<?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: The Kanban Story: Kanban Board Revisited</title>
	<atom:link href="http://blog.brodzinski.com/2010/02/kanban-board-revisited.html/feed" rel="self" type="application/rss+xml" />
	<link>http://blog.brodzinski.com/2010/02/kanban-board-revisited.html</link>
	<description>Dealing with software projects in real life</description>
	<lastBuildDate>Fri, 10 Feb 2012 19:07:34 +0000</lastBuildDate>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>http://wordpress.org/?v=3.3</generator>
	<item>
		<title>By: Pawel Brodzinski</title>
		<link>http://blog.brodzinski.com/2010/02/kanban-board-revisited.html#comment-3753</link>
		<dc:creator>Pawel Brodzinski</dc:creator>
		<pubDate>Fri, 05 Feb 2010 15:05:34 +0000</pubDate>
		<guid isPermaLink="false">http://blog.brodzinski.com/?p=1589#comment-3753</guid>
		<description>Actually I don&#039;t really get what they acutally do during this step and why tagging is so important in this case. It is hard to decide whether something adds value if you don&#039;t know context.

I may oversimplify here but the way I look at flow I want to have as few steps as possible but at the same time I want to be able to say what is happening with a feature at every moment. And I mean meaningful parts of process here.

When I look at the board and see a feature at &#039;ongoing testing&#039; station I know:
- testing has been started
- testing hasn&#039;t been finished
- either some parts of feature haven&#039;t yet been tested or there are bugs which haven&#039;t yet been fixed

And yes, it adds value after all because &#039;before&#039; a feature isn&#039;t good, i.e. there are bugs in there, and &#039;after&#039; it is good to deploy.

It is really hard for me to describe that way tagging, but it is easier with pre-production deployment which is a part of discussed stage. When I see a feature at this station I know it is deployed (or under deployment - we have no data whether there are &#039;ongoing&#039; and &#039;completed&#039; sub-columns) in pre-production environment and possibly basic sanity checks have been done.

In terms of adding value pre-production deployment is an ultimate test for production deployment which is crucial in systems which can&#039;t have long outage (or can&#039;t have outage at all). Before you aren&#039;t sure whether deployment will go smooth and after you&#039;re pretty sure it will.

However I could imagine doing test deployment during testing. It all really depends who does this (if these are separate teams it is a good idea to split columns), how much time it takes (if deployment is time-consuming you may want to know whether a team performs functional testing or test deployment) or if this is an important milestone for one of stakeholders (and you want to explicitly show it).

Anyway I wouldn&#039;t tell there is a definition of &#039;simple enough&#039; or &#039;as simple as possible&#039; - it all comes down to team&#039;s judgement in every specific case.</description>
		<content:encoded><![CDATA[<p>Actually I don&#8217;t really get what they acutally do during this step and why tagging is so important in this case. It is hard to decide whether something adds value if you don&#8217;t know context.</p>
<p>I may oversimplify here but the way I look at flow I want to have as few steps as possible but at the same time I want to be able to say what is happening with a feature at every moment. And I mean meaningful parts of process here.</p>
<p>When I look at the board and see a feature at &#8216;ongoing testing&#8217; station I know:<br />
- testing has been started<br />
- testing hasn&#8217;t been finished<br />
- either some parts of feature haven&#8217;t yet been tested or there are bugs which haven&#8217;t yet been fixed</p>
<p>And yes, it adds value after all because &#8216;before&#8217; a feature isn&#8217;t good, i.e. there are bugs in there, and &#8216;after&#8217; it is good to deploy.</p>
<p>It is really hard for me to describe that way tagging, but it is easier with pre-production deployment which is a part of discussed stage. When I see a feature at this station I know it is deployed (or under deployment &#8211; we have no data whether there are &#8216;ongoing&#8217; and &#8216;completed&#8217; sub-columns) in pre-production environment and possibly basic sanity checks have been done.</p>
<p>In terms of adding value pre-production deployment is an ultimate test for production deployment which is crucial in systems which can&#8217;t have long outage (or can&#8217;t have outage at all). Before you aren&#8217;t sure whether deployment will go smooth and after you&#8217;re pretty sure it will.</p>
<p>However I could imagine doing test deployment during testing. It all really depends who does this (if these are separate teams it is a good idea to split columns), how much time it takes (if deployment is time-consuming you may want to know whether a team performs functional testing or test deployment) or if this is an important milestone for one of stakeholders (and you want to explicitly show it).</p>
<p>Anyway I wouldn&#8217;t tell there is a definition of &#8216;simple enough&#8217; or &#8216;as simple as possible&#8217; &#8211; it all comes down to team&#8217;s judgement in every specific case.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Joshua Lewis</title>
		<link>http://blog.brodzinski.com/2010/02/kanban-board-revisited.html#comment-3751</link>
		<dc:creator>Joshua Lewis</dc:creator>
		<pubDate>Fri, 05 Feb 2010 13:42:02 +0000</pubDate>
		<guid isPermaLink="false">http://blog.brodzinski.com/?p=1589#comment-3751</guid>
		<description>Hi again,
I was reading some messages in the kanbandev Yahoo Group and came across this message: http://finance.groups.yahoo.com/group/kanbandev/message/6777 which reminded me of this conversation. In case you can&#039;t access the post, the relevant part is the columns the poster has on his board: 
&quot;waiting &#124; analysis &#124; development &#124; testing &amp; doc &#124; done &#124; tag and deploy to stag &#124; deploy to live&quot; 

