Home
Up
| |
On Holistic and Reductionist Automation
Holistic - the theory that certain wholes are to be regarded as greater
than the sum of their parts
Reductionist - 1. the tendency to or principle of analyzing complex
things into simple constituents. 2. often derog. the doctrine that a system can
be fully understood in terms of its isolated parts, or an idea in terms of
simple concepts.
If the founding fathers (a US centric reference) had personal computers, how
would they have constructed the Declaration of Independence? They did not work
for the same company and they did not have much time to plan and implement a lot
of infrastructure decisions. Thomas Jefferson would most likely be an Apple Mac
user. Ben Franklin was definitely the home configured, Pentium, SCSI, Local bus,
backed up to optical disk, OS/2 type. The rest would be a mix of Windows, DOS
and maybe an Amstrad or two for the new arrivals. Of course, someone would have
used spaces and carriage returns instead of tabs and page breaks in order to
create more havoc. They would, of course, need a wide area network with multiple
protocols installed. The revision control software would need to be very
advanced too, in order to distribute the marked up versions for review. I really
don't think it could have been done this way, its probably a good thing they did
not have PC's back then.
Managing data was simpler when there were only minicomputers and mainframes.
Mainframes, by nature of the cost, were part of a longer term holistic plan,
everything related to data was a subset of this main system. This model had
easily identified, high cost line items which required board level decisions.
The industry is discovering that comparing the costs of the previous mainframes
versus Open Systems/LAN technology is difficult at best and there does not seem
to be the massive cost savings that people thought there would be. Instead, the
customer has gained more control of the mixture of hardware and software at the
cost of quietly assuming the responsibility to make all of the diverse
components work together. The expertise needed to integrate systems is in
relatively short supply and customers are paying dearly for these services.
Why has the PC/LAN become so popular in business? One of the main reasons is
that the traditional guardians of the corporate data, the mainframers, became
complacent and said "NO" too often. The phrase 'invisible backlog' was
coined to describe the list of things that users wanted but could not have
implemented so they didn't even bother to ask for. The mainframers did not keep
up with demand, and the market, through the PC, found a way to satisfy the
demand. PC based systems tend to be reductionist and solve only a small portion
of the problem and are generally labor intensive. Implementing a series of
reductionist systems is orders of magnitude easier than solving the larger
problems but the true cost of multiple reductionist systems appears when the
attempt is made to scale up the application. The ability to produce reports on a
PC that could not be done on a mainframe led to the 'mainframes are expensive
and ineffective and PC's are better and cheaper' argument
An Open/LAN system can be either holistic or reductionist or a little of both.
It is difficult in this market to differentiate holistic, mission critical
systems from personal, reductionist systems. Many companies are trying to cobble
together reductionist systems with LANs and middleware products in order to
label them holistic systems. It is easy to make fish stew from a fish, but
nobody has found a way to make a fish out of fish stew. The data that makes up a
company is the property of that company. When a company’s real data is spread
out among multiple systems it is very hard to ensure the quality and security of
the data. Without a holistic system in place, ensuring data usefulness,
integrity and security is the same as trying to make a fish from fish stew.
Today, driven by huge marketing machines, users have access to tools which they
are unqualified to judge, implement and use (as are many of the salespeople for
these tools, unfortunately). Hidden in this situation are incredible
inefficiencies and dormant disasters. The asset that was lost in the shift from
mainframe to LAN models was the concentration of system experts (I use the word
system to mean the entire workflow process not the narrower computer 'system').
This loss of expertise has dawned on many companies and is part of the recent
drive to business process reengineering (BPR). The challenge for IT
professionals is to place themselves between the user and the technology and
learn to say "YES". There is no substitute for knowledge and
expertise!
Software should be defined as a collection of manual procedures and workflow
processes in addition to instructions to a computer. Examples of this definition
are the popular Personal Information Managers (PIMS) like ACT! from Semantic.
These products embody the procedures of successful salespersons in them. These
products impose a set of guidelines to follow in order to perform organized
contact management, it is usually possible to easily detect inexperienced cold
callers who use one of these products For many people who were previously
disorganized, these products impose much needed order on their work and they are
loyal to the product. If this level of performance is acceptable to a company
then there is no need to demand any further innovation. This level of
performance is rarely acceptable though. Successful salespeople have been
practicing the procedures now embedded in the contact software for years using
methods they developed for themselves. In the process of developing their own
procedures the salespeople became experts in their jobs and could offer new
insights to the problems of sales. Many companies are now benefiting from the
expertise developed by these innovative salespeople, when this knowledge is
combined with new information technology are able create true 'world class'
organizations. |
