__/ [ Larry Qualig ] on Thursday 25 May 2006 15:30 \__
> Michael B. Trausch wrote:
>> Larry Qualig wrote in
>> <1148522818.760840.152990@xxxxxxxxxxxxxxxxxxxxxxxxxxxx> on Wed, May 24
>> 2006 22:06:
>> >
>> > The Ghost In The Machine wrote:
>> [snip]
>> >> Note that $33.51B profit is about $91.75M/day, or $1500/day/employee.
>> >
>> > Numbers like this are absolutely staggering. The $1500/day/employee is
>> > *profit* - after the rent, employee's salary, fines, congressman, and
>> > everything else has been paid.
>> >
>>
>> Yes, that is quite an amazing amount of money. To be raking in $91
>> million
>> in profit per business day is absolutely astonishing. Unfortunately,
>> they're not putting it to good use, or they'd have fixed the problems in
>> Vista already and had it ready for release. I have to wonder if they're
>> doing so well financially, why do they have such a hard time turning out
>> quality code?
>
>
> This is an interesting statement/observation in a number of ways. My
> view on this is that having financial resources is largely (but not
> completely) orthogonal to being able to creating quality code. This
> money gives them two basic options: Hire more developers and/or hire
> better developers.
The developers need also cope withlarge amounts of code, which as we know, is
difficult to get a grip on. With morale declining and staff jumping ships,
you need to train new staff. I have worked with bad libraries in the past
and I know the consequences. There is a reason why 60% of the code need to
be rewritten and Allchin scraper Longhorn in September 2005. He said "his
engineers could not run it properly".
> Throwing more people at the problem doesn't scale very well. Eventually
> you reach the point where you have "too many chefs in the kitchen" and
> people start getting in the way of each other.
I heard/read that only a coupla' hundred work on Vista. That's nothingwhich
Novell or Red hat cannot match. Besides, they have OSDL, KDE and the other
groups working in isolation in accordance with standards (API's, modularity
frameworks/specs).
> Hiring better developers doesn't work either IMO. There are already
> some incredibly smart people at Msft. Then you need to consider
> familiarity and experience with the product. Someone who may be
> marginally more intelligent than another developer won't be as
> productive as the other developer who has 8 years of experience with
> the product.
>
> I'm also going to take issue with the "quality code" not in absolute
> terms, but in a business sense. Keep in mind that Microsoft is a
> business, one of the largest in the world, and from a business
> perspective there really isn't any incentive to create a perfect
> product. It simply has to have reasonable and sufficient quality.
Then returns the issue of morale, as well as /drive/. Ballmer preaches about
the sacking of people. This means that training of new staff becomes a
peril.
> The example that you may (or may not) have heard is that Detroit could
> build a car that was 99.999% reliable, was incredibly safe and lasted
> for 500k miles. The problem is that such a car would be so cost
> prohibitive to manufacture that few would be able to afford it.
>
> That's somewhat the issue with software. Creating a
> 'perfect/great/excellent" product takes time, resources and money. If a
> company can get 80%-85% of the way there for 50% of the cost we all
> know what they'll do.
I will always remain curious, Larry. I wonder how much money you made from
your venture/involvement with Microsoft. I suspect your life without this
reward would have been altogether different, which is why your opinions are
taken with a barrel of salt. But I can't take away the credit.
|
|
