Comments on: Rounded Corners – 143 (Kool-Kid-Aid) http://labnotes.org/2007/09/07/rounded-corners-143-kool-kid-aid/ Mon, 26 Jul 2010 17:36:41 +0000 http://wordpress.org/?v=2.9.2 hourly 1 By: raganwald http://labnotes.org/2007/09/07/rounded-corners-143-kool-kid-aid/comment-page-1/#comment-138366 raganwald Sun, 09 Sep 2007 12:06:12 +0000 http://blog.labnotes.org/2007/09/07/rounded-corners-143-kool-kid-aid/#comment-138366 Although several blogs rushed to judgment over the "Congrats, Late Adopters" ad, it is not a hoax, but it also was not placed by Apple. It was placed by someone in the Apple Store Affiliate Program. He claims it was a bit of a joke based on the Nokia ad. Either way, Apple kicked him out of the affiliate program and the ad was pulled. http://www.didntyouhear.com/2007/09/07/dyh-isnt-funny Although several blogs rushed to judgment over the “Congrats, Late Adopters” ad, it is not a hoax, but it also was not placed by Apple. It was placed by someone in the Apple Store Affiliate Program. He claims it was a bit of a joke based on the Nokia ad. Either way, Apple kicked him out of the affiliate program and the ad was pulled.

http://www.didntyouhear.com/2007/09/07/dyh-isnt-funny

]]> By: Assaf http://labnotes.org/2007/09/07/rounded-corners-143-kool-kid-aid/comment-page-1/#comment-138354 Assaf Fri, 07 Sep 2007 15:18:17 +0000 http://blog.labnotes.org/2007/09/07/rounded-corners-143-kool-kid-aid/#comment-138354 I use the term Write Consistency Database to talk about any database server that reaches a consistent state at the end up an atomic update or transactions; the end of a write from the perspective of the application. So PostrgreSQL is a Write Consistency Database. A Read Consistency Database doesn't reach a consistent state at the end of an update, but helps you reconstruct a consistent state at the end of a read. Databases like that are much easier to scale and replicate. I use the term Write Consistency Database to talk about any database server that reaches a consistent state at the end up an atomic update or transactions; the end of a write from the perspective of the application. So PostrgreSQL is a Write Consistency Database.

A Read Consistency Database doesn’t reach a consistent state at the end of an update, but helps you reconstruct a consistent state at the end of a read. Databases like that are much easier to scale and replicate.

]]> By: David Roussel http://labnotes.org/2007/09/07/rounded-corners-143-kool-kid-aid/comment-page-1/#comment-138353 David Roussel Fri, 07 Sep 2007 11:36:37 +0000 http://blog.labnotes.org/2007/09/07/rounded-corners-143-kool-kid-aid/#comment-138353 > I’d take ... if that's the first working implementation of a Read Consistency Database. Looks to me that couchDb works in exactly the same way as postgres with regard to MVCC. When you start a transaction in postgres you see a completely static view of the database that never changes. If other changes do occur, they occur on later versions of the rows, and you stay looking at the version that existed at the time you started. You can configure it to go and remove (autovaccuum) old versions of rows that are no longer visible. Subversion works in a similar way, but the old versions are never removed. What is it about CouchDB that seems better to you? > I’d take … if that’s the first working implementation of a Read Consistency Database.

Looks to me that couchDb works in exactly the same way as postgres with regard to MVCC. When you start a transaction in postgres you see a completely static view of the database that never changes. If other changes do occur, they occur on later versions of the rows, and you stay looking at the version that existed at the time you started.
You can configure it to go and remove (autovaccuum) old versions of rows that are no longer visible.

Subversion works in a similar way, but the old versions are never removed.

What is it about CouchDB that seems better to you?

]]>