Re: REST impact paper

It was thus said that the Great J. Andrew Rogers once stated:
> 
> This is simply false, there is no evidence to support it. Properly
> designed code is typically 10-100x higher throughput and efficiency than
> open source, virtually across the board. Open source doesn’t prioritize
> it. I’ve made a living demonstrating this across most parts of the tool
> chain, often with code bases that are smaller than the open source they
> replace. Open source software is atrociously inefficient. This is its
> Achilles’ Heel. There are exceptions in open source, like compilers, but
> they are rare.
> 
> Which isn’t to say there isn’t terrible closed source code, just that some
> proprietary code vastly out-performs open code in most cases.

  And I've found that proprietary code has been horrible to work with, and
the times I have seen the code (which hasn't been often) I'm horrified by
what I've found.

  At work, we paid six figures for a proprietary network stack (SS7) that is
*so* fragile that looking at it sideways will cause it to fail.  The
interface is this arcane 60s style interface that is srsly vwl mprd.  Ick. 
And then I saw the code ... external function does a linear search for a
data structure, then calls an internal function that again, does a linear
search for the same data structure, hidden global state in library code,
just horrible stuff ...

  At least with open source stuff we can *fix* the problems and not have to
wait for the vendor to provide a fix.

> Anecdotally, the only open
> source software my teams use that hasn’t been superseded by vastly better
> closed source is LLVM. 

  I was playing around with the x86 SIMD instructions.  I have code I wrote
in the early 90s (in x86 assembly) that matched (I thought) quote well to a
vectorized version.  The core was an interative application of these two
functions:

	x1 = (A * y + B) * x * (1 - x)
	y1 = (C * x + D) * y * (1 - y)

  I wrote a version using the vector instructions of the x86.  I could
easily beat clang (which uses LLVM), but try as I might, I could not beat
GCC.  

> Everything from JSON parsing to computational
> geometry to database engines, on the other hand, migrated to proprietary
> implementations. 

  So what magic things were done to make JSON parsing faster with
proprietary code than with open source code?  Or alternatively, what does
open source do wrong?  I'm asking for specific examples and not a
generalized "it sucks" argument.

  -spc (I have a friend that bitches all the time about "open sores" and how
	bad it is, but everytime he talks about working with proprietary
	code it sounds even worse than "open sores" ... )
_______________________________________________
FoRK mailing list
http://xent.com/mailman/listinfo/fork