For the mathematically sophisticated

I found this very intriguing, and in places hilarious:
http://www.cs.berkeley.edu/~wkahan/JAVAhurt.pdf

While it is on the surface an attack on Java’s Floating-Point implementation, it is realy much more in the way of a fable, or morality tale, of the hubris that I have found in all developers (including especially myself) over the years.

Be sure to read the Ariane anecdote on page 22

Thus spake pgeerkens:

I found this very intriguing, and in places hilarious:
cs.berkeley.edu/~wkahan/JAVAhurt.pdf

While it is on the surface an attack on Java’s Floating-Point
implementation, it is realy much more in the way of a fable, or morality
tale, of the hubris that I have found in all developers (including
especially myself) over the years.

This is quite an interesting read.

“A healthy balance between Stability and Progress requires an approach
to the Management of Change more thoughtful than can be expected from
business entities battling for market share.”

That’s an excellent summation of why Java is in the state it is in 2011.

Be sure to read the Ariane anecdote on page 22

You would not believe how many times I’ve told people (mainly in my day
job) not to catch Throwable without a clearly-statable reason for doing
so… None of these have yet caused any rockets to self-destruct, but
you never know.


J.

I thought you would enjoy that. Not just Java, but many development projects (and languages) have suffered from poor management of change.

Some of the specific issues he laments have been addressed (see strictfp), though I am not sufficiently knowledgeable to accurately judge how well resolved they are. But the main points I got from it are that numerical analysis is a pragmatic science often far removed from the elegance of textbook formulas assuming infinite accuracy of computation; and that there is no substitute for getting one’s hands dirty.

Geek Alert! Step away from the keyboard (actually I took a look but I’m not
admitting to anything :slight_smile: )

-----Original Message-----
From: messages-bounces@vassalengine.org
[mailto:messages-bounces@vassalengine.org] On Behalf Of pgeerkens
Sent: Sunday, September 18, 2011 2:18 PM
To: messages@vassalengine.org
Subject: [messages] [Developers] Re: For the mathematically sophisticated

I thought you would enjoy that. Not just Java, but many development
projects (and languages) have suffered from poor management of change.

Some of the specific issues he laments have been addressed (see
strictfp), though I am not sufficiently knowledgeable to accurately
judge how well resolved they are. But the main points I got from it are
that numerical analysis is a pragmatic science often far removed from
the elegance of textbook formulas assuming infinite accuracy of
computation; and that there is no substitute for getting one’s hands
dirty.


Read this topic online here:
https://forum.vassalengine.org/t/for-the-mathematically-sophisticated/4282/3