T.R | Title | User | Personal Name | Date | Lines |
---|
1001.1 | | DECWET::VOBA | | Tue Feb 04 1997 19:50 | 9 |
| Re .0, it is a requirement of NT (across all platforms - x86 and RISCs)
and it is called the system partition. It is absolutely necessary to
create this partition on the Alpha NT system for OSLOADER and
associated files for loading. It's OK to have other stuff in there
too, if you wish.
If anything, this is more MS baggage, i guess, rather than ours.
--svb
|
1001.2 | Can someone elaborate, please? | MAIL2::DERISE | | Wed Feb 05 1997 07:55 | 20 |
| re .1
Yes, I know it is called the system partition, and that is precisely
how I referred to it in my base note.
Why is it absolutely necessary for Alpha NT to have a 6MB FAT system
partition, and not necessary for Intel NT? Another way of asking the
question is why, on the only RISC platform that will continue to
support NT, do we still have to separate the System partition from the
Boot partition, requiring the system partition to be a 6MB FAT
partition?
Seems silly to have to continue to do this, and it only confuses
customers.
If there is a completely valid technical reason why NT installs like
this on RISC, or has to be installed like this on RISC, platforms, what
is it? Of course, I mean other than "Microsoft requires it..." - why
does Microsoft require it?
|
1001.3 | | DECWET::VOBA | | Wed Feb 05 1997 10:40 | 11 |
| Re .2, i did say something incorrect in .1 - that is on Intel NT, the
OSLOADER does not have to be on a FAT partition. The RISC platforms'
firmware and its interaction with NT are governed by the ARC
specification. The FAT system partition is a requirement.
Even though Alpha is the only remaining RISC platform with a future,
there will still be other RISC systems around for some time. All are
bound by the ARC specifications. Furthermore, the cost of changing it
seems to outweight the benefits or removal of annoyance.
--svb
|
1001.4 | Some customers question it. | MAIL2::DERISE | | Wed Feb 05 1997 11:34 | 17 |
| Well, considering that Alpha is the only RISC processor left to support
NT, doesn't that render the old ARC specification mute?
How much could it possibly cost to change this? What would be
required, seems a simple change to AlphaBIOS should do the trip.
If we are trying to get wide(r) acceptance of NT on Alpha, seems like a
small price to pay. Especially if we really want NT on Alpha to "look
and feel" exactly like NT on Intel, except for performance of course.
I'm raising the issue only because I've had several customers query me
on this. Bottom line is, customers perceive a difference. I know and
you know we're talking about a piddly little 6MB FAT partition that no
one even has to know is there, and can be protected via Disk
Administrator from users that don't know any better. But it is still
there. Some customers don't care; for others, it sticks out like a sore
thumb.
|
1001.5 | | DECWET::VOBA | | Wed Feb 05 1997 12:28 | 21 |
| Re .4, follows is my personal opinion and it does not necessarily
reflect DECwest thinking. Here goes...
There are many things in Windows (9x & NT) and PC that "stick out like
a sore thumb". We all can pick our favorite things on our favorite OS
and chew on them. I'd not be surprised if one is annoyed by the FAT
OSLOAD system partition on RISC platforms. That's what keeps the
religious wars going on on the news groups 8^).
DECwest has recognized some of these and have done something to
alleviate their annoyance level. For example, AlphaBIOS disk express
setup does create the 6 MB FAT OSLOAD system partition in a painless
way.
While PPC and MIPS are no longer with us on the future NT adventures,
it does not mean things can be chucked out on a whim. NT 4.0 is still
supported on all 3 RISC platforms (not counting 3.51). If you put this
question to us at the end of 1999 (after NT 5.0/SP10 ships ;^) the
answer may be altogether different.
--svb
|
1001.6 | Does "cost" include support/dissatisfaction costs? | GEMEVN::GLOSSOP | Only the paranoid survive | Wed Feb 05 1997 12:48 | 23 |
| > Furthermore, the cost of changing it
> seems to outweight the benefits or removal of annoyance.
Really? Every continuing difference from Intel NT (which is further
different from Windows 95) is something that can *and does* cause
people frustration (speaking from experience), which can further lead
to word-of-mouth criticism about "Alpha quirkiness", etc. I STRONGLY
encourage that this be fixed.
(Among other things, this can mean that your boot drive is "C:", while
your "real" system disk is "D:", and some very old installs blindly
default to C:, etc. If you forget to create it, or if a system you
inherit doesn't have it, then you don't have the option to convert
your real system area to NTFS to apply security, etc. If I remember
correctly, one V4 install around here had the small one as C, and
the "wastebasket" on C filled up, causing a later install to fail
until the cause had been determined.)
