« Media easy on one of its own | Main | Apple PIM apps are weak »

February 24, 2003

Old programming language has defenders

COBOL is one of the oldest still functioning programming language running lots of mainframes today. But even though it was invented in 1959, even though it was the hopelessly complicated COBOL code that needed to be changed in the much-hyped Y2K mess of 1999, it still has its defenders.
I've been targeted by a raft of them because of a column I wrote today for the Detroit Free Press about the antiquated computer system used by the Michigan Seceretary of State's office. The system needs a $30 million update and, I noted, "was written in the long-since-outdated COBOL programming language."
Old programmers can get pretty steamed:

"Please stop letting the latest web site, latest gadget, or latest nifty shareware program cloud your judgement. I haven't worked with COBOL in at least 10 years, but I can assure you I see business problems every day that could be solved easily with COBOL but aren't because of unfounded bias and prejudice, and the irrational desire to use the "next best thing" instead of "the thing that works and has always been working" -John Turner, Ann Arbor
Another reader sent me a link to a story that notes 85 percent of the applications managing the world's business data are written in COBOL. Someone who works for a financial institution e-mailed:
"Just because the state of Michigan's systems are written in COBOL is no reason they can't be real-time online type applications. It's possible that the hardware the state has is not powerful enough to handle that type of work, or the folks working there don't have enough experience writing real-time applications (and very likely both). Don't blame the tool because it's being improperly used."- From Jack Beckman
There you go. But I sure remember covering the Y2K changeover mess when all we heard from corporate IT execs was what a mess COBOL was and how difficult it was to go back and fix the code to make sure computers could recognize the year 2000. The billions spent on that job drained IT spending so bad it started the tech fallout we're still seeing.


in Computers | TrackBack (0)

Comments

Watch whatcha say about us old COBOL programmers and our language.

Truth is, if the stuff that needed Y2K remediation had been written in most any other language, that language would have caught the heat instead of COBOL. The name of the game when many of those programs were written was to save space on cards, tape, and disks (and the time it took to read those media in those days). The year would have been coded in two digits (unless someone were smart enough to code it in one character, using a start year and both letters and digits, binary, etc. in that character).

Some may argue that the verbosity of some COBOL makes it hard to follow. I submit that if it is properly used as it was designed, it is very self documenting and easy to follow. Look at how programs written in C (and, often, in any other language) is usually written. Without strict enforcement of standards, it's almost as bad as Greek (well, try APL).

COBOL is now object-oriented, has easy ways to manipulate screens, can easily access SQL and other data bases, ... Very complete implementations (orders of magnitude enhanced over the 1960/70's COBOL in which many of those systems were written) are available for not only "big iron" but also for UNIX, LINUX, etc., Windows, ...

Now, a real problem is to update both the user interface edge of this software to be interactive, secure, etc., and the database foundations so they can be updated and accessed in real time (so two people don't buy the same vanity plate at the same time).

A bigger problem is to change the business practices of the enterprise (in the case of your column, the State of Michigan) to take full advantage of the "new style" of access possible because of new hardware and software technologies.

It is and would be as easy to develop this software using COBOL as it would be using C or JAVA. Either can access an RDBMS, such as Oracle.

In my opinion, this is not being done today for some simple reasons. The primary reason is that the move to UNIX/LINUX... comes with C programmers inherently imbedded, setting up an uninformed decision that software should be switched to this language.

A second reason is the uninformed misconception that COBOL is not as transportable between systems as is C. Whomever believes this needs to look at the range of platforms supported by COBOL products from people such as MicroFocus.
A third reason is that colleges first taught Fortran to math and science students, then a few thought PL/I would take over from COBOL so many moved on to that language. Soon they were caught up in languages which were great for teaching, such as PASCAL, then came the solution to the world's problems - ADA. Those soon to be the decision makers never became conversant with the capabilities of COBOL.

Meanwhile, Unix was taking over mid-range systems, and C which, with BASIC, were the only non-assemble languages available early. The the mid-range started growing, pushing into the "big iron" world, and taking C along with it (actually, some implementations of BASIC, such as PowerBASIC, are better choices for "PC-based" development than C - in my opinion).

So, don't fall for the Maginot Line fallacy. COBOL may not be the popular choice, but serious and diligent research prior to making a judgement might well show COBOL in a much more favorable light.

Posted by: Harper G. Grimm, Jr. at February 24, 2003 03:15 PM

Simple enough - you really do not know what you are talking about. Read and learn the language and the technology. When you have the knowledge then state your opinion. It's apparent you have never written or looked at COBOL code.

Posted by: Dave Antonelli at February 24, 2003 04:02 PM

Heh. Don't anger us old mainframers.

Harper is absolutely correct about the storage problem. Using a 4-character year field, instead of a 2-character year field, would have meant probably literally millions of dollars spent on storage in those years. A severe no-no, and practically a fireable offense in the big code shops like EDS, IGS, CSC, etc. And we would have been fired by those same corporate IT execs who were busy blaming everyone but themselves for the Y2K "crisis."

Terry's problem is far more likely to be a backwards-looking staff dominated by one or two guys who is resistant to change, and really crappy budgeteers that think of modern networked computer systems as frippery. I've seen it a million times.

And quit taking those ass-covering IT execs at their word...:-)

Posted by: Scott Chaffin at February 25, 2003 08:25 PM

Amen to the 3 previous posts!

