A Short History of Systems Development

"If they do not have an appreciation of whence we				that would enable an integrated environment.
came, I doubtthey will have an appreciation of where				Consequently,the product ran aground not long after
we should be going."				it was launched. In fact, theprosperity of the CASE
- Bryce's Law				market was short-lived as customers failed to
INTRODUCTION				realizethe savings and productivity benefits as touted
I always find it amusing when I tell a young person in				by the vendors. By theearly 1990's, the CASE market
this industry that Iworked with punch cards and				was in sharp decline.
plastic templates years ago. Its kind of thesame				Instead, companies turned to Programmer
dumbfounded look I get from my kids when I tell				Workbenches which includedan all-in-one set of basic
them we used towatch black and white television				tools for programming, such as editing, testing,and
with three channels, no remote control,and station				debugging. Microsoft and Micro Focus did particularly
signoffs at midnight. It has been my observation that				well in offeringsuch products.
our youngerworkers do not have a sense of history;				Data Base Management Systems also took a
this is particularly apparent in thesystems world. If				noticeable turn in the 1980'swith the advent of
they do not have an appreciation of whence we				"relational" products involving tables and keys.
came,				Theconcept of the "relational" model was originally
I doubt they will have an appreciation of where we				developed by IBM
should be going. Consequently,				Fellow and mathematician Edgar (Ted) Codd in a
I have assembled the following chronology of events				paper from 1970. The conceptof a relational DBMS
in the hopes this will providesome insight as to how				was superior to the earlier network and
the systems industry has evolved to its current				hierarchicalmodels in terms of ease of use. The
state.				problem resided in the amount ofcomputer
I'm sure I could turn this into a lengthy dissertation				horsepower needed to make it work, a problem that
but, instead, I will tryto be brief and to the point.				wasovercome by the 1980's. As a result. new DBMS
Further, the following will have little concernfor				products such as Oracle and
academic developments but rather how systems				Ingres were introduced which quickly overtook their
have been implementedin practice in the corporate				older competitors. Therewas an initial effort to
world.				convert DBMS mainstays such as TOTAL, ADABAS,
PRE-1950'S - "SYSTEMS AND PROCEDURES"				and
Perhaps the biggest revelation to our younger				IDMS into relational products, but it was too little, too
readers regarding this periodwill be that there was				late. As for IBM,they simply re-labeled their flagship
any form of systems prior to the advent of				product, IMS, as a "transaction processor"and
thecomputer. In fact, "Systems and Procedures"				introduced a totally new offering, DB2, which quickly
Departments predated thecomputer by several				dominated the
years. Such departments would be concerned with				DBMS mainframe market.
thedesign of major business processes using "work				Program generators continued to do well during the
measurement" and "worksimplification" techniques as				1980's but it was duringthis period that 4GL's (fourth
derived from Industrial Engineering. Suchprocesses				generation languages) were introduced toexpedite
were carefully designed using grid diagrams and				programming. The 4GL was a natural extension of
flowcharts. Therewas great precision in the design of				the DBMS andprovided a convenient means to
forms to record data, filing systems tomanage				develop programs to interpret data in thedata base.
paperwork, and the use of summary reports to act				Another development worth noting is the evolution
as controlpoints in systems. For example,				of the Data Dictionaryinto "Repositories" (also
spreadsheets have been extensively usedfor many				referred to as "Encyclopedias") used to store
years prior to the introduction of Lotus 1-2-3 or MS				thedescriptions of all of an organization's information
Excel. There wasalso considerable attention given to				resources. One of themotivating factors behind this
human behavior during the businessprocess (the				was IBM (for AD/Cycle) who realized theyneeded
precursor to "ergonomics").				some sort of cohesive bond for the various CASE
Systems were initially implemented by paper and				tools to interface. Thisis another area pioneered by
pencil using ledgers, journals				MBA who introduced their "PRIDE"-Enterprise
(logs), indexes, and spreadsheets. We have always				Engineering Methodology (EEM) to study a business
had some interesting filingsystems, everything from				and formulate an
cards and folders, to storage cabinets.				Enterprise Information Strategy, and their
Perhaps the earliest mechanical device was the				"PRIDE"-Data Base Engineering
ancient abacus used for simplemath (which is still used				Methodology (DBEM) to develop the corporate data
even to this day). The late 1800's saw the advent of				base, both logicallyand physically. To implement these
cashregisters and adding machines as popularized by				new methodologies, their
such companies as NCR in				"PRIDE"-LOGIK Dictionary was expanded to include
Dayton, Ohio under John Patterson who also				business models, anddata models. By doing so, MBA
introduced sweeping changes interms of dress and				renamed "PRIDE"-LOGIK the "PRIDE"-IRM
business conduct. This was adopted by Thomas				(Information Resource Manager) which
Watson, Sr.who worked for many years at NCR and				complemented their concept of
carried forward these practices to IBMand the rest				Information Resource Management.
