Talk:Adaptive Huffman coding: Difference between revisions
Appearance
Content deleted Content added
Line 15: | Line 15: | ||
See also http://www.cs.duke.edu/csed/curious/compression/adaptivehuff.html |
See also http://www.cs.duke.edu/csed/curious/compression/adaptivehuff.html |
||
I have not made the change back to FGK, but for those who wonder, the algorithm is indeed an erroneous description of FGK algorithm (the error is mostly in description of the way nodes are numbered: "Numbers go down, and from right to left.") The paper mentioned by Vitter in the reference section makes this abundantly clear and there is no need to "search" the web. The reason I have not made the change is that this article needs to be written from ground up. [[Special:Contributions/24.1.53.152|24.1.53.152]] ([[User talk:24.1.53.152|talk]]) 07:16, 9 December 2009 (UTC) |
|||
== Swap == |
== Swap == |
Revision as of 07:57, 9 December 2009
Computing Start‑class Mid‑importance | ||||||||||
|
Hi, everyone.
I put up this article because, IMHO, an adaptive technique deserves a walktrough. But I'm in doubt, maybe just erase it and give a bunch of links to good sites, even merge to Huffman.
And that thing with numbers seems to be just a heap structure trick, maybe even propritary. Highest number == Topmost right.
Gnomz007 17:49, 11 Oct 2004 (UTC)
FGK -> Vitter
The algorithme described is the Vitter alg., not the FGK alg..
See also http://www.cs.duke.edu/csed/curious/compression/adaptivehuff.html
Swap
I don't understand one thing: should the nodes be swapped together with its subtree? I guess it is so so that the weights correspond and so that the tree can change its shape. Yet, I may have missed something. --Honnza 08:25, 14 July 2006 (UTC)