<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
		>
<channel>
	<title>Comments on: APR: Domain Model &#8211; UML Class Diagram</title>
	<atom:link href="http://tynerblain.com/blog/2007/04/26/apr-uml-class-diagram-1/feed/" rel="self" type="application/rss+xml" />
	<link>http://tynerblain.com/blog/2007/04/26/apr-uml-class-diagram-1/</link>
	<description>Software product success.</description>
	<lastBuildDate>Tue, 22 May 2012 20:46:27 +0000</lastBuildDate>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>http://wordpress.org/?v=3.0.4</generator>
	<item>
		<title>By: Scott Sehlhorst</title>
		<link>http://tynerblain.com/blog/2007/04/26/apr-uml-class-diagram-1/comment-page-1/#comment-92145</link>
		<dc:creator>Scott Sehlhorst</dc:creator>
		<pubDate>Sat, 28 Apr 2007 14:11:15 +0000</pubDate>
		<guid isPermaLink="false">http://tynerblain.com/blog/2007/04/26/apr-uml-class-diagram-1/#comment-92145</guid>
		<description>Thanks Roger!

Yeah, it is a little implementationish.  Probably because I&#039;m going to be doing the implementing, so adding some extra stuff/thoughts into it &quot;while I&#039;m there&quot; is a risk.  

Thanks for taking the time to look at the diagram in more detail, and reply.</description>
		<content:encoded><![CDATA[<p>Thanks Roger!</p>
<p>Yeah, it is a little implementationish.  Probably because I&#8217;m going to be doing the implementing, so adding some extra stuff/thoughts into it &#8220;while I&#8217;m there&#8221; is a risk.  </p>
<p>Thanks for taking the time to look at the diagram in more detail, and reply.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Roger L. Cauvin</title>
		<link>http://tynerblain.com/blog/2007/04/26/apr-uml-class-diagram-1/comment-page-1/#comment-92139</link>
		<dc:creator>Roger L. Cauvin</dc:creator>
		<pubDate>Sat, 28 Apr 2007 13:42:44 +0000</pubDate>
		<guid isPermaLink="false">http://tynerblain.com/blog/2007/04/26/apr-uml-class-diagram-1/#comment-92139</guid>
		<description>Scott, you asked for a specific example.  I actually hadn&#039;t studied the diagram closely enough to realize you were &quot;double documenting&quot; the concepts as attributes.  Overall, the diagram feels a bit &quot;implementationish&quot;, but I don&#039;t really have an example.  Maybe it&#039;s just that attributes feel like implementation to me.  The more I study the diagram, though, the better it looks to me.</description>
		<content:encoded><![CDATA[<p>Scott, you asked for a specific example.  I actually hadn&#8217;t studied the diagram closely enough to realize you were &#8220;double documenting&#8221; the concepts as attributes.  Overall, the diagram feels a bit &#8220;implementationish&#8221;, but I don&#8217;t really have an example.  Maybe it&#8217;s just that attributes feel like implementation to me.  The more I study the diagram, though, the better it looks to me.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Rolf</title>
		<link>http://tynerblain.com/blog/2007/04/26/apr-uml-class-diagram-1/comment-page-1/#comment-91961</link>
		<dc:creator>Rolf</dc:creator>
		<pubDate>Sat, 28 Apr 2007 05:38:51 +0000</pubDate>
		<guid isPermaLink="false">http://tynerblain.com/blog/2007/04/26/apr-uml-class-diagram-1/#comment-91961</guid>
		<description>Contact-rules: I see. Good thing.</description>
		<content:encoded><![CDATA[<p>Contact-rules: I see. Good thing.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Scott Sehlhorst</title>
		<link>http://tynerblain.com/blog/2007/04/26/apr-uml-class-diagram-1/comment-page-1/#comment-91626</link>
		<dc:creator>Scott Sehlhorst</dc:creator>
		<pubDate>Fri, 27 Apr 2007 15:41:26 +0000</pubDate>
		<guid isPermaLink="false">http://tynerblain.com/blog/2007/04/26/apr-uml-class-diagram-1/#comment-91626</guid>
		<description>Hey Rolf, thanks again for the tons of input!

Edit/resubmit: I think the rule we need to use is &quot;only if the URL changes, should the author resubmit.&quot;  I would encourage Brian (the Blogger persona) to add a review of an existing article when he edits it.

Link-last-checked.  Cool idea.  I put it in &quot;surprise and delight&quot;, but it&#039;s on the stack.

login - I use, and would love to support openID user authentication.  Will have to add that.  Not sure what&#039;s already been done with Rails to support it.  openID allows exactly what you suggest.

