THE RISE OF GIGABIT GOOF-UPS

greenspun.com : LUSENET : Grassroots Information Coordination Center (GICC) : One Thread

THE RISE OF GIGABIT GOOF-UPS

DATA ERRORS ARE INCREASING AS BUSINESS FUNCTIONS BECOME MORE AUTOMATED. MANY EXPERTS SAY IT'S TIME TO DEVELOP STANDARDS FOR INFORMATION INTEGRITY.

By Michele Fitzpatrick Tribune Staff Writer July 10, 2000 Computer data errors come in all shapes and sizes. One causes a minor mishap, like a 1-cent difference in the price of a gallon of milk because a scanner misread the bar code. Another triggers error avalanche, like the glitch by an East Coast credit reporting company whose system mislabeled a file of names as "deadbeat," thus eroding the credit rating of 1,400 upstanding citizens of Norwich, Vt.

Most are merely irritating, but some are catastrophic enough to deep-six a company's valuation, erode its profits and bury its credibility. Companies weigh bottom-line benefit against prevention cost, but those who have been bitten by a big mistake now wish they had spent enough to prevent these beauties:

- About 38,000 employees and retirees of a major lender found their nest egg balances adjusted downward by as much as 16 percent when a $70 million computer accounting system error was discovered.

- A large health insurance carrier spent $200 million consolidating its computer systems, which promptly glitched and mailed out more than $60 million in duplicate checks and overpayments.

- A bank unintentionally created 800 multimillionaires when a minor system bug resulted in more than $763 billion being mistakenly added to those accounts.

- A bar code foul-up on the return address of Cook County tax bills that went out at the end of September 1999 resulted in about 900 bills being sent to Puerto Rico.

- In 1998, 60,000 people reported telephone-bill errors to the FCC that ranged from small dollar amounts to hundreds of dollars.

As just about every process in modern life becomes automated, the number of data-driven errors is rising, and the financial and emotional costs have yet to be measured.

"We will spend years documenting some spectacular mistakes that would never have happened before the rise in technology," said Roy Schulte, an analyst for Gartner Group consultants. "I am sure we are going to see many more interesting goof-ups."

Another who says we've only seen the tip of the iceberg is Madhavan K. Nayar, who heads Naperville-based Unitech Systems, a 15-year-old software solutions company. He expects the number of major data errors to snowball and warns that corrupted data will pollute the Internet environment much as Industrial Age factories destroyed rivers and wildlife.

He argues that a universal standard is needed to ensure that any compilation of data is complete and current and that its origin and accuracy can be traced. "In the past we were able to ensure a level of accuracy, consistency and reliability of information because we either had the time to inspect information or the problems were sufficiently isolated to prevent wide-scale contamination," he said.

"Although organizations such as CommerceNet are creating standards for [business-to-business] commerce ... no defined standard exists for end-to-end information integrity throughout the transaction cycle."

At the very least, data errors may gum up the e-commerce machinery. Before the automation explosion ignited by e-commerce, human error was the key culprit when data went wrong; now software and network processes themselves are making more of their own messes.

In short: Fewer human operators plus more automation equals more computer-generated bugs.

If there is any good news in this, according to Nayar, it's that "a software program will make the same mistake 100 times. One hundred humans, on the other hand, are likely to make 100 different mistakes."

But that one software mistake can wreak havoc. For example, say an application is designed to perform a straightforward task, such as reading 1,000 names and addresses and entering them into a database. If the software should receive an empty file or a file with more than 1,000 names, there is no telling what it may do.

Unlike a human, it can't correct itself. Although safeguards to correct predictable variances are designed into most programs, Nayar suggests they are insufficient.

"People have convinced themselves that they can ensure the integrity of their systems from within," he said, "but external variables are very difficult to predict, and we have no universal certification that confirms which software is adequate to which task we may ask it to do."

Managing and protecting data already has spawned a thriving industry that comprises software solutions and software testing companies; quality control organizations; data-quality integrity scanners; as well as risk management and content management divisions in consulting companies.

