Conference 7.286::digital

    
    This is very interesting!
    
    	When I studied Finance, they taught that the only real reason for
    conglomeration was SYNERGY - something in your supply or delivery chain
    that helped you be more competitive, like a steel company buying up a
    coal mine where they got most of their coal.  This was totally
    forgotten during the merger-mania of the 80s with companies like
    Coca Cola buying movie houses and oil companies buying Montgomery Ward. 
    
    	Now it looks like they expect the reverse to happen - 
    a company should divest itself of non-synergistic divisions when that
    becomes apparent.  A lot of this is related to portfolio theory -
    management of risk.  How can you estimate the risk of a company that is
    in several diverse businesses?  If they were un-consolidated you could
    buy any combination of shares in each business.  Once consolidated you
    can only buy the proportion of each that is represented by the parent
    stock.  A good example would be UTC - if you decided that Carrier would
    thrive because the reconstruction of Florida would require lots of air
    conditioned construction you could only buy UTC of which Carrier
    represents some %. 
    
    	So how do we know which parts of the company are 'non-synergistic'? 
    I think that's what BP is trying to figure out now with the Supply
    Chain analysis.
    
    JN

    
    Re: .0 & .1
    
    Business history in the US has been characterized by waves of
    conglomeration (whether by hostile takeover or simply for
    growth) followed by waves of diversification for at least
    40 years.  This is nothing new.  
    
    Glenn

    
    
    
      I agree with .2, wait awhile and the "experts opinions" will change.
    Like most things, it's cyclical.
    
    
                                                Phil...

re Note 2084.0 by EBBCLU::BROUILLETTE:

>   "The idea is: Put all your eggs in one basket, and keep an eye on that 
> basket," 

        I thought we had done that!

        Bob

    re: .0 et al.
    
    I agree that the such advice seems cyclical. Unfortunately, in trying
    to do the right thing by giving our customers migration paths between
    architectures, we mistook that with doing everything - completing
    products with little customer interest, trying to invent things from
    scratch when better/cheaper solutions were already done elsewhere, etc.
    
    Separate from pure organizational theory, what appears to pay is focus
    and followthrough.  Our divergent businesses run counter to operating
    in that manner to date, and perhaps only by divisionalizing this
    company will we obtain the needed focus.  This is precisely the manner
    IBM, SUN and HP are running their businesses - and are spinning more and
    more off to succeed or fail on their own merits.  IMHO, we are trying
    to do so many totally 180 degree opposite strategies in the same company
    (example signing up service providing VAR's to sell networking and 
    server products while at the same time the Services are building 
    a large DESKTOP service business), we will be lucky to have the
    strategies net out positive.
    
    From a financial point of view, you can create stock for major
    divisions - like GM did with EDS, so that investors can participate in
    part of the firm's success.  You can also spin such firms out totally
    and simply retain a majority of the voting stock.
    
    Final point: Let's not confuse the need for management focus with the
    need for creating shareholder value.  I offer CML in Acton, Mass. as a
    an example of a firm with a basic mission to participate in the leisure
    time industry - having a tiny HQ staff, and enabling entrepreneurs to
    operate the various divisions: in this case Carol Reed clothing, Nordic
    Track products, Nature COmpany, Sybervision videos, etc.  CML adds value to these
    divisions only by providing access to capital and has the good sense to
    let these quality companies move ahead based on their customers needs,
    within a core operating model of freedom for results.
    
    While CML had trouble in the mid-1980's, it's focus on mission and
    creating stockholder value and its values on entrepreneurship, allowed it
    to drop unsuccessful lines and add value in new divisons.                                                                 
    
    BT

		     Why do I feel like I'm in a stew pot?

                   doesn't mean they're not out to get you.

    re Note 2084.4 by LGP30::FLEISCHER 

>   "The idea is: Put all your eggs in one basket, and keep an eye on that 
> basket," 

>>>        I thought we had done that!

  Yes, we had done that for the 80's.

  Now, for 90's, Microsoft is riding on it with WINDOW-NT for all the platform.
  I always question what DEC gain by bedding with MS beside Alpha ?
  Don't we innovate SW anymore ?

