Conference 7.286::digital

Anker, i think one could make the argument that there are "de facto standards"
in the PC and workstation area that never quite achieve the status of
architectures but that beguile customers and suppliers that are a bit leery
of the Darwinian free-for-all.  The Intel hardware architecture has assumed
this sort of de facto standard status.

One might well ask who the enemies of these worldviews are (from the standpoint
of those holding each worldview).  Architecture is vulnerable to the rapidity
of change (so Architects tend to fear and put down developments that they
didn't anticipate when they did their inital design, e.g., "PCs are toys").
Darwinians are vulnerable to the consumer's desire for investment protection
(so Darwinians tend to minimize the customer's need for legacy application
support or compatibility from one product release to another).  The "de facto
standards" worldview folks are vulnerable to revolutionary change, in terms
of totally new applications or order of magnitude decreases in price (so folks
at Intel might want to denigrate the Nintendo threat).

I don't see that much of Digital oriented toward the Darwinian model...we
just don't seem to be looking at revolutionary new applications, things that
haven't been done before.  But that's probably the long-term growth area.  We
do seem to have promoters of the Architectures and de facto standards floating
around in different positions.
								- paul

    
    I have a slightly different take on this.  I think the issue really has
    to do with customers' tolerance for complexity.  Architecture is a
    powerful tool for mastering complexity; architects cater to complexity,
    while darwinians avoid complexity.  What's happened over the past few
    years is that as computer system price/performance has improved
    exponentially, and applications have become available that are useful
    to "ordinary people", the average tolerance for complexity of the
    computer marketplace has declined dramatically.  The darwinians'
    predisposition to avoid complexity is better aligned with the bulk of
    the marketplace than the architects' predisposition to master
    complexity.
    
    It is Digital Engineering's comfort with complexity that has gotten us
    into trouble in the past few years.  What was once the majority of the
    market remains roughly the same size in absolute terms, but is now a
    niche market.  You can't continue to grow a $14B company by catering to
    that niche.  Worse, this comfort with complexity leads to an entire
    constellation of behaviours ill-suited to the current market.
    
    len (whose personal name was not set specifically for this thread).
     

re Note 3073.0 by ASABET::ANKER:

        Anker, that's a classic!

        Another way of looking at the conflict between what you have
        described as the Darwinian and the Architectural is to
        compare planned economies with free-market economies.

        A free market economy relies upon the fact individual
        decisions taking into account only local considerations
        nevertheless can outperform centrally and totally planned
        systems.  The irony is that Darwinian and free-market models
        actually produce systems of greater complexity yet demand
        less mastery of complexity on the part of the "decision
        makers".

        I have often thought that Digital's decline started in the
        early 80's when the architectural (central engineering)
        triumphed over the Darwinian (the product lines).  Or perhaps
        it was an attempt to replace the charismatic influence (as
        opposed to absolute control) of Gordon Bell with greater
        control and more central structure.

        Bob

    
    Re: .-1
    
    I'm glad I'm not the only one who felt that way.
    
    Sometimes I think the reason that 'centralism' has
    prospered (shall we say thrived at the expense of excellence?)
    so well (to our company's demise) is that the darwinistic/free market
    approach to doing business requires top-notch people who aren't afraid
    to compete - and they are precious and hard to find.
    
    Many people who cannot or will not engage in competition seem to 
    prefer to dominate/manouver politically in order to survive, rather
    than sell their services on merit in a corporate 'marketplace'.
    
    
    	-Publius

  ....Darwinian Architecture. It looks Darwinian but the 
  marketplace just chooses one architecture for whatever
  reason; defacto-standard (TCP/IP), freely available (UNIX)
  or the availability of applications for a cost that has
  a good relationship with the price of hardware and
  software (200$ application for a 2000$ PC with MSDOS/Windows).
  The complexity is not any less; it's just that the software
  maker handles all the complexity while making it easy
  for the user to use a system. Just look at any word-processing
  or a spreadsheet or a database system on an IBM PC.
  They have to be able to talk to  and be talked to from a whole
  variety of other systems. 

  Cutting through all the BS, Architecture was really invented by the
  big companies so that they can control who makes compatible
  products (in most cases, themselves). This used to work before
  with IBM, but with PCs becoming more and more powerful
  and less and less expensive, you need not wait for self-appointed
  Gods to anoint your software. If have ideas, you need not wait
  for the next five year plan to roll through. You just follow
  the Nike Process: JUST DO IT. And there are 120 million PCs
  waiting for it.  

  Nari

    Re: .-1
    
    The PC business has become such a success precisely because
    it _was_ highly architected - but not in the way the term
    has acquired meaning in Digital. It is highly architectured
    in the sense that components have clear interfaces and clear
    semantics and the synthesis strategies are clearly understood.
    
    You need architecture (strategy), you need Darwinism (tactics).
    If you don't have any backbone you fall over, if all you have is
    a backbone you don't get anywhere.
    
