Thank you Mike
  for a well expressed, informative post.  I almost didn't reply because of
my webmail probs but though this is probably going to take me quite a while
with that clunky interface I will try and attempt editting so that my
quotes are trimmed.  Really this is also inline with part of this thread
which now has branched to LVM, since my greatest complaint is it's clunky
interface.  That is perhaps the greatest achievment on PowerQuest's part
that I think in comparison and after the fact where IBM (or really these
specific coders that wrote it) totally missed the boat.  In fact I think it
even dropped below IBM's high-set bar considering that very early on OS/2
had superb context sensitive help and vluminous documentation, much of
which was actually readable and due to charts and graphs such as for
command reference switch guidelines and syntax, entirely understandable at
a glance.  OTOH the menues and what little help there is within LVM is
murky at best.

  At worst it can result in very destructive and non-sequitor behavior. 
Allow me to cite one example.  The first time I used LVM was on a system
that already had Win98 on it due to past experience with flavors of OS/2
(and btw other than poor cdrom support, I agree that W3&4 were
relatively easy to install, at least for someone familiar with OS/2
conventions. I did try to gain some converts who found it frustratingly
difficult and especially on items they felt were trite, where some little
incidental by no means important to a running system would halt the
install) where having programs installed ready for Win-OS2 migrate was a
big bonus.  Because I understood the need to create a compatibility volume
in order for LVM to recognize/use it, I did so for the windows partition. 
I decided to change it and the only way LVM would allow me to do that AFAIK
was to delete it and create it again in the manner I wanted.  I saw this as
an "undo" step where LVM should have basically given up it's 

recognition, deleted it's needed entries from the menu.  Instead it deleted
the entire partition with not a word of warning or explanation that it
would destroy something it did not create. This became a csacading problem
since it ended up wiping all partitions from the first disk and at that
time I did not have the foresite to write down *exact* size and location of
partitions to get them back.  Months of work was gone, irretrievably in an
instant and for no good reason.  It was a shock to me since one of the
things I hated about MS was it's preadory ways and loved about OS/2 that it
"played nice" with neighbors. 

   The closest Fdisk ever came to an austere interface was unexplained
greying of menuitems such as "set installable".  Just why can't I
do that whenever I want? But that is nothing compared to the rigid  control
exerted by LVM.  That would be fine if that rigid control was there to
force the user onto the correct, safe path but it is totally inconsistent. 
It keeps you from doing things you want to do and should be able to do and
doesn't even offer a note if you attempt something truly destructive. 
Either give the user complete control and good documentation, context
menues and help, so he makes informed choices or lock him into to a
foolproof path.  Don't mix.

  I disagree with you as to why windows installs are relatively painless. 
First of all I find them at least somewhat painful due to the number of
reboots required but also and more important is the ridiculous
"warnings" and predatory behaviours regarding other systems
totally non-essential to getting a running windows system and designed
simply to destroy others through the typical FUD.  However what they do
have is the foresight to not balk and dump the install for a problem that
would never harm the system and most of which can be fixed once the system
is up much more easily than during install.  So far the most painless
install while remaining intelligent and non-predatory that I've seen is
Xandros Linux.  The installer is given a choice between "Auto"
and "Expert" install which is superb.  If one creates a linux
swap partition and a sufficiently large linux partition ahead of time such
as with PQMagic, the "Auto" install is intelligent enough to go
there instead of taking over the

/any disk.  Choosing "Expert" basically asks "what, where
and how" offerring choice on partition usage/assignment, graphically
and highly friendly, admin/user names and passwords, and network info and
that is it!  It installs, and boots one time and you are at the desktop
with just about everything working including networked printers, usb mice,
cable modems, samba-windows shares, everything.  Furthermore the install CD
allows the installer to hit "Shift" at startup should one have
machine specific issues where a menu of 10 or more variations of install
modes (APM, ACPI version, Interactive, Restore, etc) are available.  It's
beautiful in it's simple logic and effectiveness.  LVM may be effective but
it is not elegant or simple.  OS/2 installs may not be as bad as a sharp
stick in the eye but IMHO it is absurd to exert show-stopping concern over
trivia that can be better fixed once the whole system is up, running and
available rather than being stuck in a vastly reduced power state dep

endant on slow, arduous boot process.  Sure once your system is up you can
then create a bootable CD, but not only was that unavailable back when it
was released but it is the eminent Catch22, only good if it is easy to
install the environment that can create such an improved bootable
environment.

  This is all so far regarding just getting installed but there is after
the fact to consider as well.  For the desktop user compatibility problems
are not over after LVM has finished it's install chores.  Afterward it
locks one into an inflexible environment regarding other systems.  An
example would be handling Fat32 file systems.  As important as Daniela
Engert's contributions have been to OS/2 users I think it behooves us all
to notice that her excellent DASD and ATAPI replacements eschew LVM usage. 
She is apparently not all that fond of LVM either and I suspect for very
good reasons, some of which I am now butting heads with after a 3 year
hiatus from having to fight with install problems.  They are by no means
insurmountable but there is just no good reason why they exist at all since
most of those problems really serve no purpose especially for the desktop
user.  A good example here is the oft repeated disabling of CPU caching
during install, a show stopper on many systems, but

 once the system is in no problems whatsoever have ever occurred to me
after enabling said caching.  That is just absurd that the install is *way*
pickier than the system is.  IMHO IBM's writing this off as a sign of
"high standards" is just posturing and spin doctoring.  M$ isn't
the only company capable of calling a bug a feature. 