of the corporate world. Also, Burroughs was a major				In terms of the MIS infrastructure, two noteworthy
player inthe early adding machine industry.				changes occurred;first was the introduction of the
The first typewriters were also introduced in the late				Chief Information Officer (CIO) as firstdescribed in
1800's which had a tremendouseffect on				the popular book, "Information Systems Management
correspondence and order processing. This was led				In Practice"
primarily by Remington				(McNurlin, Sprague) in January 1986. Basically, the MIS
Arms (later to become Remington Rand).				Director is elevatedto a higher management level
In the early 1900's, tabulating equipment was				where, theoretically, he/she is operating on thesame
introduced to support such thingsas census counting.				level as the Chief Operating Officer (COO), and Chief
This was then widely adopted by corporate America.				Financial Officer
Occasionallyyou will run into old-timers who can				(CFO) for a company. In reality, this has never truly
describe how they could program such machinesusing				happened and, in manycases, the title "CIO" is nothing
plug boards. Punch card sorters were added as an				more than a change in name, not in stature.
adjunct to tabulatingequipment.				The second change is the change in job title of
As a footnote, most of what IBM's Watson learned				"Programmer" to "Software
about business was fromhis early days at NCR.				Engineer." Again, we are primarily talking about
However, he had a falling out with Patterson who				semantics. True, manyof the programmers of the
firedhim. As a small bit of trivia, after Watson died,				1980's studied Structured Programming, butvery few
he was buried in Dayton on ahilltop overlooking NCR				truly understood the nature of engineering as it
headquarters, the company he couldn't conquer.				applies tosoftware, most are just glorified coders.
During World War II, both the U.S. military and				Nonetheless, the "Software
industrial complex relied heavilyon manually				Engineer" title is still actively used today. In contrast,
implemented systems. We did it so well that many				the last of thetrue "Systems Analysts" slowly
people, includingthe Japanese, contend it gave the				disappeared. Here too is evidence ofthe change of
Allies a competitive edge during the war.				focus from systems to software.
The lesson here, therefore, is that manually				During the 1980's we also saw the emergence of
implemented systems have been withus long before				MBA's graduating fromthe business schools and
the computer and are still with us today. To give you				working their way into the corporate landscape.
a sense ofhistory in this regard, consider one of our				Althoughthey didn't have an immediate impact on the
more popular Bryce's Laws:				systems world, they had a dramaticeffect on the
"The first on-line, real-time, interactive, data base				corporate psyche. Their work resulted in severe
system was double-entrybookkeeping which was				corporate cutbacks,downsizing, and outsourcing. This
developed by the merchants of Venice in 1200 A.D."				changed the corporate mindset to thinkshort-term as
One major development in this area was the work				opposed to long-term. Following this, companies shied
of Leslie "Les" Matthies, thelegendary Dean of				away frommajor systems projects (such as the MIS
Systems. Les graduated from the University of				projects of the 1960's) and were contenttackling
California at				smaller programmer assignments, thus the term "app"
Berkeley during the Depression with a degree in				was coined todescribe a single program application.
Journalism. Being a writer, hetried his hand at writing				Interestingly, a "quality" movement flourished in the
Broadway plays. But work was hard to come by				1980's based on theworks of W. Edwards Deming
duringthis period and when World War II broke out,				and Joseph M. Juran who pioneered qualitycontrol
Les was recruited by an aircraftmanufacturer in the				principles in the early part of the 20th century.
midwest to systematize the production of aircraft.				Unfortunately, theirearly work was unappreciated in
Relyingon his experience as a writer, he devised the				America and, consequently, they appliedtheir talents
"Playscript" technique for writingprocedures. Basically,				to help rebuild the industrial complex of postwar
Les wrote a procedure like a script to a play; there				Japan. It wasonly late in their lives did they receive
was asection to identify the procedure along with its				the recognition of their work in the
purpose; a "Setup" section to identifythe forms and				United States (after Japan became an economic
files to be used during it; and an "Operations				powerhouse). Anotherinfluential factor was the
Instructions" sectionwhich described the "actors" to				introduction of the ISO 9000 standard for
perform the tasks using verbs and nounsto properly				qualitymanagement which was originally devised by
state each operation. He even went so far as to				the British and later adopted asan international
devise rules forwriting "If" statements.				standard. Little attention would probably have been
For details on "Playscript," see "PRIDE" Special				paid to
Subject Bulletin				ISO 9000 if it weren't for the fact that European
No. 38 - "The Language of Systems" - Aug. 22, 2005				businesses started to demandcompliance in order to
"Playscript" became a powerful procedure writing				conduct business with their companies.
language and was usedextensively throughout the				Nevertheless, these factors resulted in a reorientation
world. It is still an excellent way to write				of Americanbusinesses to think in terms of
procedurestoday. Ironically, Les did not know what a				developing quality products which,inevitably, affected
profound effect his technique wouldhave later on in				how systems and software were produced. The real
the development of computer programs.				impactof the quality movement though wouldn't be
1950'S - INTRODUCTION OF THE COMPUTER				felt in the systems world untilthe next decade.