The reason it made me think of this post because considering something like tagging your source to be adding value is interesting to me. I&#039;m not sure if tagging your source adds value or is secondary waste. It may indeed value if part of adding value is being able to track versions etc (typically this would satisfy a system administrator stakeholder).

How do you define value as simply as possible? If you start off with a value statement of &#039;deliver software&#039;, do you add terms such as &#039;maintainable&#039; to it?</description>
		<content:encoded><![CDATA[<p>Hi again,<br />
I was reading some messages in the kanbandev Yahoo Group and came across this message: <a href="http://finance.groups.yahoo.com/group/kanbandev/message/6777" rel="nofollow">http://finance.groups.yahoo.com/group/kanbandev/message/6777</a> which reminded me of this conversation. In case you can&#8217;t access the post, the relevant part is the columns the poster has on his board:<br />
&#8220;waiting | analysis | development | testing &amp; doc | done | tag and deploy to stag | deploy to live&#8221; </p>
<p>The reason it made me think of this post because considering something like tagging your source to be adding value is interesting to me. I&#8217;m not sure if tagging your source adds value or is secondary waste. It may indeed value if part of adding value is being able to track versions etc (typically this would satisfy a system administrator stakeholder).</p>
<p>How do you define value as simply as possible? If you start off with a value statement of &#8216;deliver software&#8217;, do you add terms such as &#8216;maintainable&#8217; to it?</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Joshua Lewis</title>
		<link>http://blog.brodzinski.com/2010/02/kanban-board-revisited.html#comment-3701</link>
		<dc:creator>Joshua Lewis</dc:creator>
		<pubDate>Wed, 03 Feb 2010 11:35:32 +0000</pubDate>
		<guid isPermaLink="false">http://blog.brodzinski.com/?p=1589#comment-3701</guid>
		<description>Yes, but only if you understand it as one of the phases in your value stream :)</description>
		<content:encoded><![CDATA[<p>Yes, but only if you understand it as one of the phases in your value stream :)</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Pawel Brodzinski</title>
		<link>http://blog.brodzinski.com/2010/02/kanban-board-revisited.html#comment-3700</link>
		<dc:creator>Pawel Brodzinski</dc:creator>
		<pubDate>Wed, 03 Feb 2010 11:26:36 +0000</pubDate>
		<guid isPermaLink="false">http://blog.brodzinski.com/?p=1589#comment-3700</guid>
		<description>I&#039;m not sure if I saw documentation anywhere else either, but I&#039;m sure there are teams which have this stage. If you work on a system where documentation is one of key products you need to incorporate it to your workflow. If you happen to work with Kanban it will appear somewhere before &#039;done done&#039;.</description>
		<content:encoded><![CDATA[<p>I&#8217;m not sure if I saw documentation anywhere else either, but I&#8217;m sure there are teams which have this stage. If you work on a system where documentation is one of key products you need to incorporate it to your workflow. If you happen to work with Kanban it will appear somewhere before &#8216;done done&#8217;.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Joshua Lewis</title>
		<link>http://blog.brodzinski.com/2010/02/kanban-board-revisited.html#comment-3699</link>
		<dc:creator>Joshua Lewis</dc:creator>
		<pubDate>Wed, 03 Feb 2010 11:17:47 +0000</pubDate>
		<guid isPermaLink="false">http://blog.brodzinski.com/?p=1589#comment-3699</guid>
		<description>I understand. I think its a very nice example of what a value stream is, and how the Kanban Board and its evolution can assist and even direct in better understanding your value stream.

