(E-Mail Removed)lid wrote:
> (E-Mail Removed) wrote in
> news:(E-Mail Removed):
> > I'm not sure about this, but ISTR the filename is stored
> > twice in FAT16/32, once as 8.3 and once as long using
> > crosslinking. But... I'm far from sure. If true, then it
> > may be posible to change one but not the other, though
> > certainly windows couldnt do it. As jb says, it serves no
> > point, leading only to total confusion.
> >
> > The OP's request reminds me of muckysoft's general approach
> > to problems, forget fixing them, just bend something else
> > so it appears to work right. In the case of W95's long
> > filename handling, that was an excellant and successful
> > trick. But never any time since has it been at all clever.
> Do you say that M$ did that well (me not sure)?
> FWIW, M$ forgot to teach DOS how to delete LFNs completely.
The FAT FS structure is mostly beyond my knowledge limits, so I wouldnt
know. But I do think the ability to use LFNs really was a big move
forward for end users, the 8.3 system was/is a horror, and caused all
sorts of problems.
When I briefly came across a 3.1 system I didnt know what most of the
files did, and in many cases the only way to find out was click em and
see what happened. I'm betting this is precisely what many folks did,
curiosity got the better of them one day, and this is a recipe for
screwing up the OS.
GeOS addressed this problem with an 3 fields, 8.3 SFN, description, and
I forget what the 3rd field was, but this was a great help before I
switched to LFNs. Now things have moved forward again with freeware
apps that give you LFN plus an extra description field.
Managing to not only get LFN going but also do it compatibly with the
then existing FS was a perverse and smart move, and is one of the
things that made W95 such a big leap forward from 3.1. One of MS's good
moves, among the thousands of bad.
NT