    So how should a company like Digital thrive on chaos? I have
    on my desk Tom Peters (author of In Search Of Excellence and
    A Passion for Excellence) book: Thriving on Chaos which
    presents no less than 45 areas/solutions to turning companies
    the size of Digital around, targetting five key areas of
    management - Responsiveness, Innovation, People Power,
    Leadership and Systems.
    
    I have read a number of books on change management and business
    reengineering (after all that's part of the business I'm in)
    but this book in its sheer detailing of solutions and breadth
    of solutions has to win hands down.
    
    Its 557 pages thick and in it I came across every single
    major problem that Digital has and ways to combat and solve
    them.
    
    Thriving on Chaos
    Handbook for a Managment Revolution
    Tom Peters
    ISBN 0 330 30591 3
    Published by Alfred A. Knopf Inc (1987)
    Published by Pan Books Ltd (1989)
    
    re roelof
                         

    
    
    I'd like to draw your attention here to the fact that the above
    mentioned Darwinian model is based on what I'd call a 
    scientific myth, the biological evolution theory, which
    from a philosophical standpoint is largely based on
    circular arguments. I wouldn't feel quite comfortable
    with an organization built on sand, so to speak. 
    
    I don't want to say that both models could work to some
    extent;but, sooner or later relying on them will probably
    cause trouble.
    
    
    helmut

    I don't see the two approaches as mutually exclusive.
    
    The value of an architecture is to make integration of 
    components together easy.
    
    The value of the Darwinian approach is to be fast-to-market
    and innovative.
    
    However it is produced, customers will want to take the 
    new product and integrate it with existing products or with 
    other new products.  This is most easily done if the 
    products conform to an architecture.
    
    I think the problem that we and IBM have got ourselves into
    is that we have over-done the architectural work.  We've 
    taken the architectures down too far in the levels of 
    detail -- in other words we've over-engineered it.
    
    We need to get back to a broader-base architecture to allow 
    the innovative new products the freedom to develop -- we 
    need IT architectures that allow the same scope as Lego enjoy
    with their architecture.  Lego can and do change more than 
    30% of their product annualy.
    
    So the challenge is to have people say that "that's enough
    architecture, now let's get on with the products".

    Re: .-1
    
    Good points. It's not IMHO just the architectures but - far worse -
    the architectural _processes_ i.e. how is an architecture arrived
    at. I, for one, coming from Philips 2 years ago was simply
    amazed at the number of architects, senior architects, technical
    directors, senior technical directors floating around all seemingly
    with input, few seemingly with direct product or financial accountablity.
    
    At Philips the role of marketing and product management was
    far, far stronger as indeed it had to be to survive in the
    cut throat consumer electronics. Believe it or not the function
    of technical director didn't (and still doesn't) exist.
    
    Don't get me wrong: I'm not against the people in these roles 
    they are often highly intelligent people. Bob Supnik to mention
    one is a man I have a great deal of respect for. But I do
    strongly disagree with the _role_ of technical director and
    the strong technical orientation of discussions and functions
    that these roles promote and conversly by their very presence
    the downgrading of the role _and control!!!_ of market input
    for product development.
    
    Put simply: at Philips marketing _owned_ the product creation
    process at Digital marketing _inputs_ to the product creation process.
    
    re roelof
     

Anker,

   I must disagree with this note.  I also disagree with the analogy
of planned versus free market economic systems.

   Our supposedly free market economy is actually bound by a series
of architectures which need to be constantly monitored.  An example
of such an architecture is our laws on anti-trust.  Originally created
to prevent companies from ganging up to dominate markets, the rules are
now being modified to allow companies to share large research and development
investments.  This is an example of an architecture that was created to
foster Darwinian competition.

   It is clear to me that the success of a large company in the computer
industry depends largely on the proper mix of these two philosophies.  We
should take each segment of our company and examine what model is being
used by others in the industry who are successful and emulate this.

   I will use an example.  We are not currently a software company.  We
are a company which owns, produces and sells a lot of software.  We used
to be a software company.  For a brief period this company produced most
of it's software using Bliss on VAX/VMS.  This may or may not have arisen
via a architectural decision, however it is an example of how such a
decision could provide a successful framework in which a Darwinian
approach could thrive.  The two standards, bliss and VMS, allowed the
company to develop a highly skilled and productive software force which
was applied in a Darwinian fashion at the lower levels to solve customer
problems.

  Currently this company has no such overriding standards.  How can a
software company thrive or even survive with the worst of both worlds?
We have architecture at the medium level with medium level standards
which tends to crush low level Darwinian activity and we have Darwinian
at the high level which leads to low software productivity.

  In short Darwinism can only thrive in a powerful high level architectural
context.

  For example Digital might decide that all software must be written in a
single highlevel object oriented language, say C++, and that we will
focus our high level effort on providing and maintaining C++ class libraries
that span the operating systems that we support and providing C++ compilers
that support this standard.  This would provide the framework that would
support the Darwinian work of creating products and solutions for our
customers.  The architecture would provide what architectures are best
at, namely increased productivity over time and the ability to adjust the
architecture over time to reflect changes in the defacto market place.

  I am not saying that this is the solution, only an example of what a
solution would look like for a specific portion of the company combining
these two philosophies in a symbionic fashion.

