The new font code will never fail to find a font. ``The best way to deal with errors is to define them out of existence.''

Definitive 8.0 release quite soon. John went around the developers asking about the 8.0 status of major extensions. Tclx: ``on your tail''. incrTcl: end of August. BLT: sort of working now. OraTcl/SybTcl: works but doesn't exploit objects. Tix: ditto. Tcl-DP: late September. Expect: works with no exploitation. ObjTcl: autumn.

SpecTcl 1.1 is imminent. After that, effort will scale back. It hasn't been as popular as expected.

Poll around the room shows that very few developers work in a Windows environment, but many ship for it. Supporting pre-95 Windows is a pain and it will not continue long; poll says few people need it, and most of them expect to stop needing it within 1-2 years. Mac support is continuing, for now.

Eric Schmidt's departure from Sun torpedoed SunScript as a separate business unit. Bad news: no sales, no support. Good news: group still exists, substantial size, more Java stuff, all software will be free.

Neosoft will sell and support Tcl/Tk if people insist on paying for it (some do).

Doing neat things involving both Tcl/Tk and Java is big now. Unfortunately, replacing Java's cruddy graphics library with Tk did not get suggested soon enough, and is not going to happen. In the works: Java access within Tcl/Tk, Tcl/Tk as container for JavaBeans and OLE/ActiveX components, Jacl as a way of getting Tcl into pure-Java environments.

There will be an 8.1 release, hoped by January, as a consolidation release. Finish object conversion of Tcl, objectize Tk, fix bugs. Internationalization (including UTF8 as internal character set), send for Windows and Mac, modular parser to make it easier to use the parser as a tool.

Poll of the room on importance of new features particularly favored more widgets, megawidgets, and thread safety.