Nayar contends that with all this, too little attention has been paid to ensuring that the ever-rising tide of information interchanged and transferred on-line is accurate, consistent and reliable.

"Each of us are using our own smell tests to determine information integrity," he said. "Contrast this to buying milk, where the expiration date gives you reasonable assurance you don't have to take it to the lab to determine bacteria content. There is a whole system behind that date that tells us it's drinkable. We have no expiration dates for information on the Internet."

Nayar said lack of information integrity costs businesses plenty.

"My unscientific estimate is that every business is spending between 1 and 5 percent of their revenue in relation to this issue," he said.

Dean Williams, senior manager of enterprise risk services at Deloitte & Touche, concurs that information integrity is a growing and expensive concern. But he said a standard is feasible only if it's driven by business goals and not by consensus among segmented members of different industries. "Standards evolve two ways--imposed from the top-down by government or they evolve when people voluntarily get together in a room and say, `OK, let's hammer one out,'" he said.

Schulte also expressed doubts about defining a universal information integrity standard. "Within each there are literally thousands of different application areas that treat data different ways, so you wouldn't want a single standard for data integrity," he said. "What you need is something much more specific to each industry."

Nayar agreed that industry-specific standards are important but argues there is room for common agreement among computing professionals.

"Think of water," he said. "On a fundamental level there is the basic use of water for all. On the industry level, you would adopt standards that apply only to water use within a hospital. On the enterprise level, further specifics would apply only to the water processing plant that supplies that hospital; and on the application level, you have the specifics that gets the water to the laundry room, drinking fountains, and the patient's bedside."

He applied the analogy to information systems. "We all have been drinking from wells in our own back yards," he said. "Now we are tapping into the municipal water system and we're pumping our personal well water into it. All this is flowing. All of us are drinking it--and we haven't even agreed yet to call it `water.'

http://www.chicagotribune.com/business/printedition/article/0,2669,SAV-0007100019,FF.html

-- Martin Thompson (mthom1927@aol.com), July 11, 2000

Answers

"He expects the number of major data errors to snowball and warns that corrupted data will pollute the internet environment much as Industrial Age factories destroyed rivers and wildlife."

Hmmm!

Intriguing statement. It ties in almost perfectly with my "Termites Theory" of y2k.

-- JackW (jpayne@webtv.net), July 11, 2000.


Jack has added: "Intriguing statement. It ties in almost perfectly with my "Termites Theory" of y2k. "

That in of itself is intriguing. I'd like to hear more about that. (Forgive me if you've recently posted this - I've just returned from out-of-town). Thanks!

-- Jan Nickerson (JaNickrson@aol.com), July 12, 2000.


The data errors will only get worse, and this is why.

Experienced IT workers are being replaced by indentured servents so that people like Bill Gates can make more money!

The Following is written by an Indian former H-1B. He has a number of comments on the Immigrants Support Network (ISN, www.isn.org), a militant group of H-1Bs which is lobbying Congress to alleviate their long (5+ years for Indians and Chinese H-1Bs) waits for green cards, during which time they are exploitable "indentured servants."

One of the central points made by the author here is fraudulent claims by H-1Bs that they have work experience in a given programming language or platform. As noted below, the State Dept. found massive fraud, and many programmers tell me that they see H-1Bs working next to them using language X (whatever), but reading a book on X starting from scratch,learning while they go. The industry should be doing this with American workers: Hire a U.S. citizen or permanent resident who is a good generic programmer, and then let him or her learn language X on the job.

Microsoft just came out with a new programming language, "C#". "Watch what will happen. The H-1Bs will be the first ones to show up with "experience" in C# because Americans will be frozen out!

The Urban League came out last week and stated that there is no evidence that the H1B program should be expanded. Why is Congress not listening. Why is Congress selling out the voters and the consumer of IT services?

********************************* ISN, www.isn.org

A word on the H-1B lobby groups

I am writing to express concern about lobby groups that have been actively lobbying congress to speed up Green Card processing for H-1B workers

One such group is the Immigrants Support Network (ISN). Firstly, ISN is NOT an immigrant group. It's members are "guest workers" on H-1B visas who came to work in the U.S on a temporary basis.