  Architecture has it place!  But like a cancer cell, put it in the wrong
place and it can spell disaster.

Harry

RE: 3073.7 by MUNICH::HSTOECKLIN "If anything else fails, read instruction

> I'd like to draw your attention here to the fact that the above mentioned 
> Darwinian model is based on what I'd call a scientific myth, the biological 
> evolution theory, which from a philosophical standpoint is largely based on
> circular arguments. 

Darwinian evolution is also a powerful computer programming technique for
a broad range of problems.  Darwinian evolution of non-living chemicals has 
been used to create new drugs.  Darwinian evolution of bacteria is fairly
easy to show in a lab:  see this month's Scientific American.

Myth?  Hardly.


Phil

    The Darwin/Wallace theory has nothing to do with choice, opportunism,
    potential for success or any set of events where an organism evaluates
    possibilities and chooses a course of action.

    Darwin's central tenet, as set out in "The Origin of Species" and  "The
    Descent of Man" is very simple.  The environment changes randomly
    (globally at an extremely slow rate, although change may be at a
    catastrophic rate locally).  

    Physical characteristics and behavioral traits of the organisms
    inhabiting that environment also change randomly.  Natural selection
    favors those organisms where the changes maximize their potential to
    exploit the environment.  It also favors organisms that maximize their
    potential for randomization.

    The initial attacks on Darwin's theories were because he not only
    challenged creationism but also DENIED the Victorian notion of PURPOSE
    in the process of selection.  If you characterize your "Darwinians" as:

    	evaluating the environment for options
    	
    	choosing a path and planning an outcome

    	able to change course at will

    Then there is not really a comparison with the Darwin/Wallace theory.

    Regards,

    Colin

        Re:                     <<< Note 3073.12 by SMURF::WALTERS >>>

        .-*
        
        What a  great  discussion!    Thanks to the contributors who have
        commented so far.
        
        Since my intent was to create discussion this thread has lived up
        to all my expectations.
        
        I really appreciate  the  thoughts  on  the  intersection between
        Architectural and Darwinian thinking and will integrate them into
        the next revision of my paper.    I  don't remember which note it
        was,  but the one that states that  you  need  to  architect  the
        interfaces  and intersections is very close to my  thinking,  and
        restates  in  better  words  my thoughts about how you  manage  a
        Darwinian environment.
        
        And I'll ask for a favor:  Please don't clutter  this thread with
        a discussion about the merits of Darwinianism.  I chose that term
        simply  because  is  is  catchy  and  leads  to  highly  creative
        inferences.
        
        Anker

I think the discussions re: the validity of Darwin's theories are inappropriate
for this note. The topic invokes strong emotions in some people and often leads
to a highly polarized discussion that is not likely to come to any satisfactory
conclusion.

While Anker's choice of the word Darwinian may in fact be a tenuous analogy,
I think it suffices to indicate an organization that is opportunistic and
adaptable.

re: .0

>Thesis 1: Architects and Darwinians have difficulty doing business with one
>          another

The Architects probably have a very difficult time doing business with the
Darwinians: "What do you mean it's not documented!"

The Darwinians, on the other hand, not only find it easy to do business with
the Architects, they are dependent on them (and the Universities) for the
intellectual raw materials they consume.

>Thesis 2: Digital holds an Architectural World View

I suspect that the upper management would like the freedom of Darwinian
action but are thwarted by the inertia of the people who have to implement
things.

>Thesis 4: It is possible to switch

I don't think it's possible to have both coexisting in the same environment.
As suggested in the second proposal, if DEC wants to capitalize on the
economic advantages of a Darwinian business, it's going to have to completely
isolate that business from all others.

    
    Anker,
    
    Interesting stuff ...
    
    One major point hit me.  It seemed your analysis was based on our
    approach to products which does seem closer to Architectual.  However,
    if you look at our business process we, IMO, have nothing resembling an
    architecture which is why we are so hard to do business with.  The
    independent growth of business and functions created a lot of the chaos
    we live in now.  
    
    Greg
    

    Re .-1
    
    .-1 indeed brings out an important issue.
    
    (1) is getting the right product - an engineering/architectural/product
    management issue. The thrust of this thread (Darwinian Doubters Aside).
    
    However to compound the issue we indeed seem to lack a coherent
    business model (call it architecture) for
    
    (2) getting the customer to product (marketing)
    
    and
    
    (3) getting product to the customer (sales)
    
    My guess is that at the end of the day the problems surrounding
    (2) and (3) dwarf those of (1)
                                

	The differences seem obvious when you compare Intel microprocessors
to Motorola microprocessors.  Intel is 'first with the worst', while
Moto's are much cleaner designs.

	New Intel processors get designed into more products that Moto
processors, because the Intel processors are available a few months sooner.

Moto's processors are well architected, Intel's aren't as well architected.

Architecture takes time.  We used to use architecture to keep our customers
in a captive market. 

Anyone remember the phrase 'Digital has it now!'?  The current market
seems to favor whomever has it before anyone else does, if they make
enough noise about it.  Traditionally, Digital never had it first.  We
usually watched the market and designed products with features that
were better architected than our competitors.

