Comments on: A project I don’t want to do http://enfranchisedmind.com/blog/2007/05/29/a-project-i-dont-want-to-do/ Robert Fischer and Brian Hurt on Punditry, Programming Languages, and Other Religious Issues Thu, 20 Nov 2008 17:27:28 +0000 http://wordpress.org/?v=2.6.3 By: KWB http://enfranchisedmind.com/blog/2007/05/29/a-project-i-dont-want-to-do/#comment-32760 KWB Fri, 07 Mar 2008 17:20:31 +0000 http://enfranchisedmind.com/blog/archive/2007/05/29/231#comment-32760 I haven't used it yet, but Hypertable is modeled on Google's BigTable and may come close to what you're talking about: http://hypertable.org/index.html I haven’t used it yet, but Hypertable is modeled on Google’s BigTable and may come close to what you’re talking about:

http://hypertable.org/index.html

]]> By: Enfranchised Mind http://enfranchisedmind.com/blog/2007/05/29/a-project-i-dont-want-to-do/#comment-31706 Enfranchised Mind Mon, 24 Sep 2007 16:58:21 +0000 http://enfranchisedmind.com/blog/archive/2007/05/29/231#comment-31706 <strong>CouchDB — The Project Brian Didn’t Want to Do...</strong> CouchDB seems to be the implementation of the Project Brian Doesn’t Want To Do. Popularity: unranked [?]...... CouchDB — The Project Brian Didn’t Want to Do…

CouchDB seems to be the implementation of the Project Brian Doesn’t Want To Do.
Popularity: unranked [?]……

]]> By: Robert http://enfranchisedmind.com/blog/2007/05/29/a-project-i-dont-want-to-do/#comment-31063 Robert Thu, 07 Jun 2007 12:14:19 +0000 http://enfranchisedmind.com/blog/archive/2007/05/29/231#comment-31063 <blockquote>But once I start doing more than that, things start getting complex. Say I want to be able to view all the comments made by a given commenter. Or say I want some sort of index of what videos are about. Welcome to complexity, and it only gets worse from here. And, if you’re not using the relational part of the RDBMS, it’s performance is going to suck.</blockquote> I'm not so sure that's true. Check out <a href="http://lucene.apache.org/java/docs/benchmarks.html" rel="nofollow">Lucene</a>: while it's not blazing fast, it's hard to sell it as "sucking".

But once I start doing more than that, things start getting complex. Say I want to be able to view all the comments made by a given commenter. Or say I want some sort of index of what videos are about. Welcome to complexity, and it only gets worse from here.

And, if you’re not using the relational part of the RDBMS, it’s performance is going to suck.

I’m not so sure that’s true. Check out Lucene: while it’s not blazing fast, it’s hard to sell it as “sucking”.

]]> By: Brian http://enfranchisedmind.com/blog/2007/05/29/a-project-i-dont-want-to-do/#comment-31061 Brian Wed, 06 Jun 2007 21:10:26 +0000 http://enfranchisedmind.com/blog/archive/2007/05/29/231#comment-31061 Let me take one quote from that article and rip it to shreds: <BLOCKQUOTE>Why does YouTube need an RDBMS? It serves a file that people can comment on.</BLOCKQUOTE> If that's all the website does, then no, a RDBMS is probably not needed. But my second law of programming is that the complexity of a program increases over time (my first law is that kludges multiply). But once I start doing more than that, things start getting complex. Say I want to be able to view all the comments made by a given commenter. Or say I want some sort of index of what videos are about. Welcome to complexity, and it only gets worse from here. And, if you're not using the relational part of the RDBMS, it's performance is going to suck. Let me take one quote from that article and rip it to shreds:

Why does YouTube need an RDBMS? It serves a file that people can comment on.

If that’s all the website does, then no, a RDBMS is probably not needed. But my second law of programming is that the complexity of a program increases over time (my first law is that kludges multiply).

But once I start doing more than that, things start getting complex. Say I want to be able to view all the comments made by a given commenter. Or say I want some sort of index of what videos are about. Welcome to complexity, and it only gets worse from here.

And, if you’re not using the relational part of the RDBMS, it’s performance is going to suck.