Yes, I am aware that the ENIAC was developed for				To summarize the 1980's from a systems
the military at the endof World War II. More				development perspective, the focusshifted away
importantly, the UNIVAC I (UNIVversal Automatic				from major systems to smaller programming
Computer)was introduced in 1951 by J. Presper				assignments whichwere implemented using newly
Eckert and John Mauchly. The UNIVAC Iwas a				devised CASE tools. This fostered a
mammoth machine that was originally developed for				"tool-orientedapproach" to development whereby
the U.S. Bureau ofthe Census. Corporate America				companies spent considerably on thelatest
took notice of the computer and companiessuch as				programming tools but little on management and
DuPont in Wilmington, Delaware began to lineup to				upfront systemswork. In other words, they bought
experiment withit for commercial purposes. The				into the vendor's claims of improvedprogrammer
Remington Rand Corporation sponsored theproject,				productivity through the use of tools. Unfortunately,
but the company's focus and name eventually				it resultedin patchwork systems that required more
changed to "UNIVAC"				time in maintenance as opposed tomodifying or
(today it is referred to as "UNISYS," representing a				improving systems. "Fire fighting" thereby became
merger of UNIVAC with				the normalmode of operation in development.
Burroughs).				1990's - REDISCOVERY
The UNIVAC I offered a sophistication unmatched by				As the PC gained in stature, networking became
other manufacturers,most notably IBM's Mach I				very important tocompanies so that workers could
tabulating equipment. This caused IBM to inventthe				collaborate and communicate on acommon level. Local
701 and its 700 series. Other manufacturers quickly				Area Networks (LAN) and Wide Area Networks
joined the fray andcomputing began to proliferate.				(WAN)seemed to spring-up overnight. As the PC's
Although UNIVAC was the pioneer in thisregard, they				power and capacity grew, itbecame obvious that
quickly lost market share due to the marketing				companies no longer needed the burden of
muscle of				mainframesand minis. Instead, dedicated machines
IBM. For quite some time the industry was referred				were developed to control andshare computer files,
to as "IBM & the				hence the birth of "client/server computing"
BUNCH" (Burroughs, UNIVAC, NCR, CDC, and				whereclient computers on a network interacted with
Honeywell).				file servers. This did notcompletely negate the need
Programming the early machines was difficult as it				for mainframes and minis (which were alsoused as file
was performed in aseemingly cryptic Machine				servers), but it did have a noticeable impact on sales.
Language (the first generation language).				Companiesstill needed mainframes to process
Thiseventually gave way to the Assembly Language				voluminous transactions and
(the second generationlanguage) which was easier to				extensivenumber-crunching, but the trend was to
read and understand. Regardless, many ofthe utilities				move away from big iron.
we take for granted today (e.g., sorts and merges)				Thanks to the small size of the PC, companies no
simply werenot available and had to be developed. In				longer required a bigroom to maintain the computer.
other words, programming was alaborious task during				Instead, computers were keptin closets and under
this period.				desks. This became so pervasive that companies
Recognizing both the limitations and potential of the				nolonger knew where their computer rooms were
computer, the 1950'srepresented the age of				anymore. In a way, thespread of computers and
experimentation for corporate America. Here,				networks closely resembled the nervous systemof
theemphasis was not on implementing major systems				the human body.
through the computer,but rather to develop an				One of the key elements that made this all possible
assortment of programs to test the machine as				was the introductionof Intel's 30386 (or "386") chip
aviable product. As such, programmers were				which allowed 32-bit processing. To effectivelyuse
considered odd characters whomaintained "the black				this new technology, new operating systems had to
box," and were not yet considered a part of				be introduced, thefirst being IBM's OS/2 in the late
themainstream of systems development. The				1980's. OS/2 provided such things asvirtual memory,
"Systems and Procedures				multitasking and multithreading, network
Departments" still represented the lion's share of				connectivity,crash-protection, a new High
systems work in corporate				Performance File System, and a slick objectoriented
America, with an occasional foray to investigate the				desktop. Frankly, there was nothing else out there
use of the computer.				that couldmatch it. Unfortunately, Microsoft bullied its
The computer people were segregated into				way past OS/2 with
"computer departments"				Windows 95 & NT. By the end of the 1990's, OS/2
(later to be known as "EDP" or "Data Processing"				was all but forgottenby its vendor, IBM. Nevertheless,
departments).				it was the advent of 32-bit computingthat truly made
1960's - MANAGEMENT INFORMATION SYSTEMS				client/server computing a reality.
