Comments on: Expediting Synergistic Paradigms http://www.linux-mag.com/id/7298/ Open Source, Open Standards Sat, 05 Oct 2013 13:48:18 +0000 hourly 1 http://wordpress.org/?v=3.1 By: pdlaw.net http://www.linux-mag.com/id/7298/#comment-987199 pdlaw.net Fri, 07 Jun 2013 15:08:36 +0000 http://www.linux-mag.com/id/7298/#comment-987199 Hello, just wanted to say, I liked this article. It was practical. Keep on posting! Hello, just wanted to say, I liked this article. It was practical. Keep on posting!

]]> By: matt.reilly http://www.linux-mag.com/id/7298/#comment-6351 matt.reilly Wed, 30 Nov -0001 00:00:00 +0000 http://www.linux-mag.com/id/7298/#comment-6351 The interesting thing about management is the infinite variety of mistakes that can be made. Those of us on <br /> the outside have the advantage of hindsight and ignorance of the more complicated aspects of SGI's extremely complicated situation. <br /> <br /> So we'll never really know just what went wrong. <br /> <br /> On the other hand, I'll never understand SGI's inability to recognize that <br /> <br /> 1. the world didn't need another x86 cluster integrator, <br /> <br /> 2. didn't want a whole lot of Itanium shared memory systems, <br /> <br /> 3. wouldn't pay for premium service on ethernet clusters<br /> <br /> and <br /> <br /> 4. desperately needed a classy shared memory x86 multiprocessor that was bigger than 16 way.<br /> <br /> There was only one real out for SGI: build something that was differentiated from the pack and that people would pay real money for. That was the shared memory x86 widget. They stalled, screwed around, and took far to long to deliver it. SGI engineering was more than up to the task -- what was management's contribution? <br /> <br /> They finally "announced" the system this year, but will it make it to market? <br /> <br /> So many mistakes to make, so little time. The interesting thing about management is the infinite variety of mistakes that can be made. Those of us on
the outside have the advantage of hindsight and ignorance of the more complicated aspects of SGI’s extremely complicated situation.

So we’ll never really know just what went wrong.

On the other hand, I’ll never understand SGI’s inability to recognize that

1. the world didn’t need another x86 cluster integrator,

2. didn’t want a whole lot of Itanium shared memory systems,

3. wouldn’t pay for premium service on ethernet clusters

and

4. desperately needed a classy shared memory x86 multiprocessor that was bigger than 16 way.

There was only one real out for SGI: build something that was differentiated from the pack and that people would pay real money for. That was the shared memory x86 widget. They stalled, screwed around, and took far to long to deliver it. SGI engineering was more than up to the task — what was management’s contribution?

They finally “announced” the system this year, but will it make it to market?

So many mistakes to make, so little time.

]]>