Open Tech Today - Top Stories

Thursday, November 30, 2006

Lessons from the Open Source Frontier #2

Things with open source software are not always what they seem, or what is reported. Just ask the City Council of Birmingham (UK). Critics declared their open source effort a failure, but not so fast. Reports of the death of open source in Birmingham are greatly exaggerated.

According to its Head of IT, the City Council actually expects to realize cost savings over time, and contrary to press reports it plans to "significantly increase" its use of open-source.

Which brings us to a few additions to my original Lessons from the Open Source Frontier.

Lesson #3: Open source requires skills.

You can build them up or buy them. More likely, you should do both. But either way, a real migration is involved and requires experienced people--techs, installers and troubleshooters--to manage it. The level of skills on staff will impact the "team costs" (like project set-up, technical design, development, testing and training) that were so high in Birmingham. Investment to acquire those skills: real. Value of those skills for the next open source project: priceless.

Lesson #4: Objects in the mirror are less costly than they appear.

Government budgets are often short-term. The true TCO of technology projects is not. Costs for open source may differ greatly in the short term and long term. Special discounted license rates--such as offered by Microsoft to governments--affect the cost comparision in the short term. But start-up costs are not the only consideration. Vendor lock-in has its price. You never develop the technical and managerial skills needed to have more choices in the future. And the costs of your data/documents trapped in proprietary formats will always be there.

Even the iMpower Consulting report criticizing Birmingham notes:

"The extra resources involved in decision making and project management mean that the cost of this first-time open source implementation for BCC was significantly higher than for a comparable proprietary upgrade."

So, costs were higher the first time around. Is this a surprise? How much did the Windows license cost the first time around? The obvious implication is that management and training costs will be lower for future open source implementations.

Lesson #5: The price of lock-in is high, as is the price of freedom from it.

Identifying and eliminating technical "lock-in" is hard work, as Birmingham discovered. There are endless ways in which proprietary applications and configurations obstruct porting to any other technology. That is the price of lock-in, and it is permanently steep if you continue to live in a world of fewer choices. If you never move, you never feel the weight of the chains around you.

Lesson #6: Consider putting the cart before the horse.

Sequencing matters. Birmingham's rollout plans were ambitious, especially given the technical levels of its staff. A better business case may have been to begin by migrating applications before operating systems. Designing and implementing a Linux desktop system can be difficult. Yet, adoption of open source applications like Firefox and OpenOffice is often easy for users, as Birmingham learned. Either way, it should be budgeted and managed as a multi-year effort.


Solveig Haugland said...

Thanks for posting this detailed blog on what actually happened. It's easy to expect cheap from open source, when there are of course transition and retraining costs. But like you say, lock-in is painful.

I'm very excited to see what will happen now that Vista is starting to be available, at least to business customers, and with the new MS Office coming. The cost is going to be big. What will people decide to do?

As for Vista specifically, the actual operating system is mattering less and less to end users, since what do they use? Their browser, which is trivial to learn, and their office suite, and most users use 10% or fewer of the available features. Take a look at the toolbar of Writer and it's very similar to Word. Any Word user can start using the core Writer features without any training whatsoever.

The coming months will be a very interesting time in the open source world, for Microsoft, and of course for all the IT managers making their choices.

Jeff Kaplan said...


Thx for the thoughtful comments. Based on feedback from government officials in several different countries (not all are open source proponents), I think Vista's prospects are very uncertain.

First, you're right that Vista is a supersized software. Most users don't need all the stuff in it. So why should they pay for it?

Second, there is a bigger cost issue with Vista than its licensing. Apparently, its use would require many agencies and enterprises to purchase new hardware. This could be a huge expense that leads people to decline the upgrade, even if Vista is essentially given away for free, as some have suggested will happen.