Wednesday, 15 August 2007

The problem with government IT

I have worked for government agencies over several years but I cannot for professional reasons mention names of organisations. The question we often ask here in the UK is - why do government IT projects go wrong so frequently and in such a big way? Recently the British Health Service IT programme has been revealed as crippled and the system to protect the UK from terror threats was years late. The abandoned and much criticised UK ID card scheme was in turmoil with increasing over-budget costs. So what exactly are the problems?
  • Schedule overruns
  • Budget overruns
  • Systems not working when delivered
  • Poor performance of some government staff
  • Private outsourcing partners over-charging
  • Constant changes to requirements
There is not much that we can do about the last point - changes to requirements sometimes result from news media reporting of government IT plans and consequent effects on politicians. Politicians who head government departments under criticism are always quick to demand changes to carefully designed systems in development. Every system has its critics. All the news media editor needs to do if find and pay them for a good 'wasting taxpayers money' headline.

Schedule overruns result from design changes as discussed above, poor system design, poor implementation and the fact that IT consultants and suppliers see government as a gravy train. They often give unsuspecting officials systems which are over-designed and over-engineered. Poor planning also results in government IT resources being unable to deliver what has been agreed at a higher level.

Budget overruns result from the above factors and the fact that supplier companies tend to over-charge government, where departments often lack staff with IT knowledge and business acumen during the contract negotiation stages. I have heard many stories of private firms adding superfluous items  to the bill for government IT contracts.

Systems not working - this is caused by the box-ticking culture in government where officials want to tick boxes to say what their department has delivered this, this and this as per plan. They are less concerned to hear if it works well or if what has been delivered is sustainable and scalable. By contrast in business, a lot of time goes into making sure that systems have to work and to continue to work to make or save money.

Poor performance of some staff - I am told that in the UK government department which handles shipping, in order to meet diversity quotas, people with no knowledge of shipping are promoted into key posts. Thus steadily the department is losing its core competence. This is happening throughout local and central government. Also private companies are flexible on pay and can poach staff from within government departments. Government cannot react to this because staff are in rigid grades and cannot be offered extra rewards to stay. The grades of government workers do not reflect actual performance - only what their job description says they are supposed to do. What happens over time is that the best staff are invariably poached and the worst remain, and little is done to tackle poor performance. Government departments rarely have the periodic clearouts of dead wood that are so common in commerce. From what I have seen the leaders of failed projects can still expect career development - in commerce they would often be fired.

Finally - why do government IT projects fail in such a big way? The old boxing saying - the bigger they come the harder they fall. Government projects tend to be too ambitious - driven by politicians' dreams rather than the caution of commerce.

Poor Johnny Taxpayer. He can expect no relief any time soon.

Wednesday, 13 June 2007

Skills Shortage

We all hear about the 'skills shortage' and how it it going to blight the lives of those poor souls in UK PLC. And yet in the midst of the shortage some of my friends, seasoned IT professionals, are unable to find work. They are nearly all over 40 or are good people with out of date skills, or have a family to maintain and cannot afford to work for low pay. So I think that this skills shortage really means that there is a shortage of people who are young and have the right skills. Also maybe a shortage of young people who will go to college to learn about IT, get into debt with student loans, and then work very cheap afterwards.

In other words there is not really a UK skills shortage, but a shortage of people who can be readily and immediately exploited by firms who do not want to train or retrain and develop their staff. And of course the IT recruiting agencies want people here and now - under the recruitment agency model of 20% of fist year's salary there is no profit margin in offering employers people with potential.

There are of course some good firms out there who will invest in and train people. But they seem to be increasingly few in number.

In the UK we still seem to trust too much to chance in building a skills base for the future. On the other hand government attempts to build the skills base seem to waste a lot of money for any progress made.

Friday, 4 May 2007

Web Portals - don't you just love 'em

I have worked with SAP Portal, Oracle Portal and Microsoft Sharepoint And you know few of my co-workers who worked on these platforms had a good word for them....

Let me explain. A portal is supposed to be an easy way of getting content onto the web, in a controlled and safe fashion. The problem is that Dreamweaver is quite easy with a little coaching, and raw html is more flexible than what the portals tools allow. Portals allow you to make simple pages with simple theme design. But for anything more specialised and bespoke, the amount of time trying to get your site to look precisely how you want it, multiplies by a factor of about 5 on web portals. Trying to get that extra 5% functionality always takes 30% more time, so in terms of time you are not better off than if you used a full web development framework such as Python-Django, ASP.NET or PHP-Zend.

Portals seem to fail more often than multiple sites on an old fashioned web server. I theorise that this is because they often feature pages packed with components (called variously portlets, iviews, web parts) each of which though small on the page is actually a full web application. So, many users loading many pages packed with many portal apps all causes the portal foundation (often based on J2EE) to crash. We have had a lot of trouble with our portal in BigCity Yorkshire as users packed the home page with apps all of which drew unaggregated data from one database. The strain on both the portal and the database proved too much.

Personally I think an organisation is always better off with a good content management system which sits on top of a number of technologies, and which allows those technologies to be mixed, matched and mashed. Also it is not a good idea to have many sites in one CMS instance, instead spread sites over a number of CMS installations.

Computer Science

Like many people in the IT industry I have studied Computer Science. And you know, I am not so sure that I am always in good company. Let me explain.

What do we have to do to graduate in Computer Science? Pass exams on paper. Does that really prove anything? Some of us are not walking programming language syntax manuals, so we lose marks. Then there is coursework: it can all be downloaded from the Internet. There is nothing one would have to do for coursework that could not be easily found.

Finally there is the fact that some of the worst people I have ever met in our business had degrees in computing, and some of the best only had a few GCEs (exams we all take at age 16 in the UK).

Personally I think that a degree in Computer Science proves that someone can persevere at something and is intelligent. But it does not mean that they will be outstandingly good at IT.