]]> By: Andre Pang http://enfranchisedmind.com/blog/2007/05/29/a-project-i-dont-want-to-do/#comment-31052 Andre Pang Wed, 06 Jun 2007 14:56:36 +0000 http://enfranchisedmind.com/blog/archive/2007/05/29/231#comment-31052 Sounds a lot like Apple's <a href="http://en.wikipedia.org/wiki/Core_Data" rel="nofollow">Core Data</a>, although Core Data is Cocoa-only, and thus has a limited audience. It's the bees knees for Mac programmers, though! Sounds a lot like Apple’s Core Data, although Core Data is Cocoa-only, and thus has a limited audience. It’s the bees knees for Mac programmers, though!

]]> By: Robert http://enfranchisedmind.com/blog/2007/05/29/a-project-i-dont-want-to-do/#comment-31034 Robert Mon, 04 Jun 2007 20:26:28 +0000 http://enfranchisedmind.com/blog/archive/2007/05/29/231#comment-31034 Interestingly, there seems to be an undercurrent of "drop the RDBMS" in the hip Rails circles as a way of increasing performance. <a href="http://blog.nicksieger.com/articles/2007/05/23/geeksessions-i-ruby-on-rails-to-scale-or-not-to-scale" rel="nofollow">Here is an example</a>. Interestingly, there seems to be an undercurrent of “drop the RDBMS” in the hip Rails circles as a way of increasing performance. Here is an example.

]]> By: Sandro Magi http://enfranchisedmind.com/blog/2007/05/29/a-project-i-dont-want-to-do/#comment-31026 Sandro Magi Sat, 02 Jun 2007 16:37:52 +0000 http://enfranchisedmind.com/blog/archive/2007/05/29/231#comment-31026 This "object store" is the approach taken by the web-calculus as implemented in the Waterken server: http://waterken.sourceforge.net/ Bonus: each objects assigned a guid can now be trivially exported over a network as a url. Each database is essentially a singly-threaded, isolated event-loop (like a Vat in E), and you now have a distributed object system instead of a local persistent object system. However, there will always be a need to query some data. This can be solved be creating and persisting indexes of the data of interest, and keeping the index in synch with changes to the data structure. This “object store” is the approach taken by the web-calculus as implemented in the Waterken server:

http://waterken.sourceforge.net/

Bonus: each objects assigned a guid can now be trivially exported over a network as a url. Each database is essentially a singly-threaded, isolated event-loop (like a Vat in E), and you now have a distributed object system instead of a local persistent object system.

However, there will always be a need to query some data. This can be solved be creating and persisting indexes of the data of interest, and keeping the index in synch with changes to the data structure.

]]> By: Robert http://enfranchisedmind.com/blog/2007/05/29/a-project-i-dont-want-to-do/#comment-29562 Robert Thu, 31 May 2007 22:29:47 +0000 http://enfranchisedmind.com/blog/archive/2007/05/29/231#comment-29562 Adam -- Have you open sourced that project from Sharpcast? Nobody in particular -- CouchDb really wants to meet Mercurial. Adam –

Have you open sourced that project from Sharpcast?

Nobody in particular –

CouchDb really wants to meet Mercurial.

]]> By: Brian http://enfranchisedmind.com/blog/2007/05/29/a-project-i-dont-want-to-do/#comment-28664 Brian Wed, 30 May 2007 21:20:35 +0000 http://enfranchisedmind.com/blog/archive/2007/05/29/231#comment-28664 Good. The project exists, and I don't have to build it, so it's not on my conscience. :-) Good. The project exists, and I don’t have to build it, so it’s not on my conscience. :-)

]]> By: Adam Rosien http://enfranchisedmind.com/blog/2007/05/29/a-project-i-dont-want-to-do/#comment-27908 Adam Rosien Wed, 30 May 2007 01:24:48 +0000 http://enfranchisedmind.com/blog/archive/2007/05/29/231#comment-27908 We\'ve built something similar at the company I work for, Sharpcast (<a href=\"http://www.sharpcast.com\" target=\"v\" rel="nofollow">http://www.sharpcast.com</a>), though we have a different take on the synchronization protocol. <a href=\"http://www.couchdb.com\" target=\"v\" rel="nofollow">Couchdb</a> is also similar. We\’ve built something similar at the company I work for, Sharpcast (http://www.sharpcast.com), though we have a different take on the synchronization protocol. Couchdb is also similar.

]]>