<?xml version="1.0"?>
<rss version="2.0">
<channel>

  
<title>Coding the Architecture - Premature Optimisation</title>
<link>http://www.codingthearchitecture.com/2007/06/08/premature_optimisation.html</link>
<description>I&#039;d like to talk to you about a sensitive subject that has been affecting me and, I suspect, affecting many of you as well. It&#039;s embarrassing so I&#039;ll say it first and you can all nod in agreement.  &#034;We&#039;re having problems on my project with being premature ...</description>
<language>en</language>
<managingEditor>Robert Annett</managingEditor>
<lastBuildDate>Sat, 09 Jun 2007 13:35:07 GMT</lastBuildDate>
  
  

  <generator>Pebble (http://pebble.sourceforge.net)</generator>
  <docs>http://backend.userland.com/rss</docs>
  
  
  <item>
    <title>Re: Premature Optimisation</title>
    <link>http://www.codingthearchitecture.com/2007/06/08/premature_optimisation.html#comment1181595581469</link>
    <description>
      Leaving non-functionals until the end of the project ... right ... this is why I&#039;m keen on RUP style &#034;executable reference architectures&#034; that prove the key non-functionals as soon as possible. And it&#039;s also the same reason why I believe even &#034;agile&#034; projects need some form of architecture/architectural spike. I&#039;ve been burnt by non-functional requirements before. ;-)
    </description>
    <author>Simon Brown</author>
    <comments>http://www.codingthearchitecture.com/2007/06/08/premature_optimisation.html#comments</comments>
    <guid isPermaLink="true">http://www.codingthearchitecture.com/2007/06/08/premature_optimisation.html#comment1181595581469</guid>
    <pubDate>Mon, 11 Jun 2007 20:59:41 GMT</pubDate>
  </item>
  
  <item>
    <title>Re: Premature Optimisation</title>
    <link>http://www.codingthearchitecture.com/2007/06/08/premature_optimisation.html#comment1181565512098</link>
    <description>
      &lt;p&gt;
I think I&#039;m with Tom on this - &#034;optimisation&#034; can be a moot point before you&#039;ve even written a line of code!
&lt;/p&gt;
&lt;p&gt;
&#034;Tardy&#034; optimisation is just as big a problem. The key, as you suggest, is probably to test early and test often to ensure you&#039;re neither too early or too late with your optimisation. And not to restrict your testing and optimisation to the build phase.
&lt;/p&gt;

    </description>
    <author>Kevin Seal</author>
    <comments>http://www.codingthearchitecture.com/2007/06/08/premature_optimisation.html#comments</comments>
    <guid isPermaLink="true">http://www.codingthearchitecture.com/2007/06/08/premature_optimisation.html#comment1181565512098</guid>
    <pubDate>Mon, 11 Jun 2007 12:38:32 GMT</pubDate>
  </item>
  
  <item>
    <title>Re: Premature Optimisation</title>
    <link>http://www.codingthearchitecture.com/2007/06/08/premature_optimisation.html#comment1181396107065</link>
    <description>
      &lt;p&gt;It&#039;s funny, in my experience, developers tend to optimise too much. Architects too little... Developers will agonise over whether to use a String or a StringBuffer to save a few milliseconds. Meanwhile the Architect will stick the database on a server in Venezuela...&lt;/p&gt;
&lt;p&gt;Ok, I exaggerate, but I think as System Architects we are often guilty of leaving Non-Functionals to the end and then hoping for the best. The trouble is, clients always want to pay for the functionals and assume that &#039;good practice&#039; will somehow magically take care of the non-functionals for free...&lt;/p&gt;
&lt;p&gt;Tom&lt;/p&gt;
    </description>
    <author>Tom</author>
    <comments>http://www.codingthearchitecture.com/2007/06/08/premature_optimisation.html#comments</comments>
    <guid isPermaLink="true">http://www.codingthearchitecture.com/2007/06/08/premature_optimisation.html#comment1181396107065</guid>
    <pubDate>Sat, 09 Jun 2007 13:35:07 GMT</pubDate>
  </item>
  
  </channel>
</rss>