I&#039;m not certain but I think its the first time I&#039;ve read about a documentation column on a board.</description>
		<content:encoded><![CDATA[<p>I understand. I think its a very nice example of what a value stream is, and how the Kanban Board and its evolution can assist and even direct in better understanding your value stream.</p>
<p>I&#8217;m not certain but I think its the first time I&#8217;ve read about a documentation column on a board.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Pawel Brodzinski</title>
		<link>http://blog.brodzinski.com/2010/02/kanban-board-revisited.html#comment-3698</link>
		<dc:creator>Pawel Brodzinski</dc:creator>
		<pubDate>Wed, 03 Feb 2010 10:38:45 +0000</pubDate>
		<guid isPermaLink="false">http://blog.brodzinski.com/?p=1589#comment-3698</guid>
		<description>If I had to tell you which criteria makes it to the board as explicit column and which does not I&#039;d say it depends on averege time and maximum time a sticky note would spend in the column. If it is at least a day there is a column, otherwise there is not.

But you stressed one crucial thing - it doesn&#039;t really matter whether there is column or not. What is crucial here is how well people understand value stream - you nailed this one.

By the way, on the very beginning we didn&#039;t have documentation as explicit stage in our process - it appeared because team preferred to see it this way.</description>
		<content:encoded><![CDATA[<p>If I had to tell you which criteria makes it to the board as explicit column and which does not I&#8217;d say it depends on averege time and maximum time a sticky note would spend in the column. If it is at least a day there is a column, otherwise there is not.</p>
<p>But you stressed one crucial thing &#8211; it doesn&#8217;t really matter whether there is column or not. What is crucial here is how well people understand value stream &#8211; you nailed this one.</p>
<p>By the way, on the very beginning we didn&#8217;t have documentation as explicit stage in our process &#8211; it appeared because team preferred to see it this way.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Joshua Lewis</title>
		<link>http://blog.brodzinski.com/2010/02/kanban-board-revisited.html#comment-3697</link>
		<dc:creator>Joshua Lewis</dc:creator>
		<pubDate>Wed, 03 Feb 2010 10:18:08 +0000</pubDate>
		<guid isPermaLink="false">http://blog.brodzinski.com/?p=1589#comment-3697</guid>
		<description>I understand your conditions for moving features into different columns are not explicit but well understood.

My point is, you&#039;ve got documentation as a column on your board. I.e., in your context, you specify documentation as a step in your value stream. I imagine the documentation step could be specified as part of the requirements for moving a card to a column on the right (i.e. a &#039;done contract&#039;)  instead of as an explicit column. Having it as a column makes sense to me becuase it makes sense in terms of adding value.

This led me to ask the question, do people sometimes have criteria in their &#039;done contracts&#039; that should rather be explicit columns in their board?

I guess this question becomes &#039;how well do people understand their value stream?&#039;</description>
		<content:encoded><![CDATA[<p>I understand your conditions for moving features into different columns are not explicit but well understood.</p>
<p>My point is, you&#8217;ve got documentation as a column on your board. I.e., in your context, you specify documentation as a step in your value stream. I imagine the documentation step could be specified as part of the requirements for moving a card to a column on the right (i.e. a &#8216;done contract&#8217;)  instead of as an explicit column. Having it as a column makes sense to me becuase it makes sense in terms of adding value.</p>
<p>This led me to ask the question, do people sometimes have criteria in their &#8216;done contracts&#8217; that should rather be explicit columns in their board?</p>
<p>I guess this question becomes &#8216;how well do people understand their value stream?&#8217;</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Pawel Brodzinski</title>
		<link>http://blog.brodzinski.com/2010/02/kanban-board-revisited.html#comment-3695</link>
		<dc:creator>Pawel Brodzinski</dc:creator>
		<pubDate>Wed, 03 Feb 2010 09:31:24 +0000</pubDate>
		<guid isPermaLink="false">http://blog.brodzinski.com/?p=1589#comment-3695</guid>
		<description>We haven&#039;t written hard requirements for each column - it&#039;s just more of common knowledge of the team.

Development is completed when:
- code is written
- unit tests for the code are created
- developer performed set of manual tests and in all went good
- if something can&#039;t be easily tested (i.e. integration interface) there is a testing tool (simulator or something) created

Bug is fixed when:
- fix is applied to specified version of the application; if we want to fix a bug in a couple of versions we submit a couple of bugs to our bug tracker
- unit test following reproduction scenario is created
- fix is retested by QA folks in one of official builds; we also have unofficial, quick, dirty builds but they don&#039;t count here

Feature is tested when:
- there&#039;s no known bugs or, in very special cases, we agree that version can be shipped with specific minor bug(s)
- all planned functional tests were performed
- whenever change or bug fix affects any part of the application, the affected part is fully retested

Feature is live when:
- version is deployed
- all sanity-check tests pass; these verify all basic fucntions of the app
- our customers can access the feature

Documentation is done when:
- our maintenance guy is happy with it

It would be hard to have each of these details having its own column on the board. We just collectively worked out the way we do our jobs.

