<?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"
	>
<channel>
	<title>Comments on: Why I don&#8217;t like Fit</title>
	<atom:link href="http://marcus.ahnve.net/2006/08/10/why-i-dont-like-fit/feed/" rel="self" type="application/rss+xml" />
	<link>http://marcus.ahnve.net/2006/08/10/why-i-dont-like-fit/</link>
	<description>A northern, icebears-on-the-streets, view on programming, music etc. from Marcus Ahnve</description>
	<pubDate>Fri, 21 Nov 2008 02:53:34 +0000</pubDate>
	<generator>http://wordpress.org/?v=2.6</generator>
		<item>
		<title>By: Geoff Bache</title>
		<link>http://marcus.ahnve.net/2006/08/10/why-i-dont-like-fit/#comment-2413</link>
		<dc:creator>Geoff Bache</dc:creator>
		<pubDate>Mon, 02 Oct 2006 20:01:26 +0000</pubDate>
		<guid isPermaLink="false">http://marcus.ahnve.net/2006/07/29/why-i-dont-like-fit/#comment-2413</guid>
		<description>&lt;p&gt;It's long seemed obvious to me that tables are really great for ColumnFixture-style tests and mostly get in the way for more workflow-style tests. Your point about version control is a good one too - pretty essential for tests that are going to live a long time.&lt;/p&gt;

&lt;p&gt;Another recommendation: check out Exactor (I think I also mentioned this at XP2006). It's a lot like Fit but without the compulsory tables and based much more around a plain text format. It's a sourceforge project too, run by the guys at Exoftware. It has a much smaller user base though, of course.&lt;/p&gt;

&lt;p&gt;Then there's my own TextTest, but that works a little differently to Fitlibrary and Exactor - though it's also grown up around the same desire
to provide domain language tests for customers on agile projects.&lt;/p&gt;
</description>
		<content:encoded><![CDATA[<p>It&#8217;s long seemed obvious to me that tables are really great for ColumnFixture-style tests and mostly get in the way for more workflow-style tests. Your point about version control is a good one too - pretty essential for tests that are going to live a long time.</p>

<p>Another recommendation: check out Exactor (I think I also mentioned this at XP2006). It&#8217;s a lot like Fit but without the compulsory tables and based much more around a plain text format. It&#8217;s a sourceforge project too, run by the guys at Exoftware. It has a much smaller user base though, of course.</p>

<p>Then there&#8217;s my own TextTest, but that works a little differently to Fitlibrary and Exactor - though it&#8217;s also grown up around the same desire
to provide domain language tests for customers on agile projects.</p>]]></content:encoded>
	</item>
	<item>
		<title>By: Ville Oikarinen</title>
		<link>http://marcus.ahnve.net/2006/08/10/why-i-dont-like-fit/#comment-1820</link>
		<dc:creator>Ville Oikarinen</dc:creator>
		<pubDate>Fri, 18 Aug 2006 12:42:15 +0000</pubDate>
		<guid isPermaLink="false">http://marcus.ahnve.net/2006/07/29/why-i-dont-like-fit/#comment-1820</guid>
		<description>&lt;p&gt;I think the excel format isn't an essential part of Fit and shouldn't be used as an argument against it.&lt;/p&gt;

&lt;p&gt;Whoever implemented the excel reader for Fit caused negative PR for the product, because whenever you give people Office support they infect the whole process or product with the Office "philosophy" :)&lt;/p&gt;

&lt;p&gt;I also agree on the importance of textual formats, but to be honest, excel is quite usable if you save the table as CSV. In fact, CSV as a syntax is very wiki-like so excel is Just Another Editor for it.&lt;/p&gt;

&lt;p&gt;I don't know Fit very well, but IMHO the main philosophy behind it is to use tables as a metalanguage to easily create custom languages for tests. Data in a table is already separated as tokens so the lexer phase isn't needed when parsing the language.&lt;/p&gt;

&lt;p&gt;Of course, as many people have pointed out, tabular syntax isn't suitable for everything. That's why I believe Fit will be replaced by some more general metalanguage that enables the creation of a richer set of syntaxes and maybe also provides some (partly) general purpose language for more complex test semantics.&lt;/p&gt;
</description>
		<content:encoded><![CDATA[<p>I think the excel format isn&#8217;t an essential part of Fit and shouldn&#8217;t be used as an argument against it.</p>

<p>Whoever implemented the excel reader for Fit caused negative PR for the product, because whenever you give people Office support they infect the whole process or product with the Office &#8220;philosophy&#8221; <img src='http://marcus.ahnve.net/wp-includes/images/smilies/icon_smile.gif' alt=':)' class='wp-smiley' /> </p>

<p>I also agree on the importance of textual formats, but to be honest, excel is quite usable if you save the table as CSV. In fact, CSV as a syntax is very wiki-like so excel is Just Another Editor for it.</p>

<p>I don&#8217;t know Fit very well, but IMHO the main philosophy behind it is to use tables as a metalanguage to easily create custom languages for tests. Data in a table is already separated as tokens so the lexer phase isn&#8217;t needed when parsing the language.</p>

<p>Of course, as many people have pointed out, tabular syntax isn&#8217;t suitable for everything. That&#8217;s why I believe Fit will be replaced by some more general metalanguage that enables the creation of a richer set of syntaxes and maybe also provides some (partly) general purpose language for more complex test semantics.</p>]]></content:encoded>
	</item>
	<item>
		<title>By: Marcus</title>
		<link>http://marcus.ahnve.net/2006/08/10/why-i-dont-like-fit/#comment-1760</link>
		<dc:creator>Marcus</dc:creator>
		<pubDate>Tue, 15 Aug 2006 08:17:36 +0000</pubDate>
		<guid isPermaLink="false">http://marcus.ahnve.net/2006/07/29/why-i-dont-like-fit/#comment-1760</guid>
		<description>&lt;p&gt;Javid, &lt;/p&gt;