Basically, it can be a real pain, with the pain not visibile until
well after various earlier decisions were made, that doesn't exist
for Intel platforms. (One thing to remember is that while Digital
people that wind up doing more than the typical number of installs
have "adapted" to this quirkiness, customers aren't in that position.)
|
1001.7 | | 2954::FOLEY | http://axel.zko.dec.com | Wed Feb 05 1997 13:30 | 7 |
| RE: .6
You can create the partitions such that the big NTFS
partition Is C: and the piddly 6-10MB FAT partition
is D:. I do it all the time.
mike
|
1001.8 | | DECWET::VOBA | | Wed Feb 05 1997 14:51 | 18 |
| Re .6, my opinion - the "cost" does include support/dissatisfaction
costs. I'll maintain that it outweights the benefits or removal of
annoyance (to those who are annoyed).
What you said about the C:, D:, to Z: drives is only one of the many
Windows (9x & NT) and PC quirkinesses (let's chew on that, shall we?)
that has nothing to do with OSLOAD FAT partition. I can show you with
my HiNote Ultra II NT Server system (which used to run Windows 95).
Have you sorted out the DOS/Win95/WinNT multi-boot on Intel PC (yes,
that's another quirkiness and let's chew on that, too, shall we?)? If
you have, we can move on and chew on the UNIX/VMS/NT multi-boot on
Alpha... And on, and on.
Please, don't take offense with my rhetorical questions. My point,
let's fry bigger fish.
--svb
|
1001.9 | Let's lead, not follow! | MAIL2::DERISE | | Wed Feb 05 1997 14:59 | 21 |
| I certainly do not want to argue with anyone in this conference, least
of all our counterparts from DECWest.
I am simply passing information along which I have received first-hand
from customers that are new to Alpha, let alone NT-on-Alpha. Every
platform has its quirks. But keep in mind most potential new NT Alpha
customer is probably already an NT Intel user. Some don't care about
this 'quirk'; others, especially our ditractors, use it to criticize
us.
My personal opinion: WE are in a position to set the standard for NT on
RISC platforms. Let us not fail to take leadership positions because of
some consortium (ACE/ARC) that never amounted to anything or because of a
specification that is now irrelevant. Let's stop apologizing for
having superior technology - instead let's flex some muscle!!!
Besides, having to create and maintain this 6MB FAT partition at this
point in time seems quite silly indeed - it is almost laughable if you
think about it! :-)
Have a great evening!
|
1001.10 | One last thought. | MAIL2::DERISE | | Thu Feb 06 1997 06:11 | 13 |
| re .9
ditractors - detractors (sorry for the spelling)
One last comment. It is great that Digital is working with Microsoft
on 64-bit NT. It is also great that 64-bit NT will be available on
Alpha first. But I can't help wondering the look on a potential
customer's face when I tell him/her, "... oh, yes we still have to
create this piddly little 6MB FAT partition..."
:-)
Ya takes the good with the bad!
|
1001.11 | Maybe ARC is still valid? | MPOS01::naiad.mpo.dec.com::mpos01::cerling | I'[email protected] | Thu Feb 06 1997 06:20 | 10 |
| >> Well, considering that Alpha is the only RISC processor left to support
>> NT, doesn't that render the old ARC specification mute [sic]?
Article in the paper yesterday stated that Apple is working on
developing a plan for its systems (PPC) to run NT. That means
that the ARC will not be unique to Alpha, again. Maybe not
seeing other vendors' marketing plans is what makes the point
'mute'?
tgc
|
1001.12 | Apple? Please!!! | MAIL2::DERISE | | Thu Feb 06 1997 07:40 | 10 |
| re .11
That's ridiculous! The PPC has had just about 0 impact on the NT
market. Why should we cater to vendors that do not, for once, have as
strong as a position as we do? What vendors ever did that for
Digital??? Besides, how many organizations in the Fortune 500 really
care about what Apple does? Apple is seen as a company near death.
If we want to do something with Apple, let's get them to drop PPC and
adopt Alpha. Now that's a fight worth fighting.
|
1001.13 | | TURRIS::lspace.zko.dec.com::winalski | PLIT Happens... | Thu Feb 13 1997 20:01 | 16 |
| Or get Apple to implement their bootstrap procedure in a way that
will allow booting directly from a NTFS partition.
This is a wart on our implementation of NT that makes us look
inferior to Intel. This is the last thing we need when trying to
increase our market share. The fact that Intel doesn't have to boot
this way shows that there are no technical barriers to implementing
direct boot from NTFS on Alpha. The only reason for it seems to be
compliance with a now-dead consortium. That is pure silliness.
We should at least discuss with Microsoft the idea of eliminating the
ARC system partition on Alpha. I would think that from their
standpoint, too, anything that brings the Alpha implementation of NT
closer in behavior to the Intel implementation is goodness.
--PSW
|
1001.14 | | DECWET::SCHREIBER | DECeNT | Fri Feb 14 1997 09:36 | 3 |
| This issue is being raised again with Microsoft. Stay tuned.
Benn
|