Haai,

On Wed, June 6, 2007 18:02, Jason Thorpe wrote:
>
> On Jun 6, 2007, at 10:37 AM, De Zeurkous wrote:
>
>> 1) It includes features which are not necessary (GUIDs for example);
>
> Maybe not necessary in your limited world, but they're quite useful in
> the real world.

See below.

>
>>[snip]
>
> You're still confused as to what that MBR is actually there for.  I
> can't fix that; you just need to educate yourself.

Perhaps you aren't quite able to grasp my point. I *know* the MBR is there
to keep egoistical software from meddeling with the disk.

>
>>[snip]
>
> Can you please be specific?

AFAIK, Windoze and possible other OSes, reserve a facotry-fixed amount of
entries, which defies the whole purpose of GPT: flexibility.

>
>> 4) Worse, it deals with some stuff that is likely to change (fixed
>> checksum format, little-endianness, other constants which should be
>> copy-on-write)
>
> Why do you think these things are "likely to change"?

CRC32 error checking is already trivial; why should we insist on using it
for all eternity? As for endianness: see below. Besides, as far as I can
work out (documentation seems to be rather sparse), the block addressing
is limited, as well. Again, it should be copy-on-write.

>
>> 5) Doesn't seem to have non-specific metadata fields, even optional.
>
> Oh really?  Please explain specifically what you'd like to see, and
> then maybe I can show you how to do it with GPT.  I happen to have a
> fair amount of experience using GPT in interesting ways.

Nothing specific -- but I'm sure there is some point from which they're
going to useful. Having a free-form, metadata-based structure is quite a
necessity when designing something that is supposed to be unifying.

>
>>[snip]
>
> Uh... so that automatically makes it evil?  A de facto standard that
> provides basically everything that we need is certainly better than
> none at all, and adopting a de facto standard has advantages to
> rolling one's own.

Yes, perhaps this is what we need now. But what about the future? We
wouldn't want to ``standardize'' yet another system that isn't going to
last long due to coming from the minds in the consumer market, would we?

>
>>[snip]
>
> Any reason other than religion?

There is only one, definitive reason for it: The network is the
computer(TM) :)

>
>> Still, why not remove the impact of the collision entirely?
>
> As engineers, we weigh effort vs. benefit.  For the amount of effort,
> I don't see a lot of additional benefit over GUIDs.

So, because it's less effort right now, we should stick to GUIDs for life?

>
>>[snip]
>
> Quite a mind reader you are.

Thanks :) I'm using a technique called 'Experience'. You really should try
it.

>
>>[snip]
>
> What Windows-specific stuff are you talking about?

See above.

> Legacy?  You mean
> the protective MBR?  Whatever.

Yup -- it exists solely for legacy stuff to be able to boot from it and
preventing ill-behaved stuff from touching it.

Baai,

De Zeurkous
-----------

Friggin' Machines!

>
> -- thorpej
>