&lt;p&gt;I mostly agree with you. However, I do like the abstraction Fit provides which allows specs to be written independently of the implementation. &lt;/p&gt;

&lt;p&gt;Normally when working with JUnit, such abstractions are extracted &lt;em&gt;after&lt;/em&gt; a lot of tests have been created.&lt;/p&gt;

&lt;p&gt;I am very inclined to see what can be done with Ruby-based DSL instead of the tables.&lt;/p&gt;
</description>
		<content:encoded><![CDATA[<p>Javid, </p>

<p>I mostly agree with you. However, I do like the abstraction Fit provides which allows specs to be written independently of the implementation. </p>

<p>Normally when working with JUnit, such abstractions are extracted <em>after</em> a lot of tests have been created.</p>

<p>I am very inclined to see what can be done with Ruby-based DSL instead of the tables.</p>]]></content:encoded>
	</item>
	<item>
		<title>By: Marcus</title>
		<link>http://marcus.ahnve.net/2006/08/10/why-i-dont-like-fit/#comment-1750</link>
		<dc:creator>Marcus</dc:creator>
		<pubDate>Mon, 14 Aug 2006 11:47:25 +0000</pubDate>
		<guid isPermaLink="false">http://marcus.ahnve.net/2006/07/29/why-i-dont-like-fit/#comment-1750</guid>
		<description>&lt;p&gt;Anjan,&lt;/p&gt;

&lt;p&gt;I haven't looked it up, but Excels XML format will most probably be enclosed in a zipped file, much like OpenDocument. &lt;/p&gt;

&lt;p&gt;And even if it was a plain text XML-file, it would not be readable. Again, look at the XML produced by OpenDocument.&lt;/p&gt;
</description>
		<content:encoded><![CDATA[<p>Anjan,</p>

<p>I haven&#8217;t looked it up, but Excels XML format will most probably be enclosed in a zipped file, much like OpenDocument. </p>

<p>And even if it was a plain text XML-file, it would not be readable. Again, look at the XML produced by OpenDocument.</p>]]></content:encoded>
	</item>
	<item>
		<title>By: anjan bacchu</title>
		<link>http://marcus.ahnve.net/2006/08/10/why-i-dont-like-fit/#comment-1712</link>
		<dc:creator>anjan bacchu</dc:creator>
		<pubDate>Sat, 12 Aug 2006 07:20:39 +0000</pubDate>
		<guid isPermaLink="false">http://marcus.ahnve.net/2006/07/29/why-i-dont-like-fit/#comment-1712</guid>
		<description>&lt;p&gt;hi there,&lt;/p&gt;

&lt;p&gt;regarding the issue of VERSION CONTROL with excel's binary format -- soon, you will be able to save excel files in XML format!!&lt;/p&gt;

&lt;p&gt;BR,
~A&lt;/p&gt;
</description>
		<content:encoded><![CDATA[<p>hi there,</p>

<p>regarding the issue of VERSION CONTROL with excel&#8217;s binary format &#8212; soon, you will be able to save excel files in XML format!!</p>

<p>BR,
~A</p>]]></content:encoded>
	</item>
	<item>
		<title>By: Javid Jamae</title>
		<link>http://marcus.ahnve.net/2006/08/10/why-i-dont-like-fit/#comment-1699</link>
		<dc:creator>Javid Jamae</dc:creator>
		<pubDate>Fri, 11 Aug 2006 11:41:29 +0000</pubDate>
		<guid isPermaLink="false">http://marcus.ahnve.net/2006/07/29/why-i-dont-like-fit/#comment-1699</guid>
		<description>&lt;p&gt;My first experience with FITnesse was on a Valtech project. I had similar concerns with version control. I also had a problem with the FIT API as well. It didn't seem to follow common Java/OO idioms (such as encapsulation). Though there are well thought out reasons for this, I just feel uncomfortable with it.&lt;/p&gt;

&lt;p&gt;Really, the question for me comes down to: who is writing these things. If the customer/non-technical BA is, then I understand the desire for using a tool like this. But in my experience, projects that start out with the customer/non-technical BA writing the tests, end up with the developers eventually taking over the actual development and maintenance the tests. Because of this experience, my first inclination on many projects is to just skip the fancy UI and write the darn integration tests using JUnit, a tool that is known and familiar to developers.&lt;/p&gt;
</description>
		<content:encoded><![CDATA[<p>My first experience with FITnesse was on a Valtech project. I had similar concerns with version control. I also had a problem with the FIT API as well. It didn&#8217;t seem to follow common Java/OO idioms (such as encapsulation). Though there are well thought out reasons for this, I just feel uncomfortable with it.</p>

<p>Really, the question for me comes down to: who is writing these things. If the customer/non-technical BA is, then I understand the desire for using a tool like this. But in my experience, projects that start out with the customer/non-technical BA writing the tests, end up with the developers eventually taking over the actual development and maintenance the tests. Because of this experience, my first inclination on many projects is to just skip the fancy UI and write the darn integration tests using JUnit, a tool that is known and familiar to developers.</p>]]></content:encoded>
	</item>
</channel>
</rss>