We used to sell systems where everything was garunteed to work together
with anything else we sold.  Customers valued that.  Customers now seem
to accept that not all modules work together in the same PC, and that not
all software packages work together.  Architecture *enforced* by a single
company like Digital made everything work together, but the relative cost
has become very high.  

How do we address the current market?  Do we abandon architecture, or
do we just 'lower our standards'?  Is lowering architectural standards
a bad thing?  From our 'old corporate point of view' the answer is yes.
But the old viewpoint isn't working.  

Less is more.  At least less architecture appears to be more valuable.
But architecture makes products more supportable, and support used to be
important.  Is it still?

    
    Anker: 
    
    Top-notch thinking, eloquently presented. The debate you frame,
    however, is as old as mankind, in my view. Two or three years
    ago, I was consumed for a time by thoughts of whether it would
    be better for my organization to concentrate its scarce resources
    on generating *products* or improving *processes*.  That is,
    should we get to a desirable "end state", i.e., a product out
    the door, with less concern on *how* we got there, or should we
    micro-engineer the processes, and assume that a desirable
    end product would result?
    
    From that thinking and writing, I realized that the debate can
    be captured in other words as well : I call it the "ends v. means" 
    debate (which is really what it is -- and hence my comments that it
    has been ongoing forever.) 
    
    I also agree with a previous reply that the Architects and the
    Darwinists cannot work together well (whether as colleagues
    or in customer/supplier relationships.)
    
    For what it's worth, the conclusion I reached was that I am
    a Darwinist -- that is, I concentrate on getting a thing done,
    regardless of the process.  This has made life at Digital difficult
    at times.
             
    Glenn

    Re .17
    
    As soon as the discussion is about "lowering architectural
    standards" you may have already lost because you may well be on the
    wrong playing field. The goals scored may be fantastic but then they
    simply don't count because the referee isn't looking.
    
    My guess is that we need a different "architectural mindset"
    with a different set of basic assumptions and driving goals.
    Done with care you may well find yourself "raising the
    architectural standards" by supporting a completely different 
    architectural model that is better attuned to the environment 
    (market) around us.                                  
    
    re roelof

    I'm generally suspicious of dicotomies.  They strike me as the neurons'
    lazy way out of thinking about spectra as if only the end points count.
    
    If you want another two-extremes metaphor, you could do worse I think
    than Jane Jacobs' wonderful book "Systems of Survival."  She contrasts
    the ethics of territorial systems (governments, farmers) with the
    ethics of commercial systems (traders, manufacturers.)  She finds them
    fundamentally incompatible in the sense that mixing the ethics
    creates a situation commonly known as "corruption."  For just one
    example:  traders must trade equally with all comers;  organization "in
    restraint of trade" is a corrupt practice.  Governments must protect
    territorial integrity;  selling-out to the outsiders is treason.
    
    Jacobs is careful to point out that even though the two systems are
    incompatible within a single role, they are complementary when
    performed by distinct roles.  Even a single person may change roles; 
    her example is the lawyer who sells services to all comers from a law
    office, and later becomes a district court judge.
    
    
    However, all this dichotomous thinking has, as I said first, the hazard
    of convenience for those who like to keep their thinking unstressful.
    And I believe this hazard is especially abysmal when we stick to
    language as our means of considering what are fundamentally structures. 
    Structures have shape, which only the retina can see as a whole.  When
    you try and argue about structures in language, you get only the
    pieces, one piece at a time to deal with.  Because the aural channel,
    the ear and its nerves and pattern recognition equipment, is
    fundamentally serial.  You don't get random access.  You don't easily
    perceive the whole network of inter-relatedness *as* a network, when
    all you can do is take a serial tour around the network in language.
    
    When an engineer designs any "engine" of technology, she or he will at
    some stage need to draw a diagram.  Otherwise the complexities cannot
    be appreciated as a whole.
    
    In managing Digital, and in re-creating the company, I believe we are
    well past the point of complexity where language alone could suffice.
    We wind up with too many simplistic metaphors, and too many of them
    will be two-valued dichotomies to attain that simplicity.
    
    
    "Architecture" is fundamentally a *visual* metaphor.  "Evolution" is
    fundamentally a *serial process* metaphor.
    
    The two modes are available to most of us within our one head.  Let's
    use our heads.  All of our heads.  Eyes *and* ears.  Shutting the
    company's eyes because we like the sound of "evolution" would be as
    limiting as blocking our ears because we like the looks of
    "architecture."
    
    And let's get back in *touch* with the world.  I mean really:  we have
    to be out there where we can smell the warm armpits, as Ted Levitt
    told us thirty years ago we would someday have to re-learn to do.
    Taste the customers' needs.  Feel what customers feel.  See what they
    do with our own eyes, and see how they do it.  We are far far too
    comfortable just listening to what they formulate into language for us.
    
    
    I think it is time to cut this company into pieces.
    
