The basic parameters of an IT project - indeed any project - are threefold; it must be of good quality, done within an acceptable time and keep within the financial budget.
Balancing these three objectives has long been the major problem of development. Slippage in time has often blown out budget, and the anxiety to keep within budget and time has resulted in a cheap, quick job, which has not met quality expectations.
The stories of failed or unsatisfactory IT projects are legion.
A recent consultant's report indicated that government projects fall short of targets less often than projects in private industry; but information about the government projects becomes - as it should be - public property, more often than commercial IT endeavours do, and hearing or reading about another high-profile government "debacle" in the computer area does not inspire confidence in the IT plans of the rest of business.
I must say, as journalists we do try to bring you news of the private industry failures as well - and the successes on both sides.
This week we are faced with a couple of other emerging government projects that are subject to severe time pressure.
The Department of Work and Income (DWI) has just issued RFIs for a slew of IT projects, including a second crack at the already controversial student loans and allowances system.
It's late July already, with responses to this first stage in preparation, and the system - or at least its essential aspects - is slated to be complete by September 30 this year- about the time the students begin looking to register for the new academic year.
Such a compressed timescale raises the danger of DWI resorting to some "quick and dirty" solution, where a vendor promises to meet deadlines - and may or may not actually do that - but will compromise on quality.
One observer of the DWI session was frank that he saw "another disaster in the making", especially as the DWI people did not appear to have all the answers to prospective bidders' questions on its (the Department's) requirements.
The question must surely arise, why start from ground zero again? Taking last year's system - as shaky as it was - and trying to improve its bad spots may have come up with a better solution. There is a hope that some of the prospective bidders might suggest such an alternative.
Few computer systems these days exist in isolation. DWI's student loans and allowances computer system has to implement business processes which will be in immense flux over the next few months, as a result of a report just issued which is long on criticism of the processes involved in last year's administration of loans and allowances.
A somewhat similar picture is evident with high-profile Government initiative, the so-called "People's Bank", due to emerge from NZ Post by the end of this year or early next. Here an RFI was issued, earlier this month with a deadline for replies July 31.
Now an RFI is, by nature a very preliminary document.
But New Zealand Post - and Jim Anderton's lobby within Government, which is powering the initiative - doesn't even appear to have a firm idea of what kind of services it will want of such a bank - whether it will be a very basic operation, extending current Post Shop technology, or more of a "full service" bank, requiring a more major change in IT systems.
One wonders what kind of meaningful information they could possibly get.
And once again, it appears they want it all in the next few months; Post and the politicians are talking about having a system up by the end of the year.
There has always been a balancing act with RFIs; does the client specify requirements in reasonable detail - opening the possibility that brilliant alternative ideas could be missed - or do they come out with as general a proposal as possible, leaving lots of loose ends?
To take the latter course is to rely on the vendors to do a lot of the work of firming up details of the system - work that they might consider to be the province of the client. This may quickly lose some promising providers who may have done a good job on a more fully fleshed out system.
Political or commercial imperatives will always impose time limits. How to get round them and still produce quality is a tough problem. Perhaps more attention should be paid to what has already been done, elsewhere in Government, elsewhere in business, elsewhere in the world.
Companies and government departments should talk to one another more than they do, and be willing to share the aspects of one already built system that may be applicable to another.
An overseas government department may well have had to face a problem almost identical to our own, and with slight tailoring their system may fit New Zealand.
We may have to swallow Kiwi pioneering pride. Not a complete answer by any means; but it may help fit within those impossibly tight timescales.
Send email to Stephen Bell. Send letters for publication to cw_letters@idg.co.nz.