[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]

Re: [dinosaur] Dinosauria reclassification joins Ornithischia and Theropoda in Ornithoscelida

On 03/23/2017 11:11 PM, David Černý wrote:
> Right, because no serious mathematicians currently work in the field 

I can't answer that but that is a real possibility.  When I confronted
the Math department with papers and texts, they immediately ripped it
apart in confusion.  For example chapter three of Felsenstein's text
went though 2 EDITIONS with a fundamental error on the computation
calculating out the numbers of trees occur which was finally corrected
on pg 23 of the third addition.

I mean, what is that about?!?  No serious work on this subject in
mathematics work would ever have such a fundemental error escape narly a
decade before correction, let alone be released on the first edition.

> or
> contributed to its development in the past. Mike Steel, who has spent the
> last 30 years exploring the math behind phylogenetic inference
I'm an old man so I'm not impressed with 30 years of anything.  Everyone
I know has spent 30 years studying.  I'm surrounded by people like that
and they don't produce garbage like the like a complete lack of
understanding NP completeness in papers I've read in journals.

What I'm telling you, and not debating, is that the understanding of
mathematics in the field is lacking.  Instead of worrying about Mike
Steel, just try to first grasp the fundamentals of algorithms, with a
text like Cormen, Leiserson.  And then spend some serious time with
George Strung.  It's not like these guys who created the mathematics for
trees, paths and graph theory, starting with Euler, were idiots and that
any paleo expert can pick this stuff up on a Sunday afternoon.

And while you are on your way to that, understand that publishing
journals doesn't produce an developed area of knowledge.  Slapping
together bits and pieces of partially understood mathematics areas
doesn't formulate a body of knowledge and YES, I don't trust the
software and I do question both the application and basis for the
theorem that these applications are coded on.  It's not like these guys
who revolutionized humanity through computer engineering and
computational mathematics are stupid, which is what you are implying here.

Finally, and this is DEFINITELY the last I have to say about this topic,
the code examples I have seen around the net have fundamental
programming and algorithm flaws in them.   I've seen problems with
concurrency, algorithms logic, static memory bugs and more.  So that
magic box that you depend on for results needs to be explored, discussed
with experts in the field, debated, and rewritten.

And listen David, don't worry so much about what you think you know.
Education is about learning what you don't KNOW, and what I have learned
in the last year of investigating the computational mathematics of
phylogenies is that there is a LOT more that I don't know then I thought
I knew.  Furthermore and the literature on the topic is in chaos, with
incorrectly applied theorems, mistranslated texts and proofs,
misunderstood proofs and applications, and full of editing errors. Have
you ever made the effort to actually trace original articles in the
literature?  I have found that supposed papers that state things don't
actually exist, but mere notations of announcements for lectures.  That
articles that are quoted as to supposedly proving something or proposing
something simply don't do that at all.  I can't imagine where that
leaves others who have nearly ZERO foundation in the essentials of the
very mathematics they are professing to ply.

> www.math.canterbury.ac.nz/~m.steel/Non_UC/publications.html), presumably
> doesn't count. Nor does Jerzy Neyman, who was the first to apply maximum
> likelihood to phylogenetic analysis of nucleotide sequences back in 1971.

Try learning Knuth, Landau, Fibonacci, De Moirvre, Verma, Euler, Aslam,
Preparata and Shamos, Bellman, Hu and Shing, William Pugh, Dorit
Hochbaum, Robert Floyd, Driscoll, John Chambers, Bill Joy, Dewer,
Kenneth Rosen, Alonzo Church, and most of all probably, simply
understanding EW Dijkstra.

BTW - Neyman may have first tackled the problem with maximum likelihood
with nucleotides, but the problem predates that by almost 10 years, if
not more.  And yet, most of that work has been problematic and it is yet
to be shown how to best solved the application of this theorem, a
theorem which rarely anyone understands, let alone code in Python.  A
large part of the problem is that the working solutions fail to
recognize that the theoretical math has been already explored for a
little more than a century, long before Watson met Crick.

It just sucks having to reinvent the wheel constantly.  What is needed
is more mathematicians and computer scientists applying themselves to
the problem and a free software code base that is under constant peer
review.  A consensus needs to be build on knowledge, not speculation.
We need consistency in terminology, a translation from one paper to the
next, and an application of an over all methodology to simplify the
syntax and formulation of problems.  You need a bible, so to speak.



So many immigrant groups have swept through our town
that Brooklyn, like Atlantis, reaches mythological
proportions in the mind of the world - RI Safir 1998

DRM is THEFT - We are the STAKEHOLDERS - RI Safir 2002
  - Leadership Development in Free Software
  - Unpublished Archive
  - coins!

Being so tracked is for FARM ANIMALS and and extermination camps,
but incompatible with living as a free human being. -RI Safir 2013