Contact-rules:  From a community standpoint, people may want to send private emails to each other.  But they won&#039;t want to share their email addresses publicly on the site.  I think we should be able to facilitate sending emails opaquely from one user to another.  Would likely be a minor extension to the support we need for Paul&#039;s (the product manager persona) broadcast of articles to non-users by email.  But if people anticipate getting spammed with emails from &quot;non-qualified&quot; other users, they may use a secondary email address (like spam@hotmail.com) and never see the &quot;real&quot; emails.  Or even worse, not sign up.  This was a design attempt to address an implicit requirement for community participation.

Rating of reviews - got it.

Thanks again!</description>
		<content:encoded><![CDATA[<p>Hey Rolf, thanks again for the tons of input!</p>
<p>Edit/resubmit: I think the rule we need to use is &#8220;only if the URL changes, should the author resubmit.&#8221;  I would encourage Brian (the Blogger persona) to add a review of an existing article when he edits it.</p>
<p>Link-last-checked.  Cool idea.  I put it in &#8220;surprise and delight&#8221;, but it&#8217;s on the stack.</p>
<p>login &#8211; I use, and would love to support openID user authentication.  Will have to add that.  Not sure what&#8217;s already been done with Rails to support it.  openID allows exactly what you suggest.</p>
<p>Contact-rules:  From a community standpoint, people may want to send private emails to each other.  But they won&#8217;t want to share their email addresses publicly on the site.  I think we should be able to facilitate sending emails opaquely from one user to another.  Would likely be a minor extension to the support we need for Paul&#8217;s (the product manager persona) broadcast of articles to non-users by email.  But if people anticipate getting spammed with emails from &#8220;non-qualified&#8221; other users, they may use a secondary email address (like <a href="mailto:spam@hotmail.com">spam@hotmail.com</a>) and never see the &#8220;real&#8221; emails.  Or even worse, not sign up.  This was a design attempt to address an implicit requirement for community participation.</p>
<p>Rating of reviews &#8211; got it.</p>
<p>Thanks again!</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Scott Sehlhorst</title>
		<link>http://tynerblain.com/blog/2007/04/26/apr-uml-class-diagram-1/comment-page-1/#comment-91605</link>
		<dc:creator>Scott Sehlhorst</dc:creator>
		<pubDate>Fri, 27 Apr 2007 15:03:21 +0000</pubDate>
		<guid isPermaLink="false">http://tynerblain.com/blog/2007/04/26/apr-uml-class-diagram-1/#comment-91605</guid>
		<description>Roger - I forgot to ask - is there a specific example in this diagram where I could have represented something better?  That would help me understand and improve more quickly.

TIA</description>
		<content:encoded><![CDATA[<p>Roger &#8211; I forgot to ask &#8211; is there a specific example in this diagram where I could have represented something better?  That would help me understand and improve more quickly.</p>
<p>TIA</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Scott Sehlhorst</title>
		<link>http://tynerblain.com/blog/2007/04/26/apr-uml-class-diagram-1/comment-page-1/#comment-91597</link>
		<dc:creator>Scott Sehlhorst</dc:creator>
		<pubDate>Fri, 27 Apr 2007 14:47:23 +0000</pubDate>
		<guid isPermaLink="false">http://tynerblain.com/blog/2007/04/26/apr-uml-class-diagram-1/#comment-91597</guid>
		<description>Hey Roger,

I take your comment as a great suggestion, not a &quot;form of documentation&quot; critique!  You&#039;re making a suggestion about representation of the domain model - not description.  I&#039;ll have to check out Larman&#039;s book, but I was generally trying to follow exactly that precept.

Example: Rating is a first class object, not just an attribute.  The enumerations are another manifestation of this.  I realize the blog-documentation approach makes it harder to consume (because I included the easily scannable attribute-placeholders).

With the diagram, I tried to show all of the relationships as concepts - including multiplicity and navigability intentionally.  As a habit in previous OOA/OOD efforts, I started &quot;double documenting&quot; each relationship as an attribute name (userRatings is an example) to make the document easier to scan over time - reminding me that there is a relationship there.

Maybe Larman has a better approach to describing the concepts, I&#039;ll have to check.  Thanks again for the suggestion - and I will definitely turn up the volume on that as I move forward.

