Luap wrote:
We regret to inform customers that our Chief Engineer, Dr. Michael Morgan suffered a totally unexpected grand mal seizure Tuesday, while doing final pre-production testing on Einstein. Fortunately we have an EMT capable staff member who saved his life and we got him to the hospital in time. Mike is home resting now and begging to finish up his work at home despite a pea size brain mass that is thought to be the cause, and which is yet to be fully diagnosed. (...). We anticipate having our production vendor produce about a dozen Einsteins in, hopefully, full final production form, next week. Following successful evaluation of these units we should be ready to go into production.
My own work schedule is already about 14 hours - 7 days a week. We have been scouring employment agencies for additional highly capable engineering help for a year now. Seems our requirements are higher than the available work force. Fortunately we hired one very capable engineer about three weeks ago, but getting up to speed on this stuff takes time.(...)
Paul, first off I join others in wishing Michael a favorable diagnosis and speedy and full recovery.
Secondly, based on what you had described, even before Michael's incident, i.e. as of this past Monday, the promise to ship any units February clearly was TOTALLY unrealistic. The milestones for Einstein delivery that i see outstanding are:
1) finish Michael's and any other testing work before the final production run, spending 4-5 hours per day 4-5 days per week - not 14 hour days 7 days a week. Because you tend to other matters like the CyberCommander issues and this forum and staffing. And you need your rest, neglecting which simply makes one liable to overlook things in testing. Einstein is a complex product and we do not want it to miss anything like the off/reset switch on the CC.
2) if testing reveals any issues (and if it does not, i'd look again), some time needs to be set aside to work them into the design and unit-test them. Not having adequate time set aside for this simply establishes a psychological prerogative to NOT find any issues in testing.
3) the preproduction you've mentioned for a dozen units with final hardware.
4) test the units internally, once the units are in-house
5) send the units out to one and then the next group of testers. These should include power users of your stuff (who have constructively reported issues in the past rather than love everything but hardly use it), online opinion leaders, and other qualified folks.
6) rank their feedback on the 4 quadrants (risk/cost/time of implementation vs benefit) and implement the low hanging fruit plus any serious issues. Unit testing.
7) depending on the volume of changes and comfort level, another small batch may have to be made and evaluated internally.
8) first significant batch of Rev1.0 hardware can be manufactured and shipped.
How long? I'd say July-August. These steps won't guarantee a flawless v1.0, but feel that skipping steps and cutting execution time could take you to August just the same, only with more users who feel misled and with faultier hardware and deflated workforce. Right when you need them to expeditiously nail any issues that pop up (unlike what's happening with CC).
Just my opinion