Then an
interesting thing happened:
1) Microsoft released a version of Windows that
actually worked.
2) Intel consistently made new computers that were multiple
times faster than the old ones.
The
world changed. I changed my tools to run
on Windows, changed the products to run on Windows and added windows features. We dabbled in mapping for several years,
without success; then Microsoft created a mapping program that worked, and only
cost a hundred dollars a seat; it's initial interface was incomplete so I
patched it (the assembler thing comes in handy sometimes). The company grew significantly (10x) and
pumped seventy or so man-years into a new PC version of the product. I wrote the core messaging infrastructure and
solved problems whenever no-one else could figure them out; occasionally I
would be unsuccessful, so I trained management to believe that a problem I
can't solve is by definition an "unsolvable problem"; it was fun.
In those
days, if a customer had a crisis, I was the fourth person to show up (because
the previous three were unsuccessful). I
was bit of a cowboy (that's an understatement - more like Denis Rodman or Darth
Vader).
That was
phase two of my third job. I created,
wrote down, and pursued my mission and vision:
Mission:
Create software that enables the company to be highly successful.
Vision: Our
products are reliable, high performance, feature complete, scalable, easily
deployable, tightly integrated, and easily modified.
Strategic
checklist:
* Reliable
software continues to perform when something unexpected happens.
* High
Performance software has transactional response times less than 1 second under
full system load.
* Feature
Complete software has all the features a customer role player would expect in
each module.
* Scalable
software gives a consistent user experience with all customer sizes and
infrastructure.
* Easily
deployable software is installable by anyone, without training.
* Tightly
integrated software gives a consistent user experience.
* Easily
modified software has a high percentage of all software features exercised by
testing software with no manual labor
I was
motivated by the dream of doing this better.
My hobby was reading PhD and Master's theses from the Internet and
studying Bayesian statistics.
No comments:
Post a Comment