Monday, April 09, 2007

iGov ictQatar IT Forum

TOO MANY ACRONYMS! The Gulf Times (and probably every other paper) today announced there's going to be a meeting at the Ritz tomorrow to talk about developing government IT services so they can be accessed through the web. I could probably make some useful suggestions, if I were invited. But I'm not, and neither are you.

Why bother advertising a meeting if it's not open to the public? That's like me issuing a press release saying "Family meeting, 7.30, around the dinner table. We're going to talk about the state of the bathroom." Of course, Wednesday's papers will follow up the story with photos of Hessa al-Jaber shaking some guy's hand. There won't be any details of what was discussed - just a photo and a repeat of the purposes of the meeting. Grr.

Given that I'm not invited, here are my suggestions. The important ones are non-technical.

  1. Scatter the money around. Don't spend millions on a single project to do everything - it will fail, and all your money will have been wasted.

  2. All initial projects should be finished within three months. A two-year project will fail, and you will be two years behind schedule.

  3. After three months, see who's got something working - give them more money and more responsibility. Stop funding the projects that haven't achieved anything; look for new small projects instead.

  4. Replace the leaders of failed projects. Publicise the successful projects.

  5. Insist on compliance to internet standards. If someone is using an ActiveX control, Macromedia Flash, or non-validating HTML, take away their money. They don't understand the web, and eventually you will be embarrassed if people believe their solution is representative of the State of Qatar.

  6. Insist that people use a free database in their solutions. It saves on costly licenses for still-born projects, and the database code will be cleaner if programmers believe they need to swap from MySQL to Oracle at a later date.

  7. Apart from following internet standards, and using free databases, don't enforce any other standards. In particular, enforcing Microsoft products will halve your pool of talented programmers; enforcing project management and system analysis methodologies will double the bureaucratic overhead. Both of these will encourage failure.

10 comments:

sandy said...

Not being invited doesn't mean you cannot write to "them", preferably in a cool and constructive way suggesting how progress might be achieved in a step by step fashion. Also remember that even though you think you are right, you are in Qatar which does things in a different way and which you must go along with, or remain silent.

Anonymous said...

Still not convinvced by MySQL for anything other than Noddy projects. -- Paul

Nigel said...

Paul, I sort of agree with you a little bit - I think commercial databases have a place.

However, do you remember the Oracle cluster we bought for the replacement content-management system at ITV? It cost an arm and a leg, and was never used.

My argument is that software projects have a high failure rate, so new projects should always use a free database until they've proven themselves. If the coders stick to SQL, the transition to a commercial system should be easier, and the logic stays in the application server.

I know you have arguments against using a database purely as a data store, but a lot of people would disagree with you. Switching databases between prototyping and production enforces better design.

Nigel said...

Sandy, there are practical problems in writing to people who don't respond to email. I'm more likely to receive an interesting response to a blog entry here than an individual post.

You have a point in that people should be able to make their own mistakes. However, many of the failed IT projects are born out of bad advice from western IT consultancies, happy to work with daft requirements so long as the project limps onward, slowly spiraling toward inevitable death.

It's not culturally insensitive to point out that Qatar is being gouged by the major vendors and consultancies. Surely, it's important to point out that not all westerners are crooks ;)

No consultancy is ever going to mention the high failure rate of IT projects, which might push the client toward smaller, less risky ventures.

Even though consultancies are nominally independent, working on behalf of the client, everyone in the industry knows that consultants only recommend a small portion of the suitable vendors - those who will prove the most profitable for the consultancy.

It would be extraordinary to hear a consultant say "start with small, cheap projects, then see what works". They'd never sell any unnecessary products and services, which is how their business survives.

Anyway, I'm slightly off topic, talking about con-sultancies. The real point is, if you don't have an established IT industry, and you can't trust your advisors, wouldn't you appreciate a bit of honest talk?

Aisha said...