The overwhelming majority H-1B workers are Indian nationals who were brought to the U.S. by Indian bodyshops.

I know many H-1B workers, that have NOT been engaged in continuous employment whilst in the U.S and are yet demanding "quick" Green Cards. Many of them are spending 2 - 3 months of the year "on the bench" (without work). This is illegal. U.S. immigration and labor laws clearly state that a H-1B worker who is without work for an extended period of time has to leave the country. This is not happening.

Needless to say, this information is not being divulged to the media or members of congress.

Many H-1B workers were hired from Chennai (formerly Madras) and Mumbai (Bombay). A state department audit of H-1B visas issued at the U.S. consulate in Chennai in 1999, revealed that a whooping 45% of all applications were fraudulent.

I am friends with many H-1B workers that were hired from Chennai. They OPENLY acknowledge that they do not have the necessary work experience to qualify for H-1B visas. It's a well known fact that many Indian H-1B workers are brought to the U.S. on fabricated re'sume's. In extreme cases Indian nationals lacking an IT or engineering degree pay bribes to Indian bodyshops to get them a H-1B. I know of one individual who obtained his H-1B in this manner. He has no experience in the I.T. industry and now works in a pizza shop in the New York area!

The latest trend is for people in India to take a few classes in the Java programming language and the pass themselves off as e-commerce experts. Most of them succeed in getting H-1Bs' because the U.S. consulate in Chennai does not have the manpower to investigate their work experience.

The Green Card process, which many H-1Bs' complain about is full of loopholes that is to their advantage. Before a H-1B worker can apply for a Green Card, he has to prove that his job cannot be performed by a U.S. citizen. Using clever immigration attorneys, H-1B workers routinely exaggerate job duties so as to disqualify U.S. citizens that apply for their jobs. These job ads. often appear in fine print in obscure corners of Computerworld magazine and are rarely seen by U.S. jobseekers who mainly look for jobs on the internet. The job ads. frequently demand that the applicant have a Masters degree. Needless to say, there is never much response to these jobs ads. from U.S. citizens. The attorney can then argue that the job the H-1B worker is doing cannot be filled in the United States. Once this process, know as Labor Certification is complete, the H-1B can then apply for a Green Card.

I understand that the Dept of Labor is now planning to scrap the Labor Certification process and replace it with a system of attestations from employers (of H-1Bs'). This new system poses a grave threat to U.S. I.T. workers as it will only allow more abuse - we all know how honest the bodyshops are!

I urge the the journalists on your mailing list to challenge the groups that are lobbying for "quick Green Cards". Ask them how is it that a poor country like India that is largely uncomputerized can produce so many I.T. professionals claiming to be experts in e-commerce, web design and computer networks.

Ask them why their members spend several months of the year "on the bench" without work. If there really was a shortage of IT professionals, why can't their bodyshops find them projects.

Ask to see their degree certificate's and affidavits supporting their work experience. Can their work experience be authenticated? In India it is possible to openly buy false documents in the bazaars and marketplaces.

Some of these H-1B groups has embarked on a program to mislead the press and the U.S. congress. They claim they have 15,000 members. I strongly doubt they do. They probably have 15,000 people on their mailing list. They imply that without them the technological boom in the U.S. will end. Nothing could be further from the truth.

Please note that bill to speed up Green Card processing for H-1B workers and to increase the number of H-1B visas will be detrimental to U.S. I.T. workers.

1999 was one of the worst years for many I.T. workers - H-1Bs' as well as U.S. residents - with many of us spending 2 - 4 months on the bench. I myself was displaced by H-1B workers on 2 occasions. All this in the year of the millennium bug when there was supposed to be a big shortage of I.T. professionals. I dread to think what would happen in an economic downturn.



-- K (infosurf@yahoo.com), July 12, 2000.


I had never given much thought to the WHY of so many IT people-imports from India over the past few years. But, I sure found it strange.

Thanks, K, for an excellent analysis. It makes for nothing but pure common sense.

-- Uncle Fred (dogboy45@bigfoot.com), July 12, 2000.


Moderation questions? read the FAQ