Ciao, Eduardo,
We strive for MS-DOS compatibility, and MS-DOS, PC-DOS, PTS-DOS, OS/2, Windows 9x and Windows NT all use the same format that is so well described in the RBIL tables 2619-2622. So I chose that.
IMHO, the COUNTRY.SYS format does not affect compatibility, as the information is accessed through DOS interrupts.
But not everybody installs FreeDOS from the installer and uses all its external commands / files. I've always strived to make the kernel a drop-in replacement of the MS-DOS kernel (apart from Windows support). So, it now works with HIMEM.SYS, SMARTDRV.EXE *and* COUNTRY.SYS of MS-DOS! ;-) It doesn't yet work with the menu stuff of MS-DOS, but who knows, perhaps this will be changed...
Actually, one of the reasons I chose that format is that I needed to test my COUNTRY= code with a COUNTRY.SYS file that is *really* correct, and what more correct than the Windows 98 COUNTRY.SYS? Only *after* ensuring that the code works right, I composed our COUNTRY.ASM and fixed its bugs to make it compatible with the *already* correct code. Could I do it otherwise? If we had an already existing COUNTRY.SYS - yes, but debugging *both* code *and* data is far beyond my abilities... ;-)
Besides, the RBIL describes the COUNTRY.SYS format of MS-DOS/PC-DOS/PTS-DOS/OS2/Win9x/WinNT rather well, and we always try to follow the RBIL in our kernel work (which is even written in our docs).
DR-DOS, for instance, uses a different one.
DR-DOS has never been 100% MS-DOS compatible, which is one of our goals, albeit we'll never reach it, of course, but... the only true goal is the unreachable one ;-) I think that FreeDOS is already much more MS-DOS compatible than DR-DOS. (This is not to say that DR-DOS is bad, no! ;-)
I won't hide that I prefer Steffen's format, as I've already implemented NLSFUNC using that but, if we finally go for the MS format, we should consider to expand it to include a Yes/No table (for int 21.65.23), which is a nice addition in Steffen's version. It is pretty straightforward, but it will invalidate the use of COUNTRY.SYS files from those OSes in FreeDOS (apart from legal and/or ethical issues).
As far as I can understand (and I don't understand it very much), Steffen's format is similar to that of DR-DOS, which isn't well described in the RBIL. I think that the "Yes/No" character can be used only in *fully* translated, complete versions of FreeDOS. Besides, there are other responses like "Abort, Retry, Ignore, Fail?" that are not covered there. Therefore, such characters are much better translated along with the strings. So, until we have *fully* translated version of FreeDOS (like German or Italian MS-DOS, for example), those configurable "Yes/No" characters are unusable.
As for ethical and/or legal obstacles, I don't see any. The Russians of PhysTechSoft use it in their PTS-DOS and nobody has sued them yet. Do you think that if Microsoft had even the slightest chance to win, they wouldn't sue them? Remember how Adobe Systems ("the little Microsoft" ;-) sued Dmitry Sklyarov of ElComSoft, Russia. At the time of the announce of PTS-DOS, MS-DOS wasn't yet killed by Microsoft, so it was a direct competitor. FreeDOS can't be a competitor of something that Microsoft killed by their own hands. FreeDOS *cannot* be compared with Linux in this context, as Linux *is* a direct competitor of the Windows NT family, which Microsoft won't kill soon ;-)
A presto, Lucio
------------------------------------------------------- SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift. http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 _______________________________________________ Freedos-kernel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/freedos-kernel