    Some of the pieces may be mainly evolutionary with minor amounts of
    architecture.  Some of the pieces may be mainly architect-oriented with
    minor amounts of sink-or-swim project-by-project evolution.  Some of
    the pieces may need a fairly equal balance.  But all of the pieces have
    to have a lot of surface area exposed to the outside world.  That's how
    you can sense which way to push the balance (if you *must* use
    dichotomous metaphors....)  Some of the pieces may be territorial.
    Some of the pieces may be commercial.  Probably all of them need organs
    of both kinds, but in very different proportions.
    
    But at Digital's current size or any size approaching ours, I think the
    time is past when keeping the whole company to one attempted
    optimization will work well.  No matter where on the various spectrums
    of possibilities between extremes you try to position the whole.
    
    Customers are just too, too diverse!  See, several earlier replies have
    mentioned aspects of customers they have known, and mentioned a whole
    range of those customers' values.  They are *all* right!  We can't and
    shouldn't expect to decide.  We've got to be close enough to them to
    reasonably well fit the particular needs of particular customers.
    
    And, some amount of architectural strategy is always needed.  If it's
    done by people who are really in touch with the outside world, that's
    feasable even in the '90s.
    
    
    What does not work and never will, is architecture designed in an ivory
    tower, out of touch with real live people in the outside world.
    
    What does not work and never will, is evolutionary competition whose
    key metrics are disconnected from potential customers' actual values.
    
    
    								Russ

Re: .0
    
> Proposal 1: Manage the company consistently with the World View of the customer
>
>        Digital today is predominantly Architectural with a few islands
>        of Darwinians.
    
    I disagree with this statement but I believe that the statement
    describes a model of Digital that the majority of people are using,
    even if they do not articulate it so. (I am talking about the
    organisation and not the products or product architectures.)
    
    Digital believes it has a world-wide architected organisation. In
    reality the Darwins at the bottom of the hierarchy are faced with the
    effects of the time-lag between architectural definition of the
    organisation and its procedural implementation superimposed on frequent
    architectural revolutions. 
    
    An example of corporate architectural revolutions of late are business
    ownership by CBU's, which has been superseded before it really took
    effect. A local example here in Germany was the decision to set up five
    different Digital companies. (This was firm enough for all German
    customers to receive a letter that this would happen.) I haven't seen
    anything in writing that this will not be implemented, but it is common
    knowledge here that it will not.
    
    The current chaos is a product of previous chaos from an era of
    Darwinism (Motto: Do what's right for Digital i.e. make your own
    decision) coupled with the chaotic efforts to reduce the chaos by
    imposing architectural organisation on the chaos. (Cf. "If you automate
    a mess you get an automated mess".)
    
    The Darwins at all levels of the organisation are responding to the
    reigning chaos and their external stimulii each in their own way. Some
    have given up, some have adapted their way of work to fit their
    environment. Others are attempting to make changes which affects
    themselves and others. Most have short term individual planning
    horizons because they have no idea what is going to hit them next and
    when.
    
    Much has been written in this notes file about the absurdity of
    downsizing to increase profit because of the tacit assumption that
    lowering the number of employees will decrease the costs by a higher
    amount than it will decrease the revenue. Now Bob Palmer has been
    talking about selling parts of Digital. Perhaps this is the result of
    an assessment that Digital at its current size and form is not
    manageable and that it needs to be cut up into something that can be
    architected for management?
    
    
    Anke - Thank you for your stimulating article! By the way: who is to
    blame for the current mess? The Management architects (the top-downers)
    or the Darwins (the bottom-uppers)? (I'll just take cover under my desk
    whilst the arrows are flying!)
    
    Mike

    RE: -.15:  I pretty much disagree with your theory, and agree with the
    basenoter, at least more or less.  I do think that our business process
    and strategy is Architectural.  Unfortunately, the Architects used the
    wrong approach.  The reason we are too difficult to deal with is not
    because we don't have processes, but rather because we have too many,
    nearly none of which are really focused toward making our customers
    lives easier.  What we need are Architects who specialize in "Smart
    Buildings", so to speak.  A Smart Building turns off the lights when
    you leave the room, and so on.  Our processes are completely
    disconnected and far, far too complex for *us* to deal with, and yet
    are customers are expected to appreciate them.
    
    A total Darwinian approach is inappropriate, I agree with those who
    have stated so.  It would, in fact, lead to chaos, because of the time
    lags, etc., mentioned by -.21.  However, we need more Darwinians to
    work closely with our Architects, to prevent us from continuing in a
    hidebound manner.  We must be willing *at all levels* of the company to
    take a chance, to make our business and our processes fit the customers
    needs.  Today, all too often, our customers must make themselves fit to
    us.  Wrong, wrong, wrong!
    
    M.

         Very interesting base note and ensuing discussion.  I
         especially like the comments of .20.  I would add just a few
         more thoughts.

         Although the model adds understanding, we need to realize
         that the names given the two styles are heavily loaded.
         Perhaps we should assign neutral names to the styles, e.g.,
         "Style P" and "Style Q".  This is a two-edged sword since the
         names given in the base note are evocative of numerous
         characteristics, aiding our understanding.  Yet not all
         images evoked are necessarily relevant or accurate.

