To avoid a žDeath MarchÓ project we should be defining the architecture now. We need FY01 to budget correctly.
There is no IceCube funding for a software engineer. I have been consulting with several people for free. But, to have an engineering estimate of the schedule and budget, we would have to ask for a large commitment of time÷ and pay for it.
If we start now, there is time to prepare for the upcoming reviews and build the detector according to a reasonable schedule.
If we wait a year, we will make unrealistic schedules that look good only on paper÷ which slip, causing the budget to be overrun. The project will be scaled down to the budget and ultimately the analysis may be compromised÷..