WOW Nigel you really should’ve been invited! I’m impressed.. I couldn’t agree more about the media game thinggy.. and that such “meetings” have to be open for anyone.. let alone for experienced people know KNOWS what is needed, like yourself.

Till that day comes..You really should apply for a job that allows you to be invited ;-) You deserve it more than many.. Can I help with that? PLEASE Nige let me do that!!

sandy said...

There really is no point sitting on the outside criticizing. That just undermines your own opinion and ensures you wont be listened too. Be helpful and constructive. I also think your views on consultancies are a bit naive as they survive or fail based on reputation alone. How do people get on in the world - by working with the people who currenlty hold the power and influence, and not by criticizing them. If you do the latter the only support you get is from like minded people who are in similar boats.

Nigel said...

Aisha, thanks for your kind support. I'm always interested in hearing about job openings. If you hear of any, let me know - my email address is ngourlay@gmail.com.

Sandy, there were a couple more details published in the paper this morning:

"UK's gov3 managing partner Graham Walker, considered an expert on government e-transformation, maintained that the ideal e-government option was the single service option, wherein people could access all services on one platform."

In the same article, it says these services will include examining patient records, library databases, registering trademarks, etc.

The consultants have a homepage that doesn't validate under even the loosest definition of HTML. It doesn't even display in one popular browser. I checked at Companies House, and they were exempted from lodging accounts because their business was so small. The project has 'SPLAT' written in bold type across it.

Aisha said...

Nigel I know absolutely nothing about techno stuff.. but I did print the post out and gave it today to an instructor I know from our university:
http://www.qu.edu.qa/qu/colleges/arts_sciences/cas_mass_people.html

Tomorrow I’ll get his feedback .. I’d like to know what you exactly do because this guy he really is a ginous and keep talking about the things mentioned in your post.. so I figured he’ll be interested and will help us reach out to people in charge ;-) yaay Nigel you’ll be very rich!!
I have no idea why I’m so excited.. I was very sick but I made it to school today for this then I came home back and didn’t attend my class loool
I’ll write to you via your email soon

Nigel said...

Hehe. I got to the page - thank you Aisha.

I'm an open-source computer programmer. Open source is the term given to software released with relaxed copyright restrictions. The software is developed and shared freely - the operating system Linux is the most famous example of this approach to technology, but there are thousands of small projects that are less well known but which are the glue that holds the web together. QL is built exclusively with open-source software.

The companies where I've worked have been a mixed bunch. Dot-coms startups, TV companies, and a university. Most of them were happy to let me publish the software I was writing at the time.

I was involved in the IT community in London, and when I moved to Qatar I started collecting email addresses of people who were interested in open-source. That became Qatar Perl Mongers, which is really just a bunch of people who like to chat about IT.

Nigel said...

Here's some comments from a friend, sent by email. I've anonymized his company.

I agree with all the points that he makes, but the reader should bear in mind that he works with one of the best teams of developers in the world, with many years' experience of large-scale development. Such teams don't yet exist in Doha.

> > >

- Only pay the vendor once you go into production. If the project fails the vendor isn't paid.

- Ask the vendor to be a strategic partner. So not much money up front for the vendor, but license fees will come rolling in once the first few projects have succeeded and legacy systems are converted.

- whilst enforcing project management methodologies might be bureaucratic, each team must choose some way of project management and measuring progress. Both the team and management need to know what the status of each team is, to reassure them that the projects are going to succeed. If the team can see that the project is going well, then that acts as a good motivator. The best motivator is off course going live.

- long projects can work provided the people are good enough. The project I'm working on is going live in about a month. There now about 15 people working on it in three locations. Whilst we are about 4-5 months late, (management screw up allocating resources to earlier versions) I'm reasonably confident
that we can hit these dates because
1. all the tests (manual and automatic) pass
2. only data migration is left to do.
3. the people working on the project are mostly excellent, so it wasn't too difficult to get back on track.