Competition between computer manufacturers				Another major milestone during this decade was the
heated up during thisdecade, resulting in				adoption of the
improvements in speed, capacity, and capabilities.				Internet by corporate America. The Internet actually
Of importance here was the introduction of the				began in the late
much touted IBM 360				1960's under the Department of Defense and was
(the number was selected to denote it was a				later opened to othergovernment and academic
comprehensive solution -				bodies. But it wasn't until the 1990's thatcompanies
360 degrees). Other computer vendors offered				started to appreciate the Internet as a
products with comparableperformance, if not more				communications andmarketing medium.
so, but the IBM 360 was widely adopted				The first web browser was developed by Tim
bycorporate America.				Berners-Lee in 1990 whichled to the World Wide
The programming of computers was still a difficult				Web protocol on the Internet. Early web
task and, consequentially,				browsersincluded Mosaic, Netscape Navigator, and
Procedural Languages were introduced (the third				Microsoft's Internet Explorer,among others. The
generation languages). Inactuality, these languages				beauty of the Internet was that all computers
got their start in the late 1950's, but the				couldnow access the Internet regardless of the
proliferationof computers in the 1960's triggered the				operating system, making it atruly universal approach
adoption of procedural languages suchas COBOL,				to accessing data. To write a web page, a simpletag
FORTRAN, and PL/1. Interestingly, these languages				language was devised, Hyper Text Markup Language
were patternedafter Les Matthies' "Playscript"				(HTML), which wascompiled at time of request to
technique which made active use of verbs,nouns, and				display the web page. HTML was nice fordeveloping
"if" statements.				simple static web pages (not much interaction, just
The intent of the Procedural Languages was twofold:				simply view theweb page). Developers then invented
to simplify programmingby using more English-like				new techniques to make a web pagemore dynamic
languages, and; to create universal languagesthat				thereby allowing people to input data and interact
would cross hardware boundaries. The first goal was				with files,which ultimately allowed for the
achieved, thesecond was not. If the languages were				merchandising of products over the Internet.
truly universal, it would mean thatsoftware would be				Wanting to do something more sophisticated through
portable across all hardware configurations.				the web browser,
Manufacturerssaw this as a threat; making software				Sun Microsystems developed the Java programming
truly portable made the selection ofhardware				language in 1995. Javawas a universal programming
irrelevant and, conceivably, customers could migrate				language that could run under any operatingsystem.
away fromcomputer vendors. In order to avoid this,				Their mantra was "Write once, run anywhere." This
small nuances were introducedto the compilers for				was a radicaldeparture from programming in the past
the Procedural Languages thereby negating the				where it was necessary to recompileprograms to suit
conceptof portability. This issue would be ignored for				the peculiarities of a particular operating system.
many years until the adventof the Java programming				Basically,
language.				Java made the operating system irrelevant, much to
The 1960's also saw the introduction of the Data				Microsoft's chagrin. Further,
Base Management System				Java could be used in small pocket devices as well as
(DBMS). Such products were originally designed as file				in the new generationof computers powering
access methods for				automobiles. This did not sit well with Microsoft
Bill of Materials Processing (BOMP) as used in				whoultimately fought the propagation of Java.
manufacturing. The "DBMS"designation actually came				By the 1990's the Structured Programming
afterwards. Early pioneers in this area included				movement had fizzled out. Instead,
Charlie Bachman of G.E. with his Integrated Data				"Object Oriented Programming" (OOP) gained in
Store (IDS) whichprimarily operated under Honeywell				popularity. The concept of OOPwas to develop
GCOS configurations; Tom Richleyof Cincom Systems				bundles of code to model real-world entities such as
developed TOTAL for Champion Paper, and; IBM's				customers,products, and transactions. OOP had a
BOMPand DBOMP products. In 1969, IBM introduced				profound effect on Java as well as the
IMS which became theirflagship DBMS product for				C++ programming language.
several years.				During this time, source code generators faded from
With the exception of IMS, the early DBMS offerings				view. True, companies werestill using report writers
were based on anetwork model which performed				and 4GL's, but the emphasis turned to "Visual
chain-processing. IMS, on the otherhand, was a				Programming"which were programming workbenches
hierarchical model involving tree-processing.				with screen painting tools to layoutinputs and outputs.
Realizing that programming and data access was				The Relational DBMS movement was still in high gear,
becoming easier andcomputer performance being				but the use of Repositoriesand Data Dictionaries
enhanced, companies now wanted to capitalizeon this				dropped off noticeably. Of interest though was
technology. As a result, corporate America embarked				theintroduction of "Object Oriented Data Base
on the eraof "Management Information Systems"				Management System" (OODBMS)technology. Like
(MIS) which were large systemsaimed at automating				OOP, data was organized in a DBMS according to
business processes across the enterprise. These				real-worldentities. Regardless, Relational DBMS
weremajor system development efforts that				dominated the field.
challenged both management andtechnical expertise.				Also during this decade "Data Mining" became popular
It was the MIS that married "Systems and				whereby companies wereprovided tools to harvest
Procedures" departments withcomputing/EDP				data from their DBMS. This effort was basically an
departments and transformed the combined				admissionthat companies should learn to live with data
organization intothe "MIS" department. This was a				redundancy and not be concernedwith developing a
major milestone in the history of systems.				managed data base environment.
