4 ms·

Err, that's kinda greatly simplifying things. I'd say the last 10 years can be described as: - ~2005 to ~2009 was the Ray Ozzie era. Ozzie took over the role

by tdicola 11y ago

Err, that's kinda greatly simplifying things. I'd say the last 10 years can be described as: - ~2005 to ~2009 was the Ray Ozzie era. Ozzie took over the role of Chief Software Architect from Bill Gates in 2006 and the company, internally and externally, was focused on very ambitious projects. Software 'architects' were common on every team and usually doing something the 'right' way trumped shipping quickly and iterating. There were lots of interesting ideas and cool tech, but turning that stuff into viable products and businesses proved to be difficult. - ~2009 to 2013 was the Steven Sinofsky era. Sinofsky took over as the head of the Windows division after Vista and helped recover with the Windows 7 release. His influence quickly spread and most of the company adopted his management structure (very strict separation of concerns between development, test, and product management). The role of software architect quickly disappeared (and Ray Ozzie left the company, leaving the CSA position to disappear entirely) and although research was still important it was definitely not the top priority. There was much less focus on building platforms, ideas, or other intangible things vs. shipping to customers in controlled and predictable ways. Post 2013 and Steve Ballmer stepping down I don't really know much since I left around that time.

smacktoward 11y ago
> 2009 to 2013 was the Steven Sinofsky era... There was much less focus on building platforms, ideas, or other intangible things vs. shipping to customers in controlled and predictable ways. For certain definitions of "controlled and predictable," yes. For others not so much. Sinofsky's tenure was also the era of the great push to turn Windows 8 into the one Windows to rule all devices, which resulted in it being the controversial, poorly-received product it has been. "Controlled and predictable" would have involved rolling out changes like moving to Modern/Metro/Windows-Style/etc. UI, removing the Start button, booting to the Start screen instead of the desktop, etc. slowly and incrementally, but Sinofsky bet instead on rolling them all out at once in a single "big bang" release. That was a big, risky bet that completely failed to pay off, and that MS has been retreating from step by step ever since his departure.
- dpark 11y ago
  From my perspective, "controlled and predictable" with Sinofsky was about managing the release cycle, hitting the dates. e.g. We said we're shipping in May, so we're shipping in May. That means careful planning and aggressive feature cuts if necessary. The goal was on time with high quality, where quality is measured in metrics like crash rate and bug count, not public opinion surveys. Incremental releases and minimizing controversial changes might make releases more "controlled" in different ways, but those weren't the target. [Edited for clarity]
  - serge2k 11y ago
    That sounds wonderful to me right now.
goalieca 11y ago
Sometimes I find "architects" curious folks. XAML was developed during the ray ozzie era right? That certainly has a "right way" with "Model-view-view-model" levels of abstraction insanity. Perhaps simpler is better.
- wwosik 11y ago
  MVVM is a very good model though. I certainly enjoyed working with it.

by tdicola 11y ago

smacktoward 11y ago

> 2009 to 2013 was the Steven Sinofsky era... There was much less focus on building platforms, ideas, or other intangible things vs. shipping to customers in controlled and predictable ways. For certain definitions of "controlled and predictable," yes. For others not so much. Sinofsky's tenure was also the era of the great push to turn Windows 8 into the one Windows to rule all devices, which resulted in it being the controversial, poorly-received product it has been. "Controlled and predictable" would have involved rolling out changes like moving to Modern/Metro/Windows-Style/etc. UI, removing the Start button, booting to the Start screen instead of the desktop, etc. slowly and incrementally, but Sinofsky bet instead on rolling them all out at once in a single "big bang" release. That was a big, risky bet that completely failed to pay off, and that MS has been retreating from step by step ever since his departure.

dpark 11y ago
From my perspective, "controlled and predictable" with Sinofsky was about managing the release cycle, hitting the dates. e.g. We said we're shipping in May, so we're shipping in May. That means careful planning and aggressive feature cuts if necessary. The goal was on time with high quality, where quality is measured in metrics like crash rate and bug count, not public opinion surveys. Incremental releases and minimizing controversial changes might make releases more "controlled" in different ways, but those weren't the target. [Edited for clarity]
- serge2k 11y ago
  That sounds wonderful to me right now.

goalieca 11y ago

Sometimes I find "architects" curious folks. XAML was developed during the ray ozzie era right? That certainly has a "right way" with "Model-view-view-model" levels of abstraction insanity. Perhaps simpler is better.

wwosik 11y ago
MVVM is a very good model though. I certainly enjoyed working with it.