On Wed, Jul 2, 2008 at 7:48 PM, Sue Gardner <sgardner(a)wikimedia.org> wrote:
Without speaking on behalf of the tech team
necessarily, I would make
the observation that it seems generally conservative about spending.
That's not unique to Wikimedia; it's consistently been my experience
that tech people tend to be fairly cautious spenders. Our team did not
want to spend money until it needed to: that is not a bad thing.
Just wanted to underscore Sue's comment here.
I spent a number of years directly in charge of spending a technology
budget of size comparable to Wikimedia's.
The frenetic pace of new technology development, and the fairly short
lead times of equipment vendors strongly favored "just in time"
procurement practices: I regretted many of my more preemptive
purchases.
This reality resulted in an occasional bit of tension between myself
and my management, when sometimes I'd find myself two months from the
end of the fiscal year with a couple of hundred thousand still
unspent.... and a resulting scramble to get things taken care of.
Many of the external factors triggering technology purchases are
themselves bursty (new facilities, new applications), further
exacerbating the situation.
If we assume that moore's law applies directly to price/performance
(which isn't a totally outrageous assumption in many cases) money
differed on technology spending has an annualized return of 60%!
Thats astonishingly good: almost as good as printing money.
Obviously technology spending can't be differed forever, but when the
cards play out in a way which otherwise favors it, technology is
probably the best place to have a capital spending shortfall.
From that experience I think I've taken the
position that the best
results can be achieved by tolerating the bursty nature of
technology
purchases but keeping a gentle pressure towards consistency and good
forecasting.
I don't see any reason to believe that the WMF is doing a poor job
here. If there were years and years of significant shortfalls, I'd say
otherwise.