         The characteristics of the "Darwinian" are reminiscent of
         modern populist philosophy. Although there are many positives
         of this philosophy that you and I may agree with, we should
         be careful about making analogies and carrying over all of
         our cultural baggage, which may or may not apply in this
         narrower case.  We need, too, to recognize that we ALL do
         have cultural biases.

         I agree that we should avoid discussions of biological
         evolution and economic systems because these distract from
         the model.  Our many architectures have been generated in the
         context of and with the support of a free-market system. 

         I like description of the basic development of the industry
         in the model.  Some of the replies seemed to imply that this
         makes Darwinism "better" in some sense.  I just think it's
         more pronounced and critical at the moment.  With the
         addition of thousands of small-time players in the customer
         base, there would be a shift in paradigm.  But the funny
         thing about paradigm shifts is that the pioneer in one wave
         is often the stagnant hold-out in the next.  Digital exhibits
         some of this characteristic, as this company once dominated
         the emerging technology markets.  So will today's players
         lose their advantage if they do not continue to adapt.  Just
         because someone works in a small shop with a few, small
         applications, does not mean that individual is automatically
         better at adapting or more evolutionary.  The true test will
         be as that shop grows and the need for structures such as
         architectures appears.

	 Architecture building has a purpose and can be powerful if
	 used appropriately.  It is easy to lose sight of this purpose
	 and to let the architecture become its own end, however.  The
	 way I have seen architectures defined, they should be
	 open-ended, specifying as little as possible to allow maximum
	 flexibility.  Too often I have seen architectures which were
	 really system designs.  Done properly, they can be a powerful
	 aid to any size shop with any style of management.  They
	 simply need to be adjusted to the goal and environment.

         I would argue for a situational approach.  Let's not glorify
         or denigrate either style.  Rather, let's keep attuned to the
	 customer and apply the style which makes the most sense for a
	 given situation.  I believe that a flexible information
	 professional can adapt and exercise multiple styles based on
	 situations.  I hope that we don't fall into the trap of using
	 dichotomous thinking to "weed out the heretics."  A good
	 balance is called for and we may very well have to shift
	 toward Darwinism to achieve that balance.  Just remember that
	 it's not a panacea.


	 Paul

    Re: all previous
    
    I prefer the analogy of "Big Government" rather than Architects and
    "Libertarians" rather than Darwinians because it expresses the element
    of control missing from previous replies. Individuals will generally
    prefer control over their own affairs so long as the cost and
    uncertainty inherent in exercising that control is not perceived as
    excessive. In exactly the same way that the private car wins out over
    public transport most of the time, personal computing will win out over
    I.S. provided solutions. Consumer rather than provider control of the
    resource is what is key. Immediate versus deferred gratification is the
    preference.
    
    "Big Government" tends to chase after some amorphous greater good
    disregarding the values of the individuals who are supposed to
    benefit from such action. Libertarians believe you are responsible for
    your own destiny, which can lead to anarchy when there is no
    recognition of collective responsibility. So sometimes "Big Government"
    is necessary.
    
    As in politics, "Big Government"/Architects are seen as overbearing.
    Thus the "Libertarian"/Darwinian who can assist in shifting control to
    the individual consumer will prosper. 

    
    In the 1989-1991 time frame I had an analogy for the current
    predicament in subsystems that I liked to spring on people, that is
    before middle management had my head on a plate for such heresy.
    
    We were at the time _heavily_ into a central planning model.
    Projects would undergo incredible detailed planning - that was outdated
    a month after it was finished - but we had people who were more
    comfortable 'planning' and checking their planning than they were
    getting on with the business of getting a project done.
    
    My analogy was the trip to California. Deciding on a end means or
    product stratagey was a lot like deciding on a destination for a road
    trip:
    
    1. You know where you want to be.
    2. You know when you want to be there by.
    3. You have a general idea of feasibility as to whether you have
       the resources to get there.
    4. You have an idea of the general route to take to get there.
    5. You get on the road and count on your intrinsic abilities to 
       get you through any unexpected 'detours'.
    
    You don't:
    
    1. Plan on what mile marker you'll be at at 0400 hours on the third 
       day of September.
    2. You don't 'what if' the trip to death.
       'What if we have a flat, can't find the jack, and it's raining ?'
    
    We would spend more time answering that kind of question than we would
    moving forward. Sometimes, things would be on hold for weeks until
    the question either became academic, or an 'answer' was found.
    
    This wasn't meant to be an indictment of planning, per se. 
    Many things are planned in great detail - like building a Boeing
    757. But they aren't planned by people following a 'process' for 
    producing a 'process' with no intelligence to guide the results.
    
    
    How would I apply this to Digital today ?
    
    I'd get a firm grip on which market 'destinations' Digital wants to be
    be in.
    I'd find good drivers who can think on their feet and want to move
    forward rather than plan.
    I'd burn the master copy of the 'Phase Review' process.
    
    Now there might be those who will worry about what 'might happen'
    if nobody is 'watching' and there is no higher level oversight.
    
    My answer is: The open process isn't the problem: It's the quality
    of the people running (driving) that matters. You either trust them,
    or you don't.
    
