Also not wanting to rant but I've followed this discussion with interest as I have had over 10 years experience developing IT Solutions for multi nationals like IBM, GlaxoSmithKline (after 20 + years in Health management) and tended to apply my management understanding of their processes to the original specification. Using Rapid Application Development tools I would provide a quick "proof of concept demo" as a "point of departure" but it often represented 80% of the eventual design. At this point all fields were completely "open" in as much as any role could enter any value into any field and of more relevance no fields were obligatory at any stage in the workflow process. Only after this initial stage did the user have to address how to tighten up the security model underpinning the design. I, as the techie developer, would then ask them to consider what the implications would be for each additional element of control they wanted to introduce and challenged them to consider whether their existing processes needed to be revised and also what end user training might be required to make any chnage work. Before roll out the user training would be piloted and sometimes it would be realised that the best thing to do was to relax certain restrictions and to defer them until a later phase following further changes in the business process. Apparently this approached worked fine.... I kept getting asked to develop more solutions so I assume it was appreciated by the end users. Now the conventional wisdom, certainly with large scale projects is to tightly specify user requirements at the outset, including all aspects of the underlying security model. This is then allied to "must hit" deadlines and budgets imposed from on high so it's no small wonder that CSA staff for example had to "invent" bogus data just to overcome an over tight and unrealistic security model. BTW this is not just a problem with public IT projects. I once worked, for the stockbrokers Cazenove in the City of London, at the end of a project that was originally expected to take 3 weeks and cost no more than £20,000. I was involved 18 months into this project after £1.5m had been spent and was able to help deliver the first phase of what had always been a 3 phase piece of work (that's right each phase was expected to take just one week!) David Symes [log in to unmask] 0114 2303145 -----Original Message----- From: email list for Radical Statistics [mailto:[log in to unmask]] On Behalf Of Stephen Morris Sent: 13 April 2005 14:04 To: [log in to unmask] Subject: Re: Child Support Agency This is a subject dear to my heart - I'll try not to rant! It is true of course that those writing the system should know how it will be used, but even with in-house projects that is often not the case in my experience. And IT really is complicated: it is quite possible to get programmers who are so specialised that even their (very IT literate) line managers can't follow everything they can do. But putting that aside, I think a major problem is that the managers 'specifying' the system are often distracted by exceptional cases. If everyone accepted that the computerised system can only deal with 99% of cases, it is usually possible to build something which is easy to use, efficient and much much cheaper than the impossible dream of a system which will cope with everything. It is too easy for people to say, yes that's fine, but what happens when someone comes in and they've got this, and that, and that - there's no way for this to go 'onto the computer'. At that point there ought to be an exceptional process based on brains and paper. All those underspecified managers would be able to concentrate on these cases, rather than trying to control the IT projects. People only expect one solution for all problems when specifying IT, and it's just not realistic. Jill Szuscikiewicz -----Original Message----- From: email list for Radical Statistics [mailto:[log in to unmask]] On Behalf Of John Logsdon Sent: 13 April 2005 12:36 To: [log in to unmask] Subject: Re: Child Support Agency Ted's initiative in this CSA issue has perhaps come to the conclusion that the report was rather misleading although loosely based on fact - if that is possible. In the present atmosphere of course journalists will say anything to get a headline and it is a great pity that this has permeated even the BBC. But it raises the important question of the management of large IT projects by all governments. The habit these days seems to be to specify a solution then give it to some - even transatlantic - company to implement. This has the nice property that in theory the costs are ring-fenced but in practice this means that as the outsourced company does not have real ownership of the project, it is just interested in making a profit and moving on to the next scam as quickly as possible. Businesses will recognise the problem of course. If you want something done, do you hand it to a company that will quote you an overall figure then quibble about - or not provide - extras or do you pay someone essentially by the day but then find that your budget has been exceeded. Things in IT are not like repairing a car or building a house - IT is incredibly complex. Try specifying a web site. Of course lawyers will make a bob or two arguing over the contract - all paid for by the end user in the last analysis - and eventually the real owner of the project will have to cough up more to get the solution they really need. Now while we can throw stones at the company involved, isn't the real issue that the managers and specifiers of the process - they call it business process mapping - are generally totally incompetent when it comes to IT? They believe that they can separate the business process from the way in which it is implemented and are swayed by lobbying, slick salespersons and other methods into signing contracts. In fact it is unlikely that any business process is accurately mapped and even if it is, it is unlikely that the technology is being optimally used. So we get the worst of both worlds. The technology developed these days should enable a much better business to be made. The solution surely is that the people actually writing - or capable of writing - the system should be intimately involved in specifying and managing it, with some constraint on their enthusiasm to over-guild the lily. That way they can see from the bottom up how the solution meets the requirements. By handing projects to outsourcing companies, it is very difficult to see how this can be done properly. Of course this means that managers can understand science, IT and the like which leads us on to the poor technical education that many people have. I suspect most managers think themselves above such concerns but how wrong they are. I have little time for people who proudly claim they can't wire a plug and they are the equivalent. Best wishes John John Logsdon "Try to make things as simple Quantex Research Ltd, Manchester UK as possible but not simpler" [log in to unmask] [log in to unmask] +44(0)161 445 4951/G:+44(0)7717758675 www.quantex-research.com On Wed, 13 Apr 2005, Ted Harding wrote: > On 13-Apr-05 Stephen McKay wrote: > > Hope this helps. Happy to provide further details, but given > > the non-stats nature not sure if it merits whole-list attention. > > Much gratitude to Steve McKay for providing such a broad > spectrum of information about the coverage and provenance > of this story. > > Despite his last reservation (above), however, I'd like to > express the opinion that, in a group like RadStats, we're > not just concerned about the "Stats". We are -- or should > be, in my view -- concerned about the quality and provenance > of the data which underlie the uses to which the data are > put, often subsequent to technical procedures whose results > are valid if the data are valid, but questionable if the data > are invalid. > > At bottom, statstics is about information. > > Our group by nature should as much comcerned about the > information as about any technicalities of how it is handled. > > Provided, of course, that the technicalitiea are up to the job. > It looks very much as though the "CS2" system was not. > > I shall read with interest the computer press reports which > Steve has pointed out. > > Best wishes to all, > Ted. > > > -------------------------------------------------------------------- > E-Mail: (Ted Harding) <[log in to unmask]> > Fax-to-email: +44 (0)870 094 0861 > Date: 13-Apr-05 Time: 11:28:08 > ------------------------------ XFMail ------------------------------ > > ****************************************************** > Please note that if you press the 'Reply' button your > message will go only to the sender of this message. > If you want to reply to the whole list, use your mailer's > 'Reply-to-All' button to send your message automatically > to [log in to unmask] > ******************************************************* > ****************************************************** Please note that if you press the 'Reply' button your message will go only to the sender of this message. If you want to reply to the whole list, use your mailer's 'Reply-to-All' button to send your message automatically to [log in to unmask] ******************************************************* ****************************************************** Please note that if you press the 'Reply' button your message will go only to the sender of this message. If you want to reply to the whole list, use your mailer's 'Reply-to-All' button to send your message automatically to [log in to unmask] ******************************************************* ****************************************************** Please note that if you press the 'Reply' button your message will go only to the sender of this message. If you want to reply to the whole list, use your mailer's 'Reply-to-All' button to send your message automatically to [log in to unmask] *******************************************************