First comes relevance, then correctness, performance, and elegance in some
order. Not every problem is put right. For example. Question: from a long list
of numbers, how to compute the 15th, 50th and 85th percentile? Answer: Don't do
that! Compute the mean value and the standard deviation
On Thu, Sep 23, 2010 at 6:04 PM, Don Guinn dongu...@gmail.com wrote:
Yes. There are ways. I did something like that a few years ago. I extended
Ctrl-F1. If it didn't find a name in the help file then it would search for
the name in locales. Then issue and edit of the name in that locale. If it
Bill,
Thank you very much for your response.
Myodbc was too buggy for real-life use the last time I checked a few years
ago, and browsing its support forum suggested the odbc driver was
un-maintained. Has it been improved?
I do not know if the quality of MyODBC has improved. I've never used
Chris,
I included a fragment from odbc.ijs for your reference. As you can see
date and time (91 92?) are not supported. Furthermore blob data type (eg.
_1 and _4) will never be supported in ddfch, as it is impossible to
allocate (2G or more)*rowsize of memory buffer for each blob column.
NB.
On Fri, Sep 24, 2010 at 10:33 AM, bill lam bbill@gmail.com wrote:
I included a fragment from odbc.ijs for your reference. As you can see
date and time (91 92?) are not supported. Furthermore blob data type (eg.
_1 and _4) will never be supported in ddfch, as it is impossible to
allocate
About Don't do that! modern statistics texts recommend the five-number summary
Minimum Q1 M Q3 Maximum
as a reasonably complete description of center and spread. M Q1 Q3 are the
median and quartiles, see David S. Moore and George P. McCabe, Introduction to
the Practice of Statistics, p. 42,
Thanks to all who responded to this thread. I learned more than I
expected, including the fact that I have to make greater use of key (/.).
Best wishes,
John
--
For information about J forums see
http://www.jsoftware.com/jwiki/Essays/Order%20Statistics
Incidentally, the solutions to this problem are an instance
of the elegant vs. practical discussion more finely balanced
than the dyadic index-of example. There is a short and
elegant solution which is not extravagantly more costly than
Yes, the don't-do-that advice is controversial. Computing mean value and
standard deviation is elegant
E
+/ % #
ES
E , E.:*:(- E)
ES +/? 100 100$2
50.02 5.12636
--- Den fre 24/9/10 skrev Kip Murray k...@math.uh.edu:
Fra: Kip Murray k...@math.uh.edu
Emne: Re: [Jprogramming]
Bill,
Thanks again.
Very helpful and educational.
I now notice that another post previously addressed similar issues
(unsupported data types).
http://jsoftware.com/pipermail/programming/2010-September/020273.html
Perhaps it would save you a little explanation time in the future if
something
The median of list 0 1 2 3 4 is 2 (the middle element), and the median of 0 1 2
3 is 1.5 (average of the two middle elements). Therefore the definition of
median given in Phrases 10C. Math and Stats is incorrect:
median =: @-:@# { /:~
median 0 1 2 3 4
2
median 0 1 2 3
2
So,
midpt=: -:@:@#
median=: -:@(+/)@((. , .)@midpt { /:~)
This is taken from univariate.ijs in addons/stats/base
So, write a correct definition for median! --Kip Murray
--
(B=)
--
For information about J forums see
median=:[:-:[:+/]{~[:(.,.)[:-:[::#
On Fri, Sep 24, 2010 at 9:33 PM, Kip Murray k...@math.uh.edu wrote:
The median of list 0 1 2 3 4 is 2 (the middle element), and the median of 0
1 2
3 is 1.5 (average of the two middle elements). Therefore the definition of
median given in Phrases 10C. Math
median 3 5 5 3 0 NB. wrong
5
median 0 3 3 5 5 NB. right
3
Follow my advice: don't write a correct definition for median! The median is a
discontinuous function. The fix for even number of numbers solves no problem.
Don't support bad design with good programming!
--- Den lør 25/9/10 skrev
14 matches
Mail list logo