    California is waiting. Let's get on the road.
    
    -Publius
    
    
     
    
    
    

    > I'ld burn the master copy of the 'Phase Review' process.
    
    Master copy? Master copy? Harbouring _any_ copy of the Phase Review
    process should be grounds for immediate dismissal ;-)
    
    My estimate is that the phase review process addsat least 3 and 
    possibly 6 months of unneeded lag to the engineering
    process. That must add up to perhaps a billion dollars or more
    in lost revenue and perhaps 1 or 2 hundred millions in wasted
    engineering time. This process was most definitely not defined
    for and goaled against time to market considerations...
    
    re roelof
    

The main problem I see with the base note and the discussion is the lack
of time sense.

There is no right or wrong model, except what works AT A GIVEN TIME.

Any company developing a new industry, as we were for many years, must be
vertically integrated as there are NO niche players to fill in the gaps. You
must do it all. Hence any founding father will need to be architectural to
succeed.

But, as the new industry proves profitable the scramble begins and the niche
players enter. At some point in time a critical mass of niche expertise is
reached and the large "founding father" is no longer viable as a vertically
integrated, full service company. At this time the paradigm must shift in some
way.

The timing and direction of this decision is the one which we botched and are
still botching. IMHO this is due to the "darwinian tendency" inherent in our
"architectural mode", mentioned earlier. We have too many people that think they
know best and disregard any directive they don't like. So we get stagnation as
everyone pulls in different directions.


Alan

    The base note certainly helped me to a clearer understanding of some of
    the things that have bugged me for years.  Like why this company
    responds to every problem by changing the organizational structure. 
    Like why we spend more time talking about new software than other
    people take writing it.
    
    However...
    
    One of the things the dicsussion has missed is that much of the
    transition  from architecture to opportunism is a result of equivalent
    changes in post-industrial capitalism.  Service businesses, as opposed
    to manufacturers,  have a much smaller investment in specific lines of
    business.  The relevant values aren't precise planning, careful design
    and engineering;  the key is to get in fast, make some bucks and get
    out before everyone else jumps in and competition kills the profits.
    
    Architectural solutions don't play in this marketplace.  It doesn't
    matter if the customer is large or small.  I've seen this approach in
    major banking and financial operations.  Yes, there has to be a core of
    well-architected systems (like enterprise accounting), but those
    systems largely exist already.  They are not where the smart money is
    investing.  Smart money is going into "quick and dirty" solutions to
    support guerilla business opportunities.
    
    \dave

        I'm amazed at the amount of incredibly good thinking that's going
        into this thread!   The thoughts about the cycle in .-2 (I think)
        as the industry or  market evolves is very insightful, just as an
        example.
        
        The comments about the Phase  Review  Process are interesting.  I
        was personally heavily involved in it.  I was on the Phase Review
        Policy  Committee,  a  member  of  the  Phase   Review  Committee
        (Reviewed the Top 100 products and met on a monthly basis), and I
        was one of the authors of the Phase Review  Guide  Book  - so you
        probably  can't  find  anyone much more guilty than me!   I  have
        saved  the  Phase  Review  Guidebook as a souvenir.  That doesn't
        mean that I don't agree with the sentiments expressed:  The Phase
        Review Process has  been  the  excuse  for  incredible amounts of
        sillyness.  The important point to note is that Architects really
        believe in the work they  are  doing.  We were committed and when
        the  process failed to deliver the  results  we  had  hoped,  the
        problem was the users (we never considered it could be US!).
        
        Anker

From:	OUTPOS::MURPHY       "Dan Murphy, now at LKG; 226-6765" 18-MAY-1994 13:49:07.49
To:	ITRC::ASABET::ANKER
CC:	MURPHY
Subj:	RE: Architect/Darwinian paper, version 2

    I was interested to read your discussion of the two models for computer
    vendor and consumer companies.  It parallels a number of observations I
    have made over the years, such as:

    "Digital can't ever do two things at once".  We can't do 36-bit *and*
    32-bit; we can't do VMS *and* Unix; we can't do VAX *and* RISC.  We
    show great corporate discomfort whenever we can't decide that there is
    one "best" thing and do that.  This has been particularly true ever
    since the "One Company, One Strategy, One Message" initiative which, in
    your model, probably marks the point at which DEC fully changed from a
    Darwinian company (which I believe it was in the early days) to an
    Architectural company.

    It is also how remarkable how your description of the Darwinian company
    fits another well-known player in the industry today.  You didn't
    mention this company, perhaps for political reasons, but the
    description fits Microsoft perfectly.  Microsoft doesn't seem to worry
    about whether their various products all conform to some dominant
    architecture, or even consistent interface standard. Microsoft will
    drop something in an instant if it looks like it won't be the winner,
    and they somehow never find themselves with big, money-losing products
    that they have to keep developing and delivering because they
    "promised" it to some customer.

    Microsoft could change, as DEC did, but the way Gates runs it now, that
    prospect doesn't appear imminent.

    And I doubt DEC can change back unless, perhaps, it is busted up into a
    bunch of largely independent product lines under managers who aren't
    invested in the architectural world view.


