Re: Help B* and B+ Trees

["Tarjei T. Jensen" <tarjei@online.no>]

Robert Dewar wrote: 
> This is quite unfair to Don, he explains algorithms in a high level
> manner using abstract pseudo-code. MIX is only used in low level analysis
> of actual performance on a typical machine, i.e. to get a feel for the
> constants involved and go from O(n**2) to C*n**2, you need a concrete
> machine!

Not really. Any computer will suffice. If he had used Pascal like he did
with TeX and Metafont (I would of course prefer Ada 95) it would be
simple for any user to discover the differences. Pascal compilers are
widely available and it would be simple for interested parties to type
in the examples and analyze the results.

I believe he used mix simply because he likes bit twiddling. I don't
mind that, it just makes the books less readable and longer than they
could have been.

I have read his explanation for using mix on his web page and I think
the stated reasons are entirely bogus and that he will have a hard time
finding scientific evidence supporting his position.

It is strange that someone who worries so much about readability as to
create his own type setting system does not consider the impact his
choice of programming language for the examples has.


Greetings,

-- 
// Tarjei T. Jensen 
//    tarjei@online.no || voice +47 51 62 85 58
//   Support you local rescue centre: GET LOST!