Jump to content

Talk:Densely packed decimal

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Cryptic Sentence

[edit]

The proper way to parse this sentence isn't obvious to me. Could someone clarify? "For example, 38=12×3+2 decimal digits can be encoded in 12×10+7=127 bits." Sumguysr (talk) 07:34, 23 December 2011 (UTC)[reply]

12 sets of three decimal digits can be encoded using 12 sets of 10 binary bits. The remaining two decimal digits can be encoded using 7 binary bits. I'll add that to the article. mfc (talk) 15:30, 31 December 2011 (UTC)[reply]

Chen-Ho encoding

[edit]

Shouldn't the Chen-Ho encoding for 009 be 110 000 0001? --Wm243 (talk) 12:48, 17 April 2009 (UTC)[reply]

Indeed it should -- well spotted. (Fixed.) mfc (talk) 15:05, 17 April 2009 (UTC)[reply]

Yes, but what is it?

[edit]

The article should explain how the scheme actually works. Superm401 - Talk 15:25, 11 November 2007 (UTC)[reply]

You're probably right. Suggestions? mfc 15:52, 15 November 2007 (UTC)[reply]
[edit]

I also just realized many parts are identical to the original paper (http://speleotrove.com/decimal/DPDecimal.html), which says it's copyrighted by IBM. Superm401 - Talk 15:34, 11 November 2007 (UTC)[reply]

See the history of the page; I contributed the initial text for this page (on request), and I am also the author of the referenced paper and summary page and so it is my copyright. It is quite likely I used similar or identical words in the two places. IBM has the use of the text, too, as does Wikipedia, but not exclusively. mfc 15:51, 15 November 2007 (UTC)[reply]

Uses for the 24 redundant codes

[edit]

Possible uses for the 24 redundant codes.

  • HIGH-VALUES: Not useful for arithmetic, but useful for sorting and for the initial value of "min".
  • LOW-VALUES: Like HIGH-VALUES, but a better initial value for "max".
  • +INFINITY: Note: HIGH-VALUES > +INFINITY
  • -INFINITY: Note: LOW-VALUES < -INFINITY
  • -0: "Negative zero", useful for meteorologists, is slightly less than 0, i.e., 0 - epsilon
  • +0: Slightly more than 0, i.e., 0 + epsilon. For symmetry with -0. The canonical zero can then be "exactly zero" which will kill the "numberness" of 1/x and ln x, etc.
  • NaN: Not a Number.
  • NaDPD: Not a Densely Packed Decimal. Unicode has demonstrated the value of a "not one of mine" code, hasn't it? It can serve as a separator, for example. 111 111 1 111 would be quite suitable.
  • ROUND-UP: Not so much a value as a flag for the DPD ALU (arithmetic-logic unit).
  • ROUND-DOWN: ditto
  • ROUND-TO-ZERO: ditto
  • ROUND-TO-EVEN: ditto
  • ROUND-TO-ODD: ditto
  • ROUNDING: A query to the PDP ALU to report its rounding mode with one of the codes above. E.g.
    dpd old_rounding = 0 + ROUNDING;
    dpd new_rounding = 0 + ROUND-UP;
111 111 1 111 NaDCD
111 111 1 110 NaN
...

mjk (talk) 22:50, 25 March 2008 (UTC)[reply]

OK, on further thought, it is strange to have negative values for a representation that is predestined for a sign-magnitude implementation. So forget about -INFINITY, +0, -0, and add:
  • EPSILON: which can be used by meteorologists, etc. with a sign, EPSILON replacing +0 and -EPSILON replacing -0.
  • CR: to explicitly indicate non-negatives, or "credits".
  • DB: to explicitly indicate negatives, or "debits".
mjk (talk) 13:03, 30 April 2008 (UTC)[reply]

Most applications won't use DPD for anything special beyond encoding decimal numbers. 24691358r (talk) 13:23, 20 May 2017 (UTC)[reply]

Let alone all of your suggested codes. 202.79.134.100 (talk) 08:42, 13 March 2018 (UTC)[reply]

Other bases

[edit]

It's possible to have a similar system for other bases, for example storing 5 Ternary numeral system in a byte (3⁵=243 on 2⁸=256), as shown below:

000xxxxx — xxxxx
0010xxxx — xxxx2
0011xxxx — xxx2x
0100xxxx — xx2xx
0101xxxx — x2xxx
0110xxxx — 2xxxx
01110xxx — xxx22
01111xxx — xx2x2
10000xxx — xx22x
10001xxx — x2xx2
10010xxx — x2x2x
10011xxx — x22xx
10100xxx — 2xxx2
10101xxx — 2xx2x
10110xxx — 2x2xx
10111xxx — 22xxx
110000xx — xx222
110001xx — x2x22
110010xx — x22x2
110011xx — x222x
110100xx — 2xx22
110101xx — 2x2x2
110110xx — 2x22x
110111xx — 22xx2
111000xx — 22x2x
111001xx — 222xx
1110100x — x2222
1110101x — 2x222
1110110x — 22x22
1110111x — 222x2
1111000x — 2222x
11110010 — 22222

where x means 0 or 1. Uses are for example densely storing an uncompressed 3–color image. 202.79.134.100 (talk) 08:42, 13 March 2018 (UTC)[reply]

Improve table for clarity

[edit]

I don't know how to edit the table

Densely packed decimal encoding rules[1]
DPD encoded value Decimal digits
Code space
(1024 states)
b9 b8 b7 b6 b5 b4 b3 b2 b1 b0 d2 d1 d0 Values encoded Description Occurrences
(1000 states)
50.0%
(512 states)
a b c d e f 0 g h i 0abc 0def 0ghi (0–7) (0–7) (0–7) 3 small digits 51.2%
(512 states)
37.5%
(384 states)
a b c d e f 1 0 0 i 0abc 0def 100i (0–7) (0–7) (8–9) 2 small digits,
1 large digit
38.4%
(384 states)
a b c g h f 1 0 1 i 0abc 100f 0ghi (0–7) (8–9) (0–7)
g h c d e f 1 1 0 i 100c 0def 0ghi (8–9) (0–7) (0–7)
9.375%
(96 states)
g h c 0 0 f 1 1 1 i 100c 100f 0ghi (8–9) (8–9) (0–7) 1 small digit,
2 large digits
9.6%
(96 states)
d e c 0 1 f 1 1 1 i 100c 0def 100i (8–9) (0–7) (8–9)
a b c 1 0 f 1 1 1 i 0abc 100f 100i (0–7) (8–9) (8–9)
3.125%
(32 states, 8 used)
x x c 1 1 f 1 1 1 i 100c 100f 100i (8–9) (8–9) (8–9) 3 large digits,
b9, b8: don't care
0.8%
(8 states)

That table could be more easy to read if the last column had the observation that a b3=0, b3=1 and b2b1<11 for just one 9-8 digit, b3=1 and b2b1=11 and b6b5<11 for two 9-8 digits, and b3=1 and b2b1=11 and b6b5=11 for three 9-8 digits.

something like:

or

not all the lengthly text, maybe just the antecedent of the implication is needed.

Colors may also help. Is just a matter to do the changes to see what is visually more clear and aesthetic.

References

  1. ^ Cowlishaw, Michael Frederic (2007-02-13) [2000-10-03]. "A Summary of Densely Packed Decimal encoding". IBM. Archived from the original on 2015-09-24. Retrieved 2016-02-07.

-- (unsigned) 2018-09-15T10:56:20‎ 2806:107e:c:f09:218:deff:fe2b:121