Michael..

    
    	it seems that we've not done well at doing what everyone else is
    doing.... anytime we've had major successes, it has come about through
    true innovation... if we don't offer something that is different then
    we become a much different company. We essentially will become a trader
    in commodities markets where success and profits basically come from
    who can sell things faster, better , and cheaper. We won't need 90,000
    employees or even 50,000 for that matter. Just a little room with 150
    or so people answering telephones and maybe some administrators gluing
    together supply and distribution networks. Innovation is what has set
    us apart from the rest. The trouble that we're in today, is a
    combination of giving up on innovation to a large extent (riding on
    Vaxen momentum) and focusing more on processes and procedures. We can
    have the best processes and procedures in the world but without
    innovative products to go along with them, then  we're not very special
    to anyone. I believe it was Martin Minow who suggested that we would
    become much like Sears and Roebuck. We won't design/ build/ innovate
    but simply sell things built by others with our nameplace slapped on.
    Yes I know Alpha is innovative but it's simply not enough and may in
    fact be the last truely innovative thing that we do. I sure hope I'm 
    wrong though.
    

    	Haven't you heard? We're getting out of the software business.
    Minicomputers and chips are where the big bucks are in the 90's.

That would require that we first get *into* the software business.

re Note 2084.4 by LGP30::FLEISCHER

  -< one company, one strategy, one message >-

	One gun, one foot.

re .12  -< one company, one strategy, one message >-

How about     "One chicken, one egg, one basket."

    re .12
    
    what happened to the _other_ foot?
    
    
    		ouch!

    The other foot wasn't compatible, as it had 36 toes...mwr

	This reminds me of the old Toto Coalo (sp?) song

	I eat cannibals
	It's incredible
	You bring out the animal in me
	I eat cannibals.

	Just because something is a cannibal it doesn't mean its not vunerable
	from others as well as themselves.
	They could even show others the way to destroy themselves.

	Heather

    re: .15
    
    And they were numbered backwards too!
    
    Marie (who still remembers systems that matched our orange binders)
    Was it POPJ,P or POPJ P,

    re: .17  POPJ P,    ...the AC address field goes b4 the comma.
    
    	Dick
    

    Now you know why I have cat named JFCL.  She's 14, and since she is in
    good health, she will probably outlive the architecture that gave her
    her name.
    
    /Charlotte

    Re: .-1

    Probably not.  A new -10 on a chip will arise from Seattle, courtesy
    of Len Bosack.

    re: -1 Oh, and will it have a SNOBOL accelerator, Tom.?
    
    -2: JFCL pronounced "Jeh-fickle" I presume? Great name
    for an off-the-wall cat!
    
    	Dick
    

    C'mon, when is one of you microcode gurus going to come up with the
    PALcode to emulate a KL10 on an Alpha chip?
    
    Awright, so what if you throw away 28 bits at a time?  Memory's about
    1% of what it cost when the real KL was discontinued.

    It's been done.  Look at a GNU tape in about 6 months...

    I know it's continuing the rat-hole, and I apologise but . . .
    
    JFCL was (at the college I was a student at in England) pronounced 
    as "Jump and do ***-all". (Of course every MACRO-10 program needed 
    to start with a no-op such as JFCL.
    
    I remember the arguments when we upgraded from a KI-10 to KL-10, as to
    which was the "fastest no-op", and whether or not to change all our JFCL
    instructions for something else (TRN maybe??) (or was it the other way
    round?) Ah, youth!

    You obviously didn't read the PDP-10 System Reference Manual very well.
    It explicitly listed the fastest no-op for each machine.  It also had a
    section on how to make programs run fast.  I was responsible for
    getting that information in the Manual, knowing that hackers (like me)
    were just dying for the information.
    
    The microcode on the KL10 made most of the simple no-ops equally fast.
    We put some minor effort into that just to stop the foolishness
    described by .-1.

I still have my old Massachusetts license plate 'JFCL'. I had it on an old
Digital sales car (1974 White Chevy Malibu Classic) that I sold when it had 
over 100K miles on it.

Jim
(who got his start hacking TOPS10 4S72 on KA10s)

>    It explicitly listed the fastest no-op for each machine.  It also had a
    
    I remember that now! (Somewhere I still have a copy). 
    
    Actually the 'arguments' were more of a philosophical kind; there were 
    two schools of thought -- the 'hackers' who said that you had to use 
    the absolute fastest no-op for each machine, and the 'realists' who 
    maintained that for any normal (i.e. non-time-critical) program, it 
    didn't matter a JFCL which no-op was used; any one was as good
    as any other.
    
    (Incidentally, I am reliably informed me that the expression ****-all is
    not in common use in American English. It's a British expression
    meaning "nothing at all".)
    
    
    Nigel

    Re: .25
    
    I remember the day you swept thru the office confiscating all the
    copies of the PDP-10 System Reference Manual that had instruction
    timings in them, so that we'd stop spending time micro-tuning
    code.  I still have the copy that I managed to hide just before you got
    to my office :-)