Talk:CouchDB

From Wikipedia, the free encyclopedia

references needed here

• theoretical basis

StephenDeGabrielle (talk) 13:53, 2 January 2008 (UTC)

[edit] hard to read

this main paragraph is IMHO hard to read;

Like other Document-oriented database systems, such as Lotus Notes, CouchDB is not a Relational Database Management System. Instead of storing data in rows and columns, the database manages a collection of JSON documents (early versions of CouchDB used XML). As a result, it is often compared with column-oriented datastores like Google's BigTable; however, CouchDB is not a column-oriented store, since the documents in a collection need not share a schema. Aggregate functions and filters are computed in parallel as in MapReduce, rather than at query time (although queries may introduce new aggregate functions, which execute on the existing documents in parallel).

I'd propose the following;

CouchBB is a Document-oriented database system like Lotus Notes, with It manages a collection of documents (as JSON though early versions of CouchDB used XML). Aggregate functions and filters are computed in parallel as in MapReduce, rather than at query time. Queries may introduce new aggregate functions, which execute on the existing documents in parallel. What it is not CouchDB is not a column-oriented store,like Google's BigTable, since the documents in a collection need not share a schema. CouchDB is not a column-oriented store or Relational Database Management System.

I'm not happy about the Map-Reduce bit; is the CouchDB version a reimplementation of Googles original, like Hadoop?

StephenDeGabrielle (talk) 14:47, 2 January 2008 (UTC)