When we started Woobius back in 2007, we believed (rightly or wrongly) that we needed to be able to provide a relatively complex, "UI-heavy" front-end to make things simple for our non-technical users. For example, one of the key screens in Woobius is an explorer-like file browser, loved by users, and only possible because of the technology we picked.
We surveyed the field in 2007, which was in the very early days of jQuery, Dojo and other rudimentary toolkits, and came to the conclusion that it would be very hard indeed to use them to deliver our vision (particularly since IE6 support was not optional for us).
So we picked... Adobe Flex.
Based on our perspective, it was the right choice. But our perspective was wrong. There was a factor which we didn't take into account, though we knew about it. It was clear to us that, because of the greater dynamicism and wider community behind javascript toolkits, they would eventually surpass Flex's power, flexibility and speed (and indeed that is what has happened). We knew Flex was on the wrong side of history, we just didn't realise how much it mattered.
As a technology cofounder, one of your unspoken functions is to be a visionary. Businesses are long-term endeavours (or should be). When you make fundamental technology choices for a business, they are there to stay. The longer such a choice is made and not changed, the longer it will take to change it. So, when we chose Flex, we chose it not for the short term (as we may have thought), but for years to come.
And, in hindsight, that was a mistake that cost us.
The cost of using Flex
Adobe (perhaps seeing for themselves that the writing was on the wall for Flex, though I think it was probably sheer laziness) never stood behind Flex the way they should have. They let ridiculous bugs and limitations fester over the years, and in the meantime, the chaotic horde of Javascript toolkits advanced, took over, became superior, better, and became the standard.
All the community effort that took the web into the design-centric UI richness that we know today went into Javascript, not Flex. Flex is now a relic that must be extirpated at great cost to guarantee a future to the Woobius collaboration product. And this is only 4 years after starting development.
When Kublax (a Mint clone) closed its doors in February 2010, articles mentioned there was an alternative called Money Dashboard. I went to check it out. I signed up. I went to the login screen, and I felt this sinking feeling in my stomach. Money Dashboard uses Silverlight. Oh woe! In 2010? Sure enough, within a year, Silverlight was doomed by Microsoft themselves (followed, a year later, by Flex itself). The anonymous tech cofounder of MoneyDashboard made the same mistake we did, though even more flagrantly. I never even bothered to log in, in the end, put off and saddened by the choice of Silverlight.
The history test
When you're picking what technologies to use in your brand spanking new startup, your responsibility as a technology cofounder is not just to make the best choice for today, but to look 3 to 7 years into the future, and make a choice that will be sustainable for at least that long, if not longer. Don't lock your startup into a technology path that is a dead end.
Yes, technology changes rapidly and somewhat unpredictably. That is precisely why you need to make sure that the fundamental, critical pieces of technology that your company depends on are there for the long haul, not dependent on a single vendor, and, most importantly, pass the "history test".
Every time you choose a critical technology to lock yourself into, ask yourself: "Is this technology on the right side of history?"
If it's not, think again.
If you read this far, you should follow me on twitter here.