P3D Re: NTSF/FAT

[roberts@xxxxxxxxxxxxxxxxx (John W Roberts)]

>From: michaelk@xxxxxxxxxxxxxxxxxx (Michael Kersenbrock)
>Subject: P3D Re: NTSF/FAT

>> Relevance of the analogy: You stated that the two computers are *networked*
>> together, which introduces some considerable number of transmission/translation
>> protocols between the two machines. When you access the NT computer's files
>> over the network from the Win95 computer, it's really the NT computer that's
>> accessing the files and passing them on.

>As opposed to "translations" being done by file system code instead
>of networking code?  One is "proper" and the other "doesn't count"?

I read Dan's original post as being a caution to those who might consider
NTFS formatting as a way to get around cluster sizes (in order to readily
handle large 3D image files in an efficient manner) that while this approach
might help to address that particular issue, it could also introduce some
complications, one of those being access to those files if you happened to
want them while running Win95 or NT.

I read Ron's multiple posts as a strong assertion that Dan's caution was
wrong and that we shouldn't consider it worth thinking about, and my further
interpretation was that Ron thought that the fact that the files could be
accessed over a network proved that they could also be accessed directly -
certainly many readers of Ron's posts would think that was the case.
I appreciated Dan's caution (though there's a way around it, as you and Ron
noted), and didn't see any point in Dan being put down as wrong.

>Yes, an NTFS disk physically mounted on a DOS or Windows95 Machine won't
>work other than under NT.  But that only means what was said.  It doesn't mean 
>that an NTFS disk isn't readable under Window95.  Only under the circumstances
>mentioned, is that true. Under some other circumstances it is indeed
>accessable (and logically "mounted" as to seem to be a local disk
>to the user). 

Yes, and only on those special, specific circumstances when it's not Easter
do we not dress up in rabbit costumes. Maybe you tycoons think that buying
a second computer and networking it up for the sole purpose of accessing
an NTFS disk under Win95 is such a trivial undertaking that it's not even
worth mentioning as a possible complication, but I for one think it's
useful knowing that this is an issue, and I still don't see there's such an
effort to give Dan a hard time about telling us.

You're welcome to quibble about the semantics of Dan's statement all you want.
If I say "from North Dakota, go south to get to South Dakota", is that
invalidated by the fact that you can also get there by heading north, provided
you go far enough? :-)

>The original assertion was that an NTFS disk is NOT accessable by
>Windows95 (as a filesystem disk).  The examples showed an instance 
>where that is not true.  This is sufficient to prove show the original
>assertion false as stated.

>If the original assertion is constrained to the single-CPU single-machine
>scenario, then it *is* true, but it wasn't originally put that way
>as I understand it.

Here's an excerpt from Dan's original post:

>>In short, Yes, NTFS does handle files and cluster size, etc better than FAT.

>>   ***BUT*** 

>>NTFS formatted drives are only accesible from NT. You can not access data
>>on a drive formatted this way with Win95, OS/2 or DOS.

So whether this statement is "true" or "false" depends on whether "accessible"
means NT is running on the machine you're working on or just somewhere in the
"community", and whether Dan should have added "unless you buy a second
computer".

>All of this depends somewhat on point of view.  If one has a
>CPU-centric point of view and looks "up" at the system
>hierarchy, then what was originally said is "true", but
>if one looks at the system as a whole and sees CPU-boxes
>as pieces of a system that's been built and/or being used
>then what I've said is how one would see it (I think).

Fair enough. This matters a lot more than who's "right" and who's "wrong".

Readers of computer newsgroups, mailing lists, etc. may have noticed that
the protracted "I-said-this-no-you-said-that" semantic arguments are far
more common than on P3D. I believe this is partly because computer systems
are so flexible - there's almost always some workaround to get what you want,
no matter how horrible it is to implement, and a person making a comment on
how to do something is almost certain to fail to mention at least one possible
alternate way of doing it. I believe a greater-than-usual effort to make
it a constructive discussion is called for. [Note]

I believe it would also be in our interest to try to emphasize practical
implementations for stereo digital work, rather than wandering off into
far-ranging abstractions about what's possible in setting up a computer.

John R
[Note]: After I get off this parting shot, of course. :-)


------------------------------