woodelf schrieb: Considering what it does, in comparison to modern operating systems, it looks quite understandable and at a first glance simple; but you shouldn't underestimate that several very skilled people were working to fit the whole thing to the limited ressources of that machines. One does not want to produce a look alike - you can have things like DCL interpreters for modern systems as well (however, DCL is not relevant for RT-11), but you need much efforts to make the beast then *compatible* with the original. I have long experimented with cloning OS/2 after IBM abandoned it, but a lookalike is worthless, and the OS/2 kernel is full of historical kludges which one needs to reinvent to make it do at least a small leap. I have meanwhile written a number of compilers, so a macro assembler and FORTRAN itself isn't too hard, compared to semantic monsters like C++, in particular if you take such a nice and symmetric CPU like the PDP-11 into account, for which code generation is real fun. What is hard, though, is to fit the whole stuff into 8, 12, 16kW as it is common for those old machines. The idea to take public DECUS C and write it in C (maybe using lex and yacc) will unfortunately likely lead to large binaries. Also TKB and its kernel infrastructure needs some thinking when it comes to overlays and coroutines; 'ld' from Unix and the Unix process loading/scheduling/execution complex is by magnitudes simpler. -- Dr.-Ing. Holger Veit IT-Management Fraunhofer IAIS Institut f?r Intelligente Analyse- und Informationssysteme Schloss Birlinghoven D-53757 Sankt Augustin e-mail: holger.veit at iais.fraunhofer.de Tel. +49 2241 14 2448 Fax. +49 2241 14 2342