[clug-talk] M$ revenue falls, open source "blamed"

[Curtis Rempel]

On Saturday 08 August 2009, Robert Lewko wrote:
> Yeah, OK a talented programmer.  Yeah he was working on Basic.  Whoopee!!!
>
> Then how is it that when this "talented programmer" gets into the position
> of chief architect that he hires David cutler, who was fired from DEC for
> incompetence, to architect NT?  Then because the performance of NT 3.51

This is simply not the case at all.  

Cutler left DEC because Prism was cancelled in 1988 and he was fed up with the 
bungling within DEC management.  His impatience for fools is well documented.  
You don't receive high praise from the likes of Gordon Bell and become a 
recipient of the National Medal of Technology and Innovation at the White 
House if you're deemed incompetent by your peers.  Cutler was the architect 
of VMS (aka OpenVMS) in the 70's, well known for its rock solid stability, 
scalability, and clustering, technologies which other vendors are still 
grasping at.  When uptime is measured in years (cluster uptime now in decades 
for some customers), when lives are at risk, when a few seconds on a delayed 
transaction can cost hundreds of thousands of dollars, VMS is what is running 
the show.  Think banks, stock exchanges, healthcare, 911.  Think Deutsche 
Borse for example.  The baggage handling system in Frankfurt, the busiest 
airport in Germany and a major international hub, is running on VMS.  
Commerzbank during 9/11 survived when it's datacenter was destroyed in the 
twin towers because it was running a multisite VMS cluster with the other 
fully participating nodes in New Jersey.  This was not standby/failover that 
some vendors tout as "clustering" but rather the VMS single system image 
clustering technology.  A cluster transition of a few seconds occured when 
the link was lost and all transactions resumed processing on the surviving 
cluster members - absolutely zero data loss.  There are many many other 
examples of the quality, stability, scalability, and reliability of VMS.

So, it would be easy to argue that it was in fact DEC that was incompetent, 
certainly not technically, but from a marketing and managerial standpoint.  
Cutler reimplemented numerous VMS technologies in NT, however, the major 
problem as many have already noted elsewhere over the years, is the backward 
compatability baggage.  If he had started from scratch and not bothered with 
backward compatability, it's conceivable that NT could have morphed into 
something with greater stability and security, something similar to what VMS 
has always been and still is today.  An Edsel with a new paint job is still 
an Edsel.  NT was thus crippled from the starting gate.  I can remember 
firing up NT 3.51 on a 64bit AlphaServer in 1997 - fastest BSOD on the 
planet.   "Fired for incompetence" speaks not to Cutlers' skills, but to the 
inability of the bureaucracy within DEC to realize the engineering talent 
within its organization during the emergence of the RISC projects at the 
time.  Unfortunately, many of those managers retained their lucrative 
positions while the brains headed for the exits.  The same could be said for 
numerous other talented engineers... Richie Larry comes to mind for example.

Curtis