|
The North American railway system is at risk. The biggest problem is the interconnections problem. How can the trains move if each firm's computers are not compliant internally and also externally? I have been asking this question for over a year. No one answers it in print because the answer is obvious.
The breakdown in Union Pacific's service since the summer of 1997 is a fortetaste of things to come. This article discusses the problem's root cause: the incompatibility of Southern Pacific's computer system with Union Pacific's. When Union Pacific absorbed Southern Pacific for $3.9 billion, it swallowed a poison pill.
Union Pacific must update 32,000 internal programs.
Like so many huge firms, the company found out about its problem late: in 1995.
Optimists say this problem can be fixed. I ask: Where is any company with 20 million lines of code that has fixed them? Where is one industry that is compliant?
This appeared in a series of articles on y2k sucesss stories published by the Information Technology Association of America (ITAA) (Jan. 3).
I say: if this is a success, head for the hills. And don't buy a return ticket.
We are asked to believe in fairy tales: a universal international repair of a problem that not a single Fortune 500 company has solved.
* * * * * *
As the largest railroad in North America, Union Pacific depends on dates to make the trains both literally and figuratively run on time. . . .
What's the Y2K connection? Southern Pacific outsourced its IS operations. Thus, while these information systems will not make the transition, the consolidated company must still assimilate vast quantities of Southern Pacific data. In some cases, the new data must be changed to fit in the Union Pacific framework; in other instances, Union Pacific programs must be adapted to incorporate new types of information. Melding the information resources of the two companies has required both people and planning, even as Y2K efforts must speed ahead down a parallel track. . . .
Keeping dates straight is just one of the formidable end user obstacles this program manager must address. The third of six targets on Brechbill's screen, Union Pacific has a "customer computing" environment composed of 356,000 Focus programs, mostly used for decision support. Brechbill says that less than a third appear to be active. But even so this means roughly 98,000 homegrown programs, developed with no company standards whatsoever, tucked away in various libraries, some with what Brechbill termed mission critical functionality. . . .
Here Brechbill is using a "divide and conquer" strategy. His Year 2000 team will take responsibility for converting 32,000 of the end user programs; the balance, he said, will be left up to individual departments for correction. Whether or not this happens will be helped along by monthly Y2K conversion status reports. Brechbill says central IT support may eventually extend to staff and tools, but he frankly admits that his first priority is converting the 32,000 programs on his plate. . . .
Electronic data interchange is another functional area that has been pushed to the edge of the plate. Union Pacific has not tackled the issue yet, but Brechbill hopes to have a strategy in place by the end of the month. Embedded and vendor supplied software also constitute a category on the fringe of current activity. Like many organizations, Union Pacific is in the process of contacting software vendors to ascertain Y2K compliance information. Because much of what the company does takes place in "real time" mode, the Y2K conversion must also extend to the electro-mechanical, signal processing and communications devices used up and down the line. To make this complex assessment, Brechbill is not only polling vendors, but also Union Pacific Vice Presidents and operating departments.
Union Pacific launched its first Y2K pilot in 1995. Compared to what he has heard from others at a recent Y2K conference, Brechbill said he thinks his company is ahead of the game. The program manager hopes to stay ahead by cutting down the time it takes to test systems and put them back into production. He has a pilot in the works, details not available, that may make dramatic improvements in this critical area.
|