Testing time for ASB Bank

By Chris Barton

ASB Bank's Garry Fissenden confirms what the Y2K experts have been warning for some time.

"There's as much work in testing as there is in remediation - 50 per cent of our time has been in testing. We get very worried when we hear about companies that are just staring to test now."

The general manager technology operations is talking about the scope of ASB Bank's century date change problem: 2.2 million lines of code, 2000 programs, 16,000 date incidents, 167 automatic teller machines (ATMs) replaced and a forecast cost of $5.5 million to date.

The bank began in earnest on its Y2K programme early in 1997. All the six-digit dates in the Cobol core banking programs which run on the bank's Unisys Clearpath mainframes were changed to eight digits and put back into production by October last year. The system was then put through its paces in ASB Bank's "age testing environment" which runs programs through six key dates: 9/9/1999, 1/1/2000, 29/2/1999, (for the leap year) 31/3/2000, 30/6/2000 (financial year ends) and 31/12/2000. So far it's come through with only minor faults.

ASB Bank is firmly of the view that Y2K is "not an IT problem, but a business management issue." Its Year 2000 Programme manager Alan Barber brings 30 years experience across a range of branch, commercial banking and head office roles to the job rather than IT knowledge. He oversees progress - reporting directly to chief executive Ralph Norris - across seven programme streams comprising software services, technical infrastructure, property, treasury, automated banking and cards, testing services and conventional banking.

Mr Barber stesses that testing is the critical component of Y2K work.

"The majority of it is done, but it's an ongoing process. We will test and retest anything new coming into the bank for the rest of the year."

At present he is satisfied with the "first cut" of the ATM and Eftpos terminal tests using the bank's test environment, aptly named Tardis, that show the year 2000 does indeed compute. Bank to bank testing - where a transaction from a customer in one bank correctly finds its way to another customer's account at another bank - has also just been signed off. Next on the agenda is more testing of the Eftpos network and Visa and Mastercard transactions.

Then there are the contingency plans. As well as arranging for extra staff to be available for testing during the final days before banks open on January 5, there's a myriad of practicalities to organise to make sure the plans work. Such as: if there is a telecommunications failure, how to get the day's tapes to EDS to process settlements; the early booking of security firms to move extra cash to branches should there be a need; ensuring the ATMs can handle both the old paper notes and new polymer notes that will be in circulation at the time; updating cashflow cards that expire 12/99 progressively rather than all at once in December; and making sure retailers have plenty of manual Eftpos vouchers on hand should there be a network failure.

ASB estimates it will process about two million Eftpos transactions in the period from New Year's Eve until the branches reopen in 2000.

Communication strategies - "telling the good news about our programme" - are also swinging into play. Next week the bank will provide detailed Y2K information on its Web site. There will also be information coming out on bank statements, ATM screens, brochures and letters.

Chief executive Ralph Norris says while the project clearly has a cost and has taken resources thereby delaying some things the bank would have done - it's not all bad. He cites the newly created test environment which will remain as a key component in future software development, vastly improved documentation of inventory and systems, and "the opportunity to analyse business process" as unexpected spin-off benefits.

He believes the bank is doing everything possible to ensure its customers won't be affected. But he is aware there are a number of factors, such as power and water supply that are outside his control.

"In the final analysis testing only proves the presence of bugs, not their absence. But we've done everything we can to get on top of any problem quickly so that if problems do occur they'll be of a minor nature."