Talk:Crit bit tree

Merge discussion
It has been suggested that Patricia trie, Crit bit tree, and Radix tree be merged.

Google search result count:

"radix tree"   => 22300 "crit bit tree" => 29 "patricia trie" => 4300 "patricia tree" => 4010

Therefore, the 3 articles should be NOT be merged into "crit bit tree" (since the crit bit tree name is almost unknown). DADS redirects radix tree to patricia tree. At the ACM Digital Library, there are 40 results for "patricia trie", 43 for "patricia tree", and 15 for "radix tree". A quick glance through the search results shows that "patricia trie" yields papers relevant to the topic, while "radix tree" does not. This evidence indicates that "patricia trie" is the preferred name in the literature, and "radix tree" is the preferred name outside the literature. I'd like Wikipedia to look professional, so I say that the three articles should be merged into Patricia trie. - Connelly 20:26, 14 August 2005 (UTC)


 * Patricia tree term is also more familiar to me than Radix and 'Crit bit tree'. Nevertheless I feel that Patricia is not the same as Radix, since the former is generalization of Trie, which is not necessarily binary. As for Radix and 'CBT' I feel quite comfortable merging them together. Andrey Ustyuzhanin, 09 October 2005


 * The number of hits a term yields on Google is irrelevant. Furthermore, the way google works, "patricia tree" will match the text "my name is patricia. Tree leaves are nice", so what you're really measuring is the fact that the word "crit" is extremely rare, and a three-word query is less likely to happen "by chance" than a two-word query.  I support the merge; these are three terms for the same concept.  One of the best open-source packages out there for this data structure declares itself to be a "crit bit tree" package, so we can't simply pretend that this term is going to disappear. Megacz 05:07, 23 October 2005 (UTC)
 * Note that there is a difference between typing '"patricia tree"' and 'patricia tree' in Google. --Dirk 12:13, 23 October 2005 (UTC)
 * Yes Dirk, I know that. Even when you put quotes around a term, Google will still match against text where punctuation occurs within the phrase you requested.  This annoys me, but it's how it is. --Megacz 20:14, 10 November 2005 (UTC)
 * Also, my comment wasn't really meant in opposition to the merge; more of a rant on what I see as an (increasingly common) abuse of the google test. But I do feel that there should be an entry for Crit_bit_tree, even if it's just a redirect to a subheading of another article.  I think it would be beneficial for Wikipedia to point out the connection between the term used in peer-reviewed publications and the term used by a leading open-source implementation of the algorithm. Megacz 20:24, 10 November 2005 (UTC)
 * I agree with Connelly. Actually, there is no scientific reference for Crit Bit Tree given in the article. And there is no definition what it precisely is. --Dirk 12:13, 23 October 2005 (UTC)

The terms Patricia tree and radix tree are longstanding and well-known in computing science literature, probably with hundreds of references each. The term "crit bit tree" is really not known in the literature, but it does appear to mean the same as radix tree. Patricia trees are more general than radix/crit bit trees, not necessarily using a binary alphabet. My recommendation is that the main entries Patricia tree and Radix tree remain and that Crit bit tree be a pointer to the Radix tree entry. -- Rob Cameron
 * I agree with Rob. --Dirk

Oops, I'm sorry, I failed to read this discussion before I executed the merge. I decided that although they're not exactly the same thing, they're more than similar enough to discuss in the same article with some short discussion of the differences, so I did. I actually favoured the name "Patricia trie" since it was the most familiar to me, but I went by the Google test in the end. Please put future discussion on Talk: Radix tree. Thanks. Deco 01:36, 9 December 2005 (UTC)