  BTW IIRC even v2.0 of OS/2 could not be installed on an MFM drive,
requiring at least an RLL but preferring IDE or SCSI.  I did attempt to
install the 20+ diskette v2.0 on a 386 w/ 4 megs of ram and a 20Meg MFM
drive and determined on my own that the hard drive just couldn't keep up. 
Later I read that at least RLL was required which I had already done.

  Don't get me wrong I positively love OS/2 and it is a testament to it's
quality that there are quite a number of Warp 3 and even v2's still working
out there and on fairly modern equipment and can you even *imagine* trying
to install Win 3.11 on a modern box? Wow! what a nightmare that would be. 
Even Win95 would be a joke.  However I finally parted ways with M$ with 95
when they told me they would send me the "USBSUPP" file,
supporting USB, for $50 *or* of course I could spend $80 for an upgrade to
98 which was no more than a fixpak by OS/2 standards.  But, now IBM is in a
similar position even after declaring it dead and even after it reveled in
the user support that caused Warp to win an internet poll for favorite
opsys, of feeling the need to charge me for basic USB support.  I am
nearing the end of my rope, having tired of both insult and injury and this
little experiment of mine of installing on new hardware may yet end up in
my total conversion to Open Source.  My standalone b

ox is still useful but it's days are numbered and that is part of why I
attempted this install.

  Hopefully I'll discover, since I am very pleased that all it took was
Danis506 to get to the install, that FDISK will come to my rescue because I
am all but totally unwilling to unleash LVM on a 20+ partition system with
years of accumulated data.  Unless there is a *vastly* improved version now
available, I simply don't trust it and I wish to have good Fat32 support
and that seems unlikely in an LVM environment.

Sincerely
Jimmy


> rallee2{at}comcast.net wrote:
> 
> > Hello again Steve
> >  
> >
> 
> 

> Last *difficult* [almost impossible is nearer - without the IBM 
> technician hanging off the phoneline whilst I was talked through a whole 
> range of alternative scenarios! ] "OS/2"install I had was
with OS/2 V2.0 
> on an MFM hardrive on a 386DX33 from floppies - basically the old W3-W4 
> installation was a PITA because there were _so__many_ possible selection 
> choices - in my eyes it was long-winded rather than difficult - always 
> helped to have read the entire [originally large] printed User Guide and 
> reference manuals - before commencing the installation! M$ made their 
> installation so painless by hijacking the entire disk by default, there 
> never was and probably now never will be a [completely standalone] 
> logical drive installation for Windows or whatever the latest name for 
> Longhorn will be [VISTA?].
> 
>
> 

> 
>mander as Boot Manager one has to selectively "hide" actual 
> partitions, including selected logical partitions in the [Primary] 
> Extended partition and unlike in LVM where unless the Boot-volume is 

> 
> The thing about LVM is that it is *scriptable* - changing the BM menu on 
> the fly before a SETBOOT etc.
> In PQPM in order to have rhe scriting capability it was necessary to 
> purchase the Enterprise edition, and in addition to have some software 
> interaction problems, it cost a real bomb - like *quite* a few W4 full 
> client licences.
> 
> It would be informative for you to run "LVM /STARTLOG: file>" and look at just how much information it actually finds and 
> checks about all of the partitions etc,
> 
> I don't have any problems combining PQPM and LVM use - just don't use 
> PQPM with any JFS partitions and just "touch" the drives on
next bootup 
> by running LVM.EXE, say change [edit and un-edit] a volume|partition 
> label and then exit "saving changes" which invariably here
updates its 
> internal markers on the volumes.
> 
> LVM.EXE is *so* good - relative to the ultra-sluggish M$ disk Management 
> Tools in that one can run through the entire scenario of menu options 
> without actually implementing any of them - you can step-back as far as 
> you want, even exit completely and not  save the changes - as one does 
> that under M$ it's irrevocably done in nearly all cases!
> 
> Incidentally in addition to PQPM 6.0x, the final with support for HPFS, 
> I have PQPM for OS/2 - final 3.05.302 edition which I still use on 
> OS/2-eCS, caveat being that it won't run the OS/2 GUI version if greater 
> than 256MB of RAM is fitted - irrespective of how much of that RAM is 
> currently in use. Apparently the fullscreen DOS graphical version [which 
> is what W95/98 had to use] from 3.05 *will* run with >256MB RAM - I 
> suspect that's because it just doesn't recognize anything over 64MB.
> 
> -- 
> Regards,
> Mike


------------------------ Yahoo! Groups Sponsor --------------------~--> 
Get fast access to your favorite Yahoo! Groups. Make Yahoo! your home page
http://us.click.yahoo.com/dpRU5A/wUILAA/yQLSAA/9rHolB/TM
--------------------------------------------------------------------~-> 

 
Yahoo! Groups Links

 To visit your group on the web, go to:
    http://groups.yahoo.com/group/os2hardware/

 To unsubscribe from this group, send an email to:
    os2hardware-unsubscribe{at}yahoogroups.com

 Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 



---