|Re: Strahler numbers firstname.lastname@example.org (Tomasz Kowaltowski) (2010-07-15)|
|Re: Strahler numbers email@example.com (George Neuner) (2010-07-16)|
|Re: Strahler numbers firstname.lastname@example.org (Tomasz Kowaltowski) (2010-07-21)|
|Re: Strahler numbers email@example.com (BGB / cr88192) (2010-07-21)|
|Re: Strahler numbers firstname.lastname@example.org (2010-08-02)|
|From:||Tomasz Kowaltowski <email@example.com>|
|Date:||Wed, 21 Jul 2010 08:37:15 -0300|
|Keywords:||registers, optimize, history|
|Posted-Date:||22 Jul 2010 13:34:30 EDT|
>>> [How does it compare to Sethi-Ullman numbering? -John]
>>Look at http://compilers.iecc.com/comparch/article/08-05-013 :-).
>>[Oh, ah, er, right. -John]
> That post doesn't explain much. AFAIK, Ershov's (1958) work
> introduced Strahler's work (1952) into computing by applying it to the
> evaluation of expression trees. Ershov used it to identify what he
> called the "register function" of an expression.
> IMO, they ought really to be called Horton numbers because Robert
> Horton first introduced them as an analysis technique for stream
> mapping in a seminal paper in 1945. Horton died shortly afterward,
> and it was Strahler who went on to apply Horton's methodology and
> develop it into a general statistical model of streams.
You are right; as a matter of fact the Wikipedia article
cites both names: "Strahler number or Horton-Strahler number". On the
other hand, the numbering algorithm for expression trees is quite
obvious and was certainly rediscovered by many people who never heard
of the work of Strahler, Horton or Ershov.
Return to the
Search the comp.compilers archives again.