Thesystems people had to learn about computer				Because of the radical changes in computer hardware
technology and the programmershad to learn about				and software, companies becameconcerned with
business systems.				their aging "legacy" systems as developed over the
Recognizing that common data elements were used				last thirty years. Tomigrate to this new technology, a
to produce the variousreports produced from an				movement was created called "Business Process
MIS, it started to become obvious that data shouldbe				Re-engineering" (BPR). This was encouraging in the
shared and reused in order to eliminate redundancy,				sense that companies werestarting to think again in
and to promotesystem integration and consistent				terms of overall business systems as opposed to
data results. Consequently, Data Management				justprograms. I'm not sure I agree with the use of
(DM) organizations were started, the first being the				the term "Re-engineering" though;this assumes that
Quaker Oats Company in				something was engineered in the first place (which
Chicago, Illinois in 1965. The original DM organizations				was hardly thecase in these older systems).
were patterned after				Nonetheless, CASE-like tools were introduced to
Inventory Control Departments where the various				define business processes. Suddenly,companies were
components were uniquelyidentified, shared and				talking about such things as "work flows,"
cross-referenced. To assist in this regard, such				"ergonomics," and
organizationsmade use of the emerging DBMS				"flowcharts," topics that had not been discussed for
technology. Unfortunately, many DMorganizations lost				twenty years during the frenzy ofthe Structured
sight of their original charter and, instead, became				Programming movement. Ultimately, this all led to the
obsessedwith the DBMS. Data as used and				rediscovery ofsystems analysis; that there was more
maintained outside of the computer waserroneously				to systems than just software. But by this time,all of
considered irrelevant. Even worse, the DBMS was				the older corporate Systems Analysts had either
used asnothing more than an elegant access method				retired or been put out to pasture,leaving a void in
by programmers. Consequently,data redundancy				systems knowledge. Consequently, the industry
plagued systems almost immediately and the				started to relearnthe systems theory, with a lot of
opportunity toshare and reuse data was lost. This is				missteps along the way.
a serious problem that persists incompanies to this				Companies at this time were still struggling with
day.				devising a suitable developmentenvironment. Most
1970's - AWAKENING				were content with just maintaining their current
Although the MIS movement was noble and				systemsin anticipation of the pending Y2K (Year
ambitious in intent, it floundereddue to the size and				2000) problem (where date fields were tochange
complexity of the task at hand. Many MIS projects				from 19XX to 20XX which could potentially shutdown
sufferedfrom false starts and botched				companies). However,a few companies began to
implementations. This resulted in a period wherea				consider how to apply more scientific principles to
series of new methods, tools and techniques were				theproduction of systems. Since people were already
introduced to reign in thesehuge development				talking about "Software
efforts.				Engineering," why not apply engineering
The first was the introduction of the "methodology"				manufacturing principles to thedevelopment of total
which provided a road mapor handbook on how to				systems?
successfully implement systems				Back in the early 1980's, Japan's Ministry of
developmentprojects. This was pioneered by MBA				International Trade & Industry (MITI)coordinated a
with its "PRIDE" methodology in				handful of Japanese computer manufacturers in
1971. Although the forte of "PRIDE" was how to build				establishing aspecial environment for producing
systems, it was initiallyused for nothing more than				system software, such as operating systemsand
documentation and as a means to manage projects.				compilers. This effort came to be known as
Following "PRIDE" was John Toellner's Spectrum I				Japanese "Software Factories"which captured the
methodology and SDM/70from Atlantic Software.				imagination of the industry. Although the
Several CPA based methodologies followed				experimentended with mixed results, they discovered
thereafter.				organization and disciplinecould dramatically improve
Also during this time, mainframe based Project				productivity.
Management Systems werecoming into vogue				Why the experiment? Primarily because the Japanese
including Nichols N5500, PAC from International				recognized there arefundamentally two approaches
Systems,and PC/70 from Atlantic Software.				to manufacturing anything: "one at a time" ormass
The early methodologies and Project Management				production. Both are consistent approaches that can
Systems give evidence ofthe orientation of systems				produce a highquality product. The difference resides
departments of that time: a heavy emphasis on				in the fact that mass production offersincreased
Project				volume at lower costs. In addition, workers can be
Management. Unfortunately, it was a fallacy that				easily trainedand put into production. On the other
Project Management wasthe problem; instead people				hand, the "one at a time" approach isslower and
simply didn't know how to design and buildsystems in				usually has higher costs. It requires workers to be
a uniform manner. As companies eventually learned,				intimate withall aspects of the product.
