??? 1/11/2002 5:48:42 ??, ?/? "John Sadler" <[EMAIL PROTECTED]> ??????:
> >Perhaps the answer is to build a QL with PC hardware. > >It is possible to create an assembler which will take 680?0 code and create >Intel machine code output. > >I think Marcel and allied programmers have the knowledge to create a virtual >machine which can use the Windows drivers. > >Then you have the advantages of PC hardware with the advantage QL code >running directly. > >Then we have solved the QL Hardware problem for ever. > > Not really.... and that's the big debate... There are indeed several Assembler translators out there that can make SMSQ/E run natively on practically everything. This will generate two additional problems: 1. Someone will still have to write drivers for these devices (SMSQ/E is not exactly compatible with the device driver writing practices of other systems) 2. The old software won't work (not a problem if you ask me) without emulating the old hardware once more... so what's really the point? (the answer IMHO below) The other approach to create the same effect is to have SMSQ/E completely re-written in C, while modifying it to be more compatible with developer methods for other platforms... The drawback in such an approach will be the creation of a Unix-ish system, which will lack in effect ALL the benefits of SMSQ/E (extremely tight code, superb multitasking etc). I however have no problems with a UNIX like SMSQ/E if concessions by ALL users are made... (not to turn it into Vapourware - I do believe Unix is vapourware to a big extent btw- etc....) IMVHO the main problems of SMSQ/E is the total lack of a GUI and the disgrace of a filesystem we have.... I also disagree with the S*BASIC being so tightly integrated with the system (IT CAN be removed though and once I have my systems back in order I will be attempting to use the shell (or something similar) and running S*Basic separately). If the above were to be corrected to a manner acceptable by all, I really don't see why we should move to a different platform base as what we have now would be totally acceptable for all computing work (and compatibility wouldn't pose so many problems. To illustrate what I mean I tried recently Nasta's "Pizza challenge" (from IQLR ca. 1996) using a Windows XP w/ 1 GB ram and FreeBSD w/ KDE desktop and 256 Meg RAM. Still QPC and QXL outperformed both other systems (approx 100 open QD'98s with all its sub thingies approx. 200 processes in all) as opposed to the 45 Notepads open in XP before it crashed and the 50 in FreeBSD before the system became so slow it wasn't usable... while SMSQ/E still was usable...) That still says something ;-) Phoebus