Posted by: Ingo at February 27, 2003 08:38 PM

ok dokie

Posted by: ip address at May 4, 2003 03:44 AM

On line Real time development on a mainframe is a "snap". In addition, distributing that same capability to both mid-range platforms as well as desk top P C s can be easily accomplished as well. Please check some of the projects listed below :

Jeff von Horn
3233 Gayhart Street
Racine, WI 53406
262 - 554 - 6431
[email protected]
BACKGROUND SUMMARY

Computer System Developer with extensive expertise in integrating a variety of vendors' software and hardware offerings into an On Line Real Time IBM Mainframe environment responsible for processing $7.5 million of daily North American sales volume. Average individual transaction response time requirements for all systems delivered were in the under 12 second range, while supporting users operating in several client/server environments.

EMPLOYMENT HISTORY

1998 to Present
A major global producer of both Agricultural and Construction Equipment with annual sales of $9 billion
Application Analyst II e-business
Responsible for the design, programming, and implementation, of functional computerized interfaces from various vendors' systems into On Line Real Time IBM Systems operated by my employer. Also required to identify, and approve the use of various technically based software products needed to transmit critical business data between unlike systems, operating on completely different computer platforms.
* Perform all of the design and mainframe applications programming necessary to allow over 2,000 North American retail outlets to communicate directly in an On Line Real Time manner over the Internet from their PC's JAVA applet into my employer's IMS based systems. During any given business day, more than 7,000 transactions originate from these PC's, and an additional 4,000 spare parts orders are received, and processed for same day shipment to end customers.
* Interact with more that a dozen different computerized business system vendors, guiding their efforts during a conversion of systems to support an expanded file key format required for the submission of orders, inquiries, and return requests in an On Line Real Time setting.
* Provide on call support for a subscription driven internet based daily recommended order calculation system, optimizing inventory levels, and improving dealer profitability.

1993 to 1998

Senior Applications Developer
Responsible for developing systems that support users operating in a client/server environment. Strong ability to create interfaces that expedite global communications, improve business processes and increase profitability. In addition, was often utilized as the lead IT expert assigned to numerous project teams.
* Programmed, designed, and implemented a complete data interface between an On Line Real Time Global Inventory Control System, and a group of UNIX based process-control computers. This particular interface was needed to manage the daily receipt and shipment of product housed in a totally automated warehouse storage complex, holding $30,000,000 of company inventory.
*
* Developed specialized systems to produce a complete geographic analysis of global spare parts demand patterns. The time period covered was all of 1996 and 1997, and involved performing a detailed analysis of over 30,000,000 records of confidential daily wholesale demand data. A specialized linear programming product called SAILS was also utilized. At the conclusion of this effort, it was found that these data extracts were accurate to within one tenth of one percent of the actual reported shipments processed. Actuals had been derived from manually kept warehousing records. Senior Executive Management launched a complete $86,000,000 redesign of their global facilities network for spare parts distribution as a direct result of this study.
* Designed and developed all hardware and software interfaces required to support the production of bar coded shipping documents by an On Line Real Time IMS based system. Next, a warehouse control system was enhanced to accept bar code scanners and platform scales as input devices directly into IMS. The modified equipment was then used to transmit individual package shipping data directly from this classical shop floor environment as it was recorded. This project eliminated the costly and error prone keyboard based reentry of hand written data by office personnel, and gained a full one to two days in the customer invoicing cycle. All modified workstations completely paid for themselves within three months.
* Worked with many separate companies who were in the business of selling generalized business system software. Was able to develop a common data interchange format, and telecommunications protocol with them. Top management was interested in allowing all 2,500 of their North American based retail dealerships to have a way to electronically communicate their customers' material needs as determined by their business system's point of sale subsystems directly into an On Line Real Time Mainframe based spare parts shipping system operated by my employer. This effort resulted in the electronic receipt of 95% of daily sales volume from all retail outlets. Participating locations were able to transmit orders, returns, and many inquiries, with a minimum amount of effort, straight into corporate headquarters.
* Conceptualized, developed, implemented, and marketed, a complete ASP parts inventory system to independent businessmen who owned equipment dealerships all across North America. Managed a field force of system installers. Product features included point of sale, ordering, physical inventory, bill of materials processing, customer tracking, return goods, performance reporting, EOQ stocking level calculations, site to site messaging, and a nationwide product locator. Over 200 locations rented access to this centralized service for $1,600 per month.
EDUCATION
BS, Mathematics, Purdue University, West Lafayette, IN
BSIM, Industrial Management, Krannert School of Business, Purdue University, West Lafayette, IN
SPECIAL SKILLS
Windows Word Excel Access E-Mail Internet TSO ISPF CLISTS MVS FTP JCL DYNACOM PANVALET COBOL BAL FORTRAN JAVA (light) Visual Basic GPSS IMS DB MPP BMP MFS BTS DC Tuning DB2 MQ Series (light) IMS DC UNIX (light) AIAG OLE PL/I CCN AS/400 OCL (light) AMAX ACF/2 Client/Server Insight SAILS Integer Programming MS Project (light) CICS (light)
COURSEWORK Go to http://webdev.gtc.edu/pearson/vonhornj/resume.htm
HTML XHTML JAVA Script CSS DHTML XML VB Script ASP Perl Python SAP VB.NET

Posted by: Jeff von Horn at July 19, 2003 05:10 AM

Quick Links


    mike!

Google

Web This site

Site Credits