Project				MBA took it a step further by introducing their
Management is useless without a clear road map for				concept of an "Information
how to build something.				Factory" in the early 1990's. The Information Factory
In the mid-to-late 1970's several papers and books				was a comprehensivedevelopment environment
were published on howto productively design				which implemented MBA's concept of Information
software thus marking the beginning of the				Resource Management. Basically, they drew an
"Structured				analogy between developingsystems to an
Programming" movement. This was a large body of				engineering/manufacturing facility, complete with
work that included suchprogramming luminaries as				assembly lines,materials management and production
Barry Boehm, Frederick P. Brooks, Larry Constantine,				control. These concepts were proveneffective in
Tom DeMarco, Edsger Dijkstra, Chris Gane, Michael A.				companies throughout Japan, most notably Japan's
Jackson, Donald E. Knuth,				BEST project,which was sponsored by the Ministry
Glenford J. Myers , Trish Sarson, Jean Dominique				of Finance. As background, the ministrywanted to
Warnier, Generald M. Weinberg,				leapfrog the west in terms of banking systems. To
Ed Yourdon, as well as many others. Although their				do so, theyassembled a team of over 200 analysts
techniques were founduseful for developing				and programmers from four of the toptrust banks in
software, it led to confusion in the field				Japan; Yasuda Trust & Banking, Mitsubishi Trust &
differentiatingbetween systems and software. To				Banking,
many, they were synonymous. In reality,they are				Nippon Trust & Banking, and Chuo Trust & Banking.
not. Software is subordinate to systems, but the				By implementing MBA'sconcepts they were able to
growing emphasison programming was causing a				deliver over 70 major integrated systems in lessthan
change in perspective.				three years. Further, because they had control over
The only way systems communicate internally or				their informationresources using a materials
externally to other systemsis through shared data; it				management philosophy, the Y2K problemnever
is the cohesive bond that holds systems (and				surfaced.
software)together. This resulted in the introduction				In terms of infrastructure, development organizations
of Data Dictionary technology. Again,this was				essentially wentunchanged with a CIO at the top of
pioneered by MBA with its "PRIDE" methodology				the pyramid and supported by
(which included a manuallyimplemented Data				Software Engineers and DBA's. But there was one
Dictionary) and later with its "PRIDE"-LOGIK product				slight difference,instead of being called an MIS or IS
in				department, the organization wasnow referred to as
1974. This was followed by Synergetics' Data				"IT" (Information Technology). Here again, the name
Catalogue, Data Manager from				hintsat the direction most organizations were taking.
Management Software Products (MSP), and Lexicon				Finally, the 1990's marked a change in the physical
by Arthur Andersen & Company.				appearance ofthe work force. Formal suit and ties
The intent of the Data Dictionaries was to uniquely				gave way to casual Polo shirts and
identify and track wheredata was used in a				Docker pants. At first, casual attire was only allowed
company's systems. They included features for				on certain days
maintainingdocumentation, impact analysis (to allow				(such as Fridays), but it eventually became the
the studying of a proposed change),and redundancy				normal mode of dress. Unfortunately,many people
checks. "PRIDE"-LOGIK had the added nuance of				abused the privilege and dressed slovenly for work.
catalogingall of the systems components, thereby				This hada subtle but noticeable effect on work
making it an invaluable aid fordesign and				habits, including how we build systems.
documentation purposes.				2000's - GADGETS
The Data Dictionary was also a valuable tool for				We are now past the halfway point in this decade
controlling DBMS productsand, as such, several				and there is nothing ofsubstance to report in terms
adjunct products were introduced, such as UCC-10,				of computer hardware, other than our machineshave
DB/DC Data Dictionary, and the Integrated Data				gotten faster, smaller, with even more capacity.
Dictionary (IDD) from				Perhaps the biggest innovationin this regard is the
Cullinet. Unlike the other general purpose Data				wide variety of "gadgets" that have been introduced,
Dictionaries, these productswere limited to the				all ofwhich interface with the PC, including: Personal
confines of the DBMS and didn't effectively track				Digital Assistants (PDA's), iPods,
dataoutside of their scope.				MP3 players, digital cameras, portable CD/DVD
DBMS packages proliferated during this period with				players (and burners), cell phones,
many new productsbeing introduced including				PS2 and XBox game players. These devices are
ADABAS, Image, Model 204, and IDMS from				aimed at either communications orentertainment,
Cullinet (which was originally produced at BF				giving us greater mobility, yet making us a bit
Goodrich). All were basedon the network-model for				dysfunctionalsocially. All of this means the computer
file access which was finally adopted as anindustry				has become an integral part ofour lives, not just at
standard (CODASYl).				work but at home as well.
