There's a Second Life cover story in today's USA Today that mentions an interesting aspect of our development previously not covered before - the challenges created by consumer movement from desktops to laptops.
In the early days, we were modeling a number of assumptions relevant to our adoption and addressable base. Consumer broadband was the big one, followed by graphics cards. I was keeper of these charts and with each new Dell catalog, would excitedly benchmark what $500, $750 and $1000+ could buy. We needed the entry level PC to run us out of the box. It was happening too -- desktop prices were falling quickly and with the exception of enterprise models (which often left out a graphics card to save $), every new computer was shipping with solid technology to run SL at reasonable levels.
Then WiFi started getting hot and laptops became the new PC of choice faster than we, or analysts, had predicted.
These notebooks were powerful enough to run most applications but 12-18 mths behind in terms of processor and graphics cards. D'oh! We hoped that these notebooks would become the second computer in a household, but instead they started to serve dual purposes as both an office and portable device. We were in danger of missing a hardware upgrade cycle, especially troubling because these new laptops were disproportionately being purchased by a younger demographic and web creatives, target markets for SL.
As the article notes, "that wait severely tested Linden Lab's financial reserves" and was one of the factors in the 2003 cash crunch.