[Sussex] A Quality product from Microsoft

[Steve Dobson]

John

On Tue, Jan 25, 2005 at 12:46:03PM +0000, John D. wrote:
> I've got a quicker and easier challenge for Steve, and that's to find
> any location whatsoever, where the M25 intersects the M27 ?????
> 
> Fortunately, being a charitable person, I suspect he actually meant the
> M26/M20/M's 3,4,40,1 or 11 (or at a scrape even the M23) :-P

You are (of course) correct.  I was confusing the M27 with the M23.

That's the great think about this list, there is always someone to point
out your mistakes and guide you back to the "one true path".  :-P
 
> Sorry Steve, couldn't resist.

That's okay.  Enjoy your moment of fun (I enjoy mine) :-)
 
> Besides, if I really wanted to know the quickest/easiest route to
> Haugesund, I'd just ask my mate Daves' ex-wife. She was born there and
> as most of her family are still resident, well .........

Aarh, local knowledge - the best.  So ask!  What is the best route?
 
> Still, I thought that it was a good "snipe" at MS despite the apparent
> difficulties of producing an accurate/efficient alogrithm!

I'm sure that Microsoft's algorithm is good.  It did a perfectly reasonable
job on the route between Eastbourne and Ashford - not the route I use,
but then no computer planner can beat human expertise.  

To me it points to something worse:  Microsoft has not enter all the
appropriate data for Norway.  Geoff pointed out the lack of bridges and
the need to use lots of ferries.  If M$'s database lacks ferry data 
because they made the same bad assumption I did (that for a short
geographical distance road would be faster) then what else have the
left out of their database?  What have they left out that is important
on my route?

In this instance it failed silly.  I doubt anyone that saw the route
concidered it reasonable.  But what about a failure that looks reasonable?
On a route I don't know what is it not telling me?

Steve