|
I am doing some interim testing on a personal project to check its accuracy. It has dozens of slowly converging polynomials, and lots of separate complex functions to be integrated into the final output. I am expecting errors, since most polynomials are approximations, getting closer and closer to the truth, without ever getting there.
My accuracy requirement is basically less than 1 part in 100,000, which doesn't sound like much, but to some of the polynomials, that looks more like 1 in 4 billion.
One of the critical components is producing figures well within my requirements - but the error range is very small, and is all positive. One quick trig function, and I could improve the accuracy by an order of magnitude. The trig function, of course, would be completely made up, have no basis in reality, totally outside any possible justification.
It would just make it all work so prettily!
Should I?  No - it would be completely out of order.
Yes, but it's just one very small function.  Not a chance, you fraudster!
But... but... but...  You're fired!
Sigh! 
|
|
|
|
|
I'd do it, but then I run with scissors.
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
Yeah - but I would know...
|
|
|
|
|
Hah! The traditional problem that I should be old enough and wise enough to be sceptical about!
Just because one of my reference works declared something to be a constant, that does not mean I should believe them. It may have been a constant when people were using an abacus to add with, but it ain't now. looks like I have to do some research and then write yet another polynomial.
the first and second sets of errors, moved to an understandable range and rounded:-
1.7 -0.1
1.6 -0.1
1.7 0.0
1.8 0.1
1.9 0.2
2.1 0.3
2.2 0.3
1.7 0.1
1.5 -0.2
1.5 -0.3
1.7 -0.2
2.0 0.1
1.8 0.0
My target accuracy was -3.0 >> +3.0
I took the so called 'constant' and adjusted it by bracketing down to new values, and have learned that although the 'constant' changes very slowly, it does change. I tweaked the digit in the eighth
significate position.
Constants ain't!
[edit]Tabs don't stick![/edit]
|
|
|
|
|
the only constant is change. and I'm sure i'm wrong about that being constant.
we are adrift.
my code doesn't even execute the same way twice.
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
|
I think I'll need all that and some more by the time I finish this.
|
|
|
|
|
Hi,
I've always been interested in learning parsing and compiler design, and especially parser generation just like the stuff you're doing.
I never had the time to sit down and study it. (It's not something I can pick up on the fly.)
You seem to have loads of practical hands-on experience based upon the projects you're describing here.
How did you first learn it? Is that what you did at Microsoft?
Thank you
The difficult we do right away...
...the impossible takes slightly longer.
|
|
|
|
|
I was in your shoes actually, although I had played around with it some when I was young.
At microsoft I did not work on compiler stuff. I wish I did, I was hoping for a spot on the compiler team - even testing - over at MSR (back when Herb Sutter was there) to the point where I punted a separate position at MSR to hold out for it (i didn't get it but i'm glad i tried)
I taught myself. I can teach you.
It's a pain in the backside until you "get it" - then it's still a pain but you can at least navigate.
I can break down LL(1) parsing for you pretty easily though.
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
honey the codewitch wrote: It's a pain in the backside until you "get it"
Oh I really know what you mean. A lot of things are difficult for me until they click. But once they click I'm able to do them well.
What's MSR?
The difficult we do right away...
...the impossible takes slightly longer.
|
|
|
|
|
it's microsoft research. sorry. the lingo still sticks with me. MS is acronym hell. years of them being drilled into i use them without thinking.
have you used parser generators before at all?
if so I can probably teach you the birds-eye of generating an LL(1) parser in 10-20 minutes.
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
Yes, I once experimented with lex and Yacc but I just couldn't find enough real world examples of "scripts" for them. I understood BNF, but not how to use lex and yacc.
The difficult we do right away...
...the impossible takes slightly longer.
|
|
|
|
|
lex and YACC are hard.
if i gave you a simple ruleset that's basically really stripped BNF, you might get it, eh?
How about
E -> E + T | T
T -> T * F | F
F -> ( E ) | int
or longhand
E -> E + T
E -> T
T -> T * F
T -> F
F -> ( E )
F -> int
This represents a very simple expression grammar with positive ints, +, * and () supported
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
Well, I did understand it at the time. Only by now I have forgotten it all. <nervous laugh>
The difficult we do right away...
...the impossible takes slightly longer.
|
|
|
|
|
well that's a variant. I mean, things like XBNF are a lot friendlier but the computer can't do much with them directly.
how about if i pretty it up some and put it in xbnf?
expr= expr "+" term | term;
term= term "*" factor | factor;
factor= "(" expr ")" | int;
it works using substitution. Every time it finds either "term" or "expr" "+" "term" it replaces that with "expr"
(note it's cyclical here, that makes a loop)
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
Ah yes, I see. Each successive line defines the symbol that was declared on the line above it?
The difficult we do right away...
...the impossible takes slightly longer.
|
|
|
|
|
in this case yes, but not necessarily. They can be in any order. They do reference each other though and that's how you figure out what replaces what
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
Back when I was in undergraduate school, lexical analysis and compiler design were all taught in Computer Science. They don't seem to do that much anymore, though.
"One man's wage rise is another man's price increase." - Harold Wilson
"Fireproof doesn't mean the fire will never come. It means when the fire comes that you will be able to withstand it." - Michael Simmons
"You can easily judge the character of a man by how he treats those who can do nothing for him." - James D. Miles
|
|
|
|
|
Ain't that the truth. Now it's only about learning how "to code".
The difficult we do right away...
...the impossible takes slightly longer.
|
|
|
|
|
i never went to school for it. I wound up at microsoft at 18 instead. But compilers and parsers have always been an interest of mine.
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
Isn't the Roslyn compiler open source?
And maybe Honey can contribute 
|
|
|
|
|
A bug in Microsoft's CodeDOM for the VB language forced me to change my code if I wanted to support VB.
Well, I feel like i need a shower, but the next rev will support VB code generation
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
If changing your code so it works with VB feels like you need a shower, just wait until you write actual VB code to test it out.
Bring out the peroxide and pressure washer.
|
|
|
|
|
i have a "friend" i can fob that off on.  poor guy
When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.
|
|
|
|
|
Okayyyy...You've already said more than enough. 
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