There were a few other notable innovations				Shortly after taking the reigns of IBM in 2003, CEO
introduced, including IBM's				Sam Palmisano introduced
Business Systems Planning (BSP) which attempted to				"On-Demand Computing" as the company's thrust for
devise a plan forthe types of systems a company				the years ahead and, inevitably,it will mark his legacy.
needed to operate. Several other				The concept as described by Palmisano was
comparableofferings were introduced shortly				simple,treat computing like a public utility whereby a
thereafter. Interestingly, many companiesinvested				company can draw upon IBM forcomputing resources
heavily in developing such systems plans, yet very				as required. "On-Demand Computing" made a
few actuallyimplemented them.				nicecatch-phrase and was quickly picked up by the
Program Generators were also introduced during this				press, but many peoplewere at a loss as to what it
period. This includedreport writers that could interpret				was all about. Some of the early
data and became a natural part of therepertoire of				developmentsresulting from IBM's "e-Business On
DBMS products. It also included products that could				Demand" research included balancing the loadon file
generateprogram source code (COBOL				servers, which makes sense. But IBM is carrying the
predominantly) from specifications. Thisincluded such				analogy perhapstoo far by stressing that "on
products as System-80 (Phoenix Systems),				demand" is the manner by which companies shouldrun
GENASYS (Generation				in the future. Basically, the theory suggests we
Sciences), and JASPOL (J-Sys of Japan), to mention				abandon capacity planningand rely on outside vendors
but a few.				to save the day. Further, it implies
MBA also introduced a generator of its own in 1979 -				computerssupersede the business systems they are
a Systems generatorinitially named ADF (Automated				suppose to serve. Instead ofunderstanding the
Design Facility) which could automaticallydesign whole				systems which runs a business, just throw as much
systems, complete with an integrated data base.				computerresources as you need to solve a problem.
Based oninformation requirements submitted by a				This is like putting the cart beforethe horse.
Systems Analyst, ADF interactedwith the				The "on-demand" movement has evolved into
"PRIDE"-LOGIK Data Dictionary to design new				"Service Oriented Architectures"
systems and, whereappropriate modify existing				(SOA) where vendors are introducing "on-demand"
systems. Because of its link to LOGIK,				applications that will takecare of such tasks as payroll,
ADFemphasized the need to share and reuse				marketing, etc. through the Internet. Again, it
information resources. Not onlywas it useful as a				allsounds nice, but as far as I can see, this is
design tool but it was a convenient tool for				essentially no different than servicebureaus like ADP
documentingexisting systems. The only drawback to				who for years provided such processing facilities.
ADF was that the mindset of theindustry was				Now,companies are being asked to swap out their
shifting from systems to software. Consequently,				internal programs for third partyproducts. I fail to see
programgenerators captured the imagination of the				how this is different than buying any other
industry as opposed to ADF.				packagedsolution, other than an outsider will be taking
The increase in computer horsepower, coupled with				care of your software.
new programmingtools and techniques, caused a shift				The need to build software faster has reached a
in perspective in MIS organizations. Now,such				feverish pitch. So much so,full-bodied development
departments became dominated by programmers,				methodologies have been abandoned in favor ofwhat
not systems people. Itwas here that the job titles				is called "Agile" or "Extreme Programming" which are
"Systems Analyst" and "Programmer" were marriedto				basicallyquick and dirty methods for writing software
form a new title of "Programmer/Analyst" with the				using power programmingtools. To their credit, those
emphasis being onprogramming and not on front-end				touting such approaches recognize this is limitedto
systems design. Many managers falselybelieved that				software (not total systems) and is not a substitute
developers were not being productive unless they				for a comprehensivemethodology. Agile/Extreme
wereprogramming. Instead of "Ready, Aim, Fire," the				Programming is gaining considerable attentionin the
trend became "Fire, Aim,				press.
Ready."				Next, we come to "Enterprise Architecture" which is
Data Management organizations floundered during this				derived from a paperwritten by IBM's John A.
period with theexception of Data Base Administrators				Zachman who observed that it was possible to
(DBA's) who were considered thehandmaidens of the				applyarchitectural principles to the development of
DBMS.				systems. This is closelyrelated to consultants who
The proliferation of software during this decade was				extoll the virtues of capturing "business rules"which is
so great that itgave rise to the packaged software				essentially a refinement of the Entity Relationship
industry. This went far beyondcomputer utilities and				(ER) Diagrammingtechniques popularized a decade
programming tools. It included whole systemsfor				earlier using CASE tools.
banking, insurance and manufacturing. As a result,				As in the 1990's, concepts such as "Enterprise
companies wereinclined to purchase and install these				Architecture" and "business rules"is indicative of the
systems as opposed to reinventingthe wheel. Among				industry trying to reinvent systems theory.
their drawbacks though was that they normally				CONCLUSIONS
requiredtailoring to satisfy the customer's needs				Like computer hardware, the trend over the last fifty
which represented modification to theprogram source				years in systemsdevelopment is to think smaller.
