<?xml version="1.0" encoding="ISO-8859-1"?>
<rss version="2.0">
  <channel>
    <title>Phritz Homepage :: Forum</title>
    <link>http://phritz.fritz-works.com/</link>
    <description> :: XOOPS Community Bulletin Board</description>
    <lastBuildDate>Tue, 7 Sep 2010 04:54:24 -0400</lastBuildDate>
    <docs>http://backend.userland.com/rss/</docs>
    <generator>CBB 1.16</generator>
    <category>Forum</category>
    <managingEditor>mrfritz379@gmail.com</managingEditor>
    <webMaster>mrfritz379@gmail.com</webMaster>
    <language>en</language>
        <image>
      <title>Phritz Homepage :: Forum</title>
      <url>http://phritz.fritz-works.com/modules/newbb/images/xoopsbb_slogo.png</url>
      <link>http://phritz.fritz-works.com/</link>
      <width>92</width>
      <height>52</height>
    </image>
            <item>
      <title>Re: Phritz kernel todo [by mrfritz379]</title>
      <link>http://phritz.fritz-works.com/modules/newbb/viewtopic.php?topic_id=1&amp;forum=1</link>
      <description>Kernel Development:: Phritz kernel todo&lt;br /&gt;
Here are some more ideas that I&amp;#039;m considering.&lt;br /&gt;&lt;br /&gt;Regarding the View aspect of the framework, I&amp;#039;m considering following the pattern established by most other frameworks by having the controller classes instantiate their own views. As mentioned in my previous post, I think I want to have a single View class that has its own private methods to compile the templates in the appropriate format. I&amp;#039;m also considering, now, having each controller specific View compile the content of that View and returning that View to the front controller (the bootstrap file) via some accessor method to compile the complete template at display time. This would allow for customized handling of the View&amp;#039;s data for each controller, making the framework more flexible. The only issue that I see is the Ajax response, since Ajax merely returns the XML tree that is compiled over the course of the script execution. We can&amp;#039;t really return an entire XML document everytime the view is instructed to fetch a template.&lt;br /&gt;&lt;br /&gt;Also, regarding the View, I&amp;#039;m thinking that there should be a compile() method that can be bound in the Event management. The View class can have a template_body property that can be accessed by filtering methods that would be responsible for performing the pre/post filter functions. The filtering of templates are easy enough to implement now using the register_pre(post)filter methods of the TPl class, but some more dynamic method of filtering may be a good idea (it would certainly help to support the attempts to cut down on the passing of objects around as method parameters or even the calling of static accessor methods such as Phritz::getView()).</description>
      <pubDate>Sat, 30 Sep 2006 07:28:11 -0400</pubDate>
      <guid>http://phritz.fritz-works.com/modules/newbb/viewtopic.php?topic_id=1&amp;forum=1</guid>
    </item>
        <item>
      <title>Phritz events [by mrfritz379]</title>
      <link>http://phritz.fritz-works.com/modules/newbb/viewtopic.php?topic_id=3&amp;forum=1</link>
      <description>Kernel Development:: Phritz events&lt;br /&gt;
Currently, in the source, there&amp;#039;s a candidate for a Phritz event management system. I thought about instituting a typical Observer/Observable interface to handle events, but I think I like the idea of a centralized event managent system that allows objects to bind themselves to other objects and their methods. Of course the observable object needs to send a global notification of the event using the PhritzEvent::notify() method. This allows an element of control on the part of the observable regarding which methods can be bound. &lt;br /&gt;&lt;br /&gt;The current bind() method allows observers to bind themselves to events triggered by specific objects (using a GUID created at the object&amp;#039;s instantiation) or by any object of a given class (including subclasses). The idea is to save a certain amount of passing objects around as parameters as the observer object never has to have &amp;#039;contact&amp;#039; with the observed object until the event is triggered. &lt;br /&gt;&lt;br /&gt;One plan that I have for implementing this interface is to allow the User management system (that hasn&amp;#039;t been created yet) to change display themes according to user preferences while running in the backround. It offers a lot of room for customization of the running application without having to modify the kernel. Loaded modules could interact with any of the kernel objects without having to every really have contact with the kernel objects.</description>
      <pubDate>Wed, 27 Sep 2006 21:44:45 -0400</pubDate>
      <guid>http://phritz.fritz-works.com/modules/newbb/viewtopic.php?topic_id=3&amp;forum=1</guid>
    </item>
      </channel>
</rss>