[clug] Stories of Microsoft technical meltdown: T-Mobile Sidekick/Danger 'Hiphop', all data lost

[Paul Wayper]

On 13/10/09 18:10, Daniel Pittman wrote:
> I disagree: this is the "eat your own dogfood" model, taken to the logical
> endpoint for a company that is big enough to buy the competition:
>
> You don't just hang on to the, say, Oracle box: you replace it with your own
> DBMS, so that your tools work on your own platform.

See, the funny thing here is that normally the reason for "eating your own 
dogfood" with software is to test it, to get feedback and to improve it.  This 
works best when the people using the software are also its programmers; it 
works badly when the people using it have no control over the software at all 
and are treated as customers (or, worse, rival factions) in the development 
process.

Guess what happens at Microsoft?  Yep, lots of internal politics, lots of code 
and political separation, and - worst of all for improving the software - 
resource accountability means that even if you did have access to the code for 
WinCE, and the knowledge to fix the bug you've just found, you wouldn't be 
allowed to fix it because you couldn't charge your time to the right cost 
centre and no middle manager is going to allow her budget to pay for someone 
else's improvements.

And I note with wry amusement that AppleInsider doesn't put Maemo or Moblin on 
its timeline...

This isn't the failure of "the cloud" per se, it's just a failure in who you 
trust.  "The Cloud" can be basically treated as a big data center - albeit one 
that you can't drive up to and pick up all your servers from when they go 
bust.  You look at the services they provide and what they charge, and decide 
what you can afford from there.

Have fun,

Paul