code. Further, the customer's data requirements had				Developers operate in a mad frenzy towrite
tobe considered to assure there were no conflicts in				programs within a 90 day time frame. Interestingly,
how the customerused and assigned data. After the				they all know thattheir corporate systems are large,
package had been installed, thecustomer was faced				yet they are content to attack them oneprogram at
with the ongoing problem of modifying and				a time. Further, there seems to be little concern that
enhancingthe system to suit their ever-changing				their workbe compatible with others and that
needs.				systems integration is someone else'sproblem. Often
1980's - THE TOOL-ORIENTED APPROACH				you hear the excuse, "We don't have time to do
As big iron grew during the 1960's and 1970's,				thingsright." Translation: "We have plenty of time to
computer manufacturersidentified the need for				do things wrong." Any shortcutto get through a
smaller computers to be used by small to				project is rationalized and any new tool promising
medium-sizedbusinesses. In the 1970's, people were				improvedproductivity is purchased. When companies
skeptical of their usefulness butby the 1980's their				attempt to tackle large systems
power and sophistication caused the "mini"				(which is becoming rare) it is usually met with
computerto gain in popularity as either a general				disaster. Consequently, companiesare less confident in
purpose business machine or dedicatedto a specific				their abilities and shy away from large system
system. Among the most popular of the "mini"				developmentprojects.
computers were:				Corporate management is naive in terms of
- IBM's System 36/38 series (which led to the AS				comprehending the value ofinformation and have not
400)				learned how to use it for competitive advantage
- DEC PDP Series (which gave way to the DEC VAX				(unlike their foreign competitors). Further, they are
VMS)				oblivious to the problems insystems development.
- Hewlett-Packard's HP-3000 series with MPE				They believe their systems are being developed with
- Data General Eclipse series with AOS				a highdegree of craftsmanship, that they are
- PRIME				integrated, and that they are easy to maintainand
The competition was fierce in the "mini" market				update. Executives are shocked when they discover
which resulted inconsiderable product improvements				this is not the case.
and better value to the customer.				The problems with systems today are no different
Instrumental to the success of the mini was the				than fifty years ago:
adoption of UNIX asdeveloped by Bell Labs, a				- End-user information requirements are not satisfied.
powerful multi-user, multitasking operating systemthat				- Systems lack documentation, making maintenance
eventually was adopted by most, if not all, mini				and upgrades difficult.
manufacturers.				- Systems lack integration.
But the major development in computer hardware				- Data redundancy plaques corporate data bases.
was not the mainframe,nor the mini; it was the				- Projects are rarely delivered on time and within
"micro" computer which was first popularized by				budget.
Apple in the late 1970's. IBM countered with the its				- Quality suffers.
Personal Computer (PC)in the early 1980's. At first,				- Development personnel are constantly fighting fires.
the micro was considered nothing more thana				- The backlog of improvements never seems to
curiosity but it quickly gained in popularity due to its				diminish, but rather increases.
inexpensive cost,and a variety of "apps" for word				Although the computer provides mechanical leverage
processing, spreadsheets, graphics, anddesktop				for implementingsystems, it has also fostered a
publishing. This caught on like wildfire as micros spread				tool-oriented approach to systems development.
throughcorporate desktops like the plague.				Insteadof standing back and looking at our systems
By the mid-1980's the "micro" (most notably the PC)				from an engineering/manufacturingperspective, it is
had gained in powerand sophistication. So much so,				seemingly easier and less painful to purchase a tool to
that a series of graphical based productswere used				solve aproblem. This is like taking a pill when surgery
for software development in support of the				is really required. What isneeded is less tools and
Structured Programmingmovement of the 1970's.				more management. If we built bridges the same
Such tools were dubbed "CASE" (Computer Aided				waywe build systems in this country, this would be a
Software Engineering) which allowed developers to				nation run by ferryboats.
draw their favorite softwarediagramming technique				The impact of the computer was so great on the
without pencil and paper. Early CASE pioneersincluded				systems industry that itelevated the stature of
Index Technology, Knowledgeware, Visible Systems,				programmers and forced systems people to
Texas				nearextinction. Fortunately, the industry has
Instruments, and Nastec, as well as many others.				discovered that there is moreto systems than just
CASE tools took the industryby storm with just				programming and, as a result, is in the process
about every MIS organization purchasing a copy				ofrediscovering basic systems theory. Some of the
either forexperimental use or for full application				ideas being put forth are trulyimaginative, others are
development. As popular as the toolswere initially,				nothing more than extensions of programming
there is little evidence they produced any major				theory,and others are just plain humbug. In other
systems but,instead, helped in the design of a single				words, the systems world is still goingthrough
program.				growing pains much like an adolescent who questions
Recognizing the potential of the various CASE tools,				things and learnsto experiment.
IBM in the late				I have been very fortunate to see a lot of this
1980's devised an integrated development				history first hand. I haveobserved changes not just in
environment that included IBM'sproducts as well as				terms of systems and computers, but alsohow the
third parties, and entitled it "AD/Cycle." However,				trade press has evolved and the profession in
IBMquickly ran into problems with the third party				general. Ithas been an interesting ride.
vendors in terms of agreeingon technical standards

Xinit Systems

A Short History of Systems Development