    Cheers,


    dlm

re: .31

	, albeit in a BIG niche.  In the beginning all computer companies had to
do it all; hardware, software, middleware, networks....Then as the industry
developed, niche players came along and specialized in one area or another.

This is similar in some ways to the recording or TV industries; is anyone else
old enough to remember when the same companies that built the TVs put on the
programs? This is why all founding fathers are architects in the sense that
Anker meant in the base note. You couldn't have a software company before there
was hardware and the hardware won't work without software hence the earlies
companies had to do both.  Then after time, the niche players could move in on
those areas that could be done independently.

	The Darwinian approach "A Darwinian will
        also assume that someone else is working on all the other
        problems of the world" can't work when there is no-one else in the 
	industry.

We had to be architecturally driven, even if the people here were Darwinians at
heart, years ago. One of the hardest things to do is how to break up a
successful, vertically integrated operation to compete in a changed industry.

Spinning-off our niche expertise in someway is probably the right answer. But,
does that mean total divestiture, wholly owned subsidiary, division or what?
THAT is the top question for TOP management today.

What businesses do we want to be in and how?

The fact that you're all replying to this note is evidence of the success 
of quite a few architectures -- VAX and DECnet IV, to name but
two of the many.

Architecture in itself is not a bad thing.   What is a bad thing is
the insistence on Grand Unified Everything architectures.

    Anker and Colleagues;
    
    Thank you for such a stimulating paper and delightful discussion.
    
    I was reminded of a paper I read 10 years ago and long since lost that
    characterized 3 management styles. The connection is that the 3
    management styles appeared to be a function of whole brain integration
    and stages of enlightenment. I have since lost the paper and it may
    have been written by Noel Tichy et al but I'm not sure.
    
    What if the two paradigms that Anker suggests are merely two dimensions
    of what may be a three dimensional model. Perhaps I can be so bold as
    to suggest a higher level paradigm that may embrace both of Ankers
    views.
    
    If we view the dimensions similar to Mazlow's hierarchy of needs with
    the vertical axis as degrees of whole brain integration I might suggest
    at least three of the levels.
    
    At the bottom level, the primitive level we have behaviour that is
    survival of the fittest. In the management style paper this was called
    fear state management because it suggested tactics of control through
    style. The question becomes,"Are Darwinian companies motivated by
    fear."
    
    The level above the fear state level was solid state. The paper
    suggested that "vision" was a key element of this style. Managers
    viewed their world in structural and architectural terms, rather like
    the military command model with everyone knowing their place in the
    order and the performance management system providing friendly
    controls. These managers were participative facilitators whose job it
    was to explain and elucidate the structure or architecture. The
    managers were the custodians of the architecture.Such devices as "visions"
    and "missions" with a loose planning framework abound in this style.
    
    The last style was termed "flow state" and was most akin to a Zen
    model. There is an underlying central order or architecture. Everything
    is connected and affects everything else but the system operates on the
    energy flowing through it. The role of the manager was to nurture and
    facilitate the energy flows, much like an acupuncturist.
    
    My thesis here is that Digital needs to comprehend and faciliate our
    architectural views but exploit opportunities based on where the
    energy is blocked. I envisage this as an opportunity search solving
    immediate problems but within the context of architecture, not
    disregarding architecture and letting someone else worry about it.
    
    This model may be similar to the observation of the free market system
    within the US. The energy flows are the opportunists working within the
    architectural structure created by anti trust and other laws. In this
    model the Darwinian and Architectural views coexist in support of one
    another.
    
    Let's make Digital a Zen company!! and not a fear state or solid state
    organization. Anyone volunteer the mantra?

    >Let's make Digital a Zen company!! and not a fear state or solid state
    >organization. Anyone volunteer the mantra?
    
    How 'bout "O(h)m's Law?"  ;-)

    Interesting that some of the observations herein are not limited to DEC
    insiders. The following is a quote from a report issued recently by
    Forrester Research, Inc. which has been circulating widely on the
    enet:

        Digital's management seems to believe that there must be a top-down 
        strategy for the whole company - "client/server," or "open," or
        whatever - and that everyone must then hew to that strategy. The
        markets now move too fast for this. The strategy should be an
        "anti-strategy" - a statement to every group that they should
        create the best product in their market or get out. Let a thousand
        flowers bloom. As Jack Welch stated in the 1993 GE annual report,
        "You must take your foot off the brake before you can step on the
        gas."

        We believe that there is still gas in the warehouse to accelerate.
        If you cut into the heart of Digital, you will find engineering
        excellence. But this excellence is in chains - shackled by fear of
        layoffs, overcontrolled by the bean counters, and not given the
        room to succeed (or fail). Only radical moves will smash the chains
        - like breaking up the company. Create 30-50 "Digital Labs" -
        independent businesses with some shared resources, ranging from $10
        million-$50 million in size. Each would be self-contained with
        autonomous presidents, boards of directors, sales forces, and
        engineers. Run Digital like Emerson Electric - 40 independent
        companies orbiting an aggressively managed holding company.


    dlm