And thanks again for your sustained readership and multiple contributions to discussions basically since the beginning of the blog.  I truly appreciate it.</description>
		<content:encoded><![CDATA[<p>Hey Roger,</p>
<p>I take your comment as a great suggestion, not a &#8220;form of documentation&#8221; critique!  You&#8217;re making a suggestion about representation of the domain model &#8211; not description.  I&#8217;ll have to check out Larman&#8217;s book, but I was generally trying to follow exactly that precept.</p>
<p>Example: Rating is a first class object, not just an attribute.  The enumerations are another manifestation of this.  I realize the blog-documentation approach makes it harder to consume (because I included the easily scannable attribute-placeholders).</p>
<p>With the diagram, I tried to show all of the relationships as concepts &#8211; including multiplicity and navigability intentionally.  As a habit in previous OOA/OOD efforts, I started &#8220;double documenting&#8221; each relationship as an attribute name (userRatings is an example) to make the document easier to scan over time &#8211; reminding me that there is a relationship there.</p>
<p>Maybe Larman has a better approach to describing the concepts, I&#8217;ll have to check.  Thanks again for the suggestion &#8211; and I will definitely turn up the volume on that as I move forward.</p>
<p>And thanks again for your sustained readership and multiple contributions to discussions basically since the beginning of the blog.  I truly appreciate it.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Rolf</title>
		<link>http://tynerblain.com/blog/2007/04/26/apr-uml-class-diagram-1/comment-page-1/#comment-91480</link>
		<dc:creator>Rolf</dc:creator>
		<pubDate>Fri, 27 Apr 2007 09:49:44 +0000</pubDate>
		<guid isPermaLink="false">http://tynerblain.com/blog/2007/04/26/apr-uml-class-diagram-1/#comment-91480</guid>
		<description>A submitted (rated, reviewed) article could be changed by its author, hopefully turning it into an even better article. Should the author resubmit it? Leave a comment?

How about a link-last-checked timestamp? Checking can be done automatically, I guess.

reviewTitle and reviewBody:
It should be possible to fill in the Title only (for short comments)

User metadata: Isn&#039;t there a technology that relieves me of entering my data in dozends of sites?

Should user roles be private? No.

Is there a reason to have a distinction between reviewers and submitters? No.

Any other contact rules? I can&#039;t see the point in having contact rules. Please explain.

Should people be able to “Rate” a review? Yes!
 
Threaded commenting? I think rating reviews is enough.</description>
		<content:encoded><![CDATA[<p>A submitted (rated, reviewed) article could be changed by its author, hopefully turning it into an even better article. Should the author resubmit it? Leave a comment?</p>
<p>How about a link-last-checked timestamp? Checking can be done automatically, I guess.</p>
<p>reviewTitle and reviewBody:<br />
It should be possible to fill in the Title only (for short comments)</p>
<p>User metadata: Isn&#8217;t there a technology that relieves me of entering my data in dozends of sites?</p>
<p>Should user roles be private? No.</p>
<p>Is there a reason to have a distinction between reviewers and submitters? No.</p>
<p>Any other contact rules? I can&#8217;t see the point in having contact rules. Please explain.</p>
<p>Should people be able to “Rate” a review? Yes!</p>
<p>Threaded commenting? I think rating reviews is enough.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Roger L. Cauvin</title>
		<link>http://tynerblain.com/blog/2007/04/26/apr-uml-class-diagram-1/comment-page-1/#comment-91269</link>
		<dc:creator>Roger L. Cauvin</dc:creator>
		<pubDate>Fri, 27 Apr 2007 00:12:31 +0000</pubDate>
		<guid isPermaLink="false">http://tynerblain.com/blog/2007/04/26/apr-uml-class-diagram-1/#comment-91269</guid>
		<description>It&#039;s not agile to get hung up on the format of documentation, but I would look into Craig Larman&#039;s thoughts on domain/conceptual models.  In particular, you might err more on the side of including a greater number of full-fledged concepts instead of attributes:

&quot;Perhaps the most common mistake when creating a conceptual model is to represent something as an attribute when it should have been a concept.&quot;

Larman offers a guideline:

&quot;If we do not think of some concept X as a number or text in the real world, X is probably a concept, not an attribute.&quot;

However:

&quot;If in doubt, make it a separate concept.&quot;

This information comes from Larman&#039;s APPLYING UML AND PATTERNS.</description>
		<content:encoded><![CDATA[<p>It&#8217;s not agile to get hung up on the format of documentation, but I would look into Craig Larman&#8217;s thoughts on domain/conceptual models.  In particular, you might err more on the side of including a greater number of full-fledged concepts instead of attributes:</p>
<p>&#8220;Perhaps the most common mistake when creating a conceptual model is to represent something as an attribute when it should have been a concept.&#8221;</p>
<p>Larman offers a guideline:</p>
<p>&#8220;If we do not think of some concept X as a number or text in the real world, X is probably a concept, not an attribute.&#8221;</p>
<p>However:</p>
<p>&#8220;If in doubt, make it a separate concept.&#8221;</p>
<p>This information comes from Larman&#8217;s APPLYING UML AND PATTERNS.</p>
]]></content:encoded>
	</item>
</channel>
</rss>

