CompuServe Thread

Don't worry to much about speeding up the optimizing pass of the Linker. When you are developing, your aren't too worried about the size of the executable (especially if your are including debugging info.) Since most programmers will recompile everything once it is debugged, mainly to remove debug hunks, it is easy to say OPT after everything is working. On the single pass compiler. I agree with you about the advantages of multi-pass compiling. I LIKE the flexibility of my declarations. FORWARD's are a pain, and sometimes dont help (like two-way interdependancies). Here is what I suggest. Have a compiler option something like CHECK where the compiler simply does the syntax/error checking it normally does without the code generation. This would be a simple way to take care of much of the problem of speed, since for 90% of the time we are invoking the compiler for just this purpose (at least in a strict language like M2). As for this *new* Amiga product, what kind of upgrade paths and costs will there be to existing owners. (I don't object to paying a little for significant improvments in functionality/power, upgrades -vs- updates.) Also, a good debugger is essential. The only people who don't want a good debugger are the ones who have never used one. In the VAX environment, where all languages share the same object code formats and debugging specs, ALL languages are fully supported for source level step by step debugging by the SAME debugger. If only this were true in the Amiga world… Give me a good debugger and a good listing file, I can clean up and get some really bad code working in a few hours. Give me no listing file and no debugger, (post mortem doesn't count.) and I won't even try. Take care, Mark Gooderum