the Tcl Workshop, asking for more detail on the recent restructuring of SunScript. Here's what happened.
The story goes back to late Summer 1996. By that time we were just about "ready for prime time" in that we had several pieces of software that could become products within a few months. So, we began looking around for opportunities to productize our work. Over the Fall, we talked to several groups within Sun's main product organizations (SunSoft and JavaSoft) but weren't able to find a good fit.
Around the end of the fall Eric Schmidt, Sun's Chief Technology Officer, suggested that we spin the Tcl team out of Sun Labs into an independent small business unit under his supervision to commercialize Tcl/Tk products. Although this is something relatively unusual for Sun, Eric had been starting several of these as experiments to develop new businesses for Sun. We spent the next several months putting together a business plan, and after presentations to Eric, Scott McNealy, and others, we got an official "go ahead" in late February. At that point we started recruiting and preparing for a launch; the official announcement was made at the end of April.
Unfortunately, around this same time Eric decided to leave Sun to become CEO at Novell. He made his announcement in late March and started at Novell in early April. When he left, no other senior management person stepped in to manage the fledgling business units. After about six weeks it became clear that SunScript couldn't succeed in its commercial aspirations without high-level Sun management support. So, around the beginning of June we restructured SunScript back into a Sun Labs project again.
SunScript is still alive and well, but our mission is different from what we'd initially anticipated. The bad news is that all our commercial projects are on hold: for the immediate future we have no plans to offer products or services for sale. The good news is that the development team is still well supported (our headcount is now higher than ever) and we're starting on some interesting new projects.
Since we don't have near-term commercial aspirations we've decided to make *all* of our software freely available in source form. This includes things we'd previously hoped to sell, such as the plugin, SpecTcl, and other Web tools such as WebTk and an embeddable Web server. As we make new releases of these programs the copyright notices will change to follow the standard Tcl/Tk notices, which allow arbitrary reuse without license or fee. We're also releasing the Tcl logos freely, plus the Java integration stuff described below.
The restructuring also freed up resources for new projects. We're using these resources for two substantial projects related to Java. The first is a package that will integrate Java with the existing C Tcl interpreter: you'll be able to load a Java virtual machine into any Tcl application and easily invoke Java classes from Tcl. There will also be support for JavaBeans. The second project is an extension of the Jacl work started by Ioi Lam at Cornell. He's now working for us, and we're starting a major project to build a new Tcl interpreter implemented entirely in Java. When it is done, it will allow you to execute Tcl/Tk scripts in Pure Java environments such as NCs and off-the-shelf Web browsers. Jacl will provide the same interfaces between Tcl and Java as those that we're building for the C Tcl interpreter, so applications developed for one will also work with the other.
We think that the Java/Tcl integration will provide an incredibly powerful way for people to migrate to Java. If you want to migrate your application base to Java, the best way to do it is with Tcl! You can start off using the C Tcl interpreter, which will work great with all your existing C and C++ code. Then you can use our Java library to incorporate the Java virtual machine into your applications; you can use a few Java classes to start, and use Tcl as the glue to make the Java classes work together with your existing C/C++ code. Over time, you'll write more and more code in Java, and you can gradually swap out the old C/C++ code, again using Tcl as the glue so everything still works together. Eventually, you'll be left with only Tcl and Java code. At this point, you'll be able to migrate to 100% Pure Java environments by using your Java code with Jacl.
To sum it all up, SunScript isn't doing what we'd initially hoped for, and we're disappointed not to be able to offer any commercial products right now. However, SunScript will be doing some neat things to integrate Tcl with Java, and all of our stuff will now be freely available, which should make it more accessible and easier to build on.