﻿<?xml version="1.0" encoding="utf-8"?><rss version="2.0"><channel><title>Ayende @ Rahien</title><link>http://ayende.com</link><description>Ayende @ Rahien</description><copyright>Copyright (C) Ayende Rahien  2004 - 2021 (c) 2026</copyright><ttl>60</ttl><item><title>Ayende Rahien commented on Designing a document database: Scale</title><description>configurator,
  
Yes, it is pretty easy to create a diff format to json.
  
That is not the problem, I don't want to write one.
</description><link>http://ayende.com/3899/designing-a-document-database-scale#comment10</link><guid>http://ayende.com/3899/designing-a-document-database-scale#comment10</guid><pubDate>Wed, 11 Mar 2009 05:35:15 GMT</pubDate></item><item><title>configurator commented on Designing a document database: Scale</title><description>"How do you know that a removed field is not really removed?"
  
  
The format for partial updates should send a json change object and a list of fields that should be removed.
  
  
"There is no standard way of diffing json"
  
I understand that - I was just saying that it should be relatively trivial to implement as far as I understand json (and I don't, really).
</description><link>http://ayende.com/3899/designing-a-document-database-scale#comment9</link><guid>http://ayende.com/3899/designing-a-document-database-scale#comment9</guid><pubDate>Wed, 11 Mar 2009 04:32:32 GMT</pubDate></item><item><title>Ayende Rahien commented on Designing a document database: Scale</title><description>A field can be removed for many reasons.
  
For instance, because its value is missing.
  
If I am planning to support generic Json docs, then it is pretty obvious that this is a requirement.
</description><link>http://ayende.com/3899/designing-a-document-database-scale#comment8</link><guid>http://ayende.com/3899/designing-a-document-database-scale#comment8</guid><pubDate>Wed, 11 Mar 2009 03:22:56 GMT</pubDate></item><item><title>Ayende Rahien commented on Designing a document database: Scale</title><description>configurator,
  
There is no standard way of diffing json
</description><link>http://ayende.com/3899/designing-a-document-database-scale#comment7</link><guid>http://ayende.com/3899/designing-a-document-database-scale#comment7</guid><pubDate>Wed, 11 Mar 2009 03:03:53 GMT</pubDate></item><item><title>Thomas Eyde commented on Designing a document database: Scale</title><description>I don't know why a field should be removed. If that is part of your spec, I guess I would use a special json document just for that. 
  
  
My thought was that json, and javascript, is dynamic. So we could take advantage of that feature to build up documents which only have changed fields.
</description><link>http://ayende.com/3899/designing-a-document-database-scale#comment6</link><guid>http://ayende.com/3899/designing-a-document-database-scale#comment6</guid><pubDate>Tue, 10 Mar 2009 23:39:54 GMT</pubDate></item><item><title>configurator commented on Designing a document database: Scale</title><description>json has a simple enough syntax to diff, doesn't it? You can either diff it per-field or normalize the string's whitespace to easily string-diff it. Where's the hard part with diffing json?
</description><link>http://ayende.com/3899/designing-a-document-database-scale#comment5</link><guid>http://ayende.com/3899/designing-a-document-database-scale#comment5</guid><pubDate>Tue, 10 Mar 2009 17:10:30 GMT</pubDate></item><item><title>Ayende Rahien commented on Designing a document database: Scale</title><description>Thomas
  
How do you know that a removed field is not really removed?
</description><link>http://ayende.com/3899/designing-a-document-database-scale#comment4</link><guid>http://ayende.com/3899/designing-a-document-database-scale#comment4</guid><pubDate>Tue, 10 Mar 2009 15:38:38 GMT</pubDate></item><item><title>josh commented on Designing a document database: Scale</title><description>I'm sure you're already thinking this and probably say so in a more recent post, but you can of course use another table to index the documents as is talked about in the FriendFeed article you posted a couple days ago. That index table(s) doesn't have to reside on the same node as the document table; just give you a way to find the ID of the document you are looking for.
</description><link>http://ayende.com/3899/designing-a-document-database-scale#comment3</link><guid>http://ayende.com/3899/designing-a-document-database-scale#comment3</guid><pubDate>Tue, 10 Mar 2009 14:39:18 GMT</pubDate></item><item><title>Thomas Eyde commented on Designing a document database: Scale</title><description>Can't you use a partial json document for partial edits? Just pass along the fields that have changed. Missing fields are not updated.
</description><link>http://ayende.com/3899/designing-a-document-database-scale#comment2</link><guid>http://ayende.com/3899/designing-a-document-database-scale#comment2</guid><pubDate>Tue, 10 Mar 2009 13:58:37 GMT</pubDate></item><item><title>Yitzchok commented on Designing a document database: Scale</title><description>I know that for now you are only supporting JSON but you might also support compressed JSON to save some space and then you will really get into problems with partial updates.
</description><link>http://ayende.com/3899/designing-a-document-database-scale#comment1</link><guid>http://ayende.com/3899/designing-a-document-database-scale#comment1</guid><pubDate>Tue, 10 Mar 2009 05:12:10 GMT</pubDate></item></channel></rss>