The other thing is I believe the Kanban Board should have as little columns as possible but at the same time the board should reflect real workflow.

If we document our features and we do that after deployment and it takes significant time (up to a couple of days in some cases) there should be a column where a sticky note spends these two days.

If it happens, rarely but it happens, that a feature is at research and design stage for a week we add a column which shows it because that&#039;s not really development since work isn&#039;t progressing here in terms it is while the feature is in development.

But we don&#039;t add deployment stage since it takes on average less than hour to deploy the app. We don&#039;t move a feature between testing and bug fixing stages because these two happen in parallel and even if they didn&#039;t we would need to move a sticky notes up to few times an hour which doesn&#039;t make any sense.

It all depends on how you work. Consider the way of work as a hand and make Kanban Board a glove. It should just fit.</description>
		<content:encoded><![CDATA[<p>We haven&#8217;t written hard requirements for each column &#8211; it&#8217;s just more of common knowledge of the team.</p>
<p>Development is completed when:<br />
- code is written<br />
- unit tests for the code are created<br />
- developer performed set of manual tests and in all went good<br />
- if something can&#8217;t be easily tested (i.e. integration interface) there is a testing tool (simulator or something) created</p>
<p>Bug is fixed when:<br />
- fix is applied to specified version of the application; if we want to fix a bug in a couple of versions we submit a couple of bugs to our bug tracker<br />
- unit test following reproduction scenario is created<br />
- fix is retested by QA folks in one of official builds; we also have unofficial, quick, dirty builds but they don&#8217;t count here</p>
<p>Feature is tested when:<br />
- there&#8217;s no known bugs or, in very special cases, we agree that version can be shipped with specific minor bug(s)<br />
- all planned functional tests were performed<br />
- whenever change or bug fix affects any part of the application, the affected part is fully retested</p>
<p>Feature is live when:<br />
- version is deployed<br />
- all sanity-check tests pass; these verify all basic fucntions of the app<br />
- our customers can access the feature</p>
<p>Documentation is done when:<br />
- our maintenance guy is happy with it</p>
<p>It would be hard to have each of these details having its own column on the board. We just collectively worked out the way we do our jobs.</p>
<p>The other thing is I believe the Kanban Board should have as little columns as possible but at the same time the board should reflect real workflow.</p>
<p>If we document our features and we do that after deployment and it takes significant time (up to a couple of days in some cases) there should be a column where a sticky note spends these two days.</p>
<p>If it happens, rarely but it happens, that a feature is at research and design stage for a week we add a column which shows it because that&#8217;s not really development since work isn&#8217;t progressing here in terms it is while the feature is in development.</p>
<p>But we don&#8217;t add deployment stage since it takes on average less than hour to deploy the app. We don&#8217;t move a feature between testing and bug fixing stages because these two happen in parallel and even if they didn&#8217;t we would need to move a sticky notes up to few times an hour which doesn&#8217;t make any sense.</p>
<p>It all depends on how you work. Consider the way of work as a hand and make Kanban Board a glove. It should just fit.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Joshua Lewis</title>
		<link>http://blog.brodzinski.com/2010/02/kanban-board-revisited.html#comment-3691</link>
		<dc:creator>Joshua Lewis</dc:creator>
		<pubDate>Wed, 03 Feb 2010 06:52:21 +0000</pubDate>
		<guid isPermaLink="false">http://blog.brodzinski.com/?p=1589#comment-3691</guid>
		<description>I like the live and doc columns, with WiP limits because they force you to frequently perform those tasks otherwise the entire value stream gets held up. It makes a lot of sense to impose small WiP limits on these columns. I think a lot of people would be reluctant to implement a WiP limit on a &#039;live&#039; column because they&#039;re scared of deployments - which is indicative of other problems.

This also implies to me that to deliver value its not sufficient to just deploy a feature, but also to document it. This leads me to the question: &#039;what else is required to provide value?&#039;

Also, what is your opinion of including the documentation requirement as part of the &#039;done is done&#039; contract for a specific column, perhaps dev or test (I&#039;m not sure). Should certain requirements be part of &#039;done&#039; contracts or have their own columns on the board? How do you distinguish between them?</description>
		<content:encoded><![CDATA[<p>I like the live and doc columns, with WiP limits because they force you to frequently perform those tasks otherwise the entire value stream gets held up. It makes a lot of sense to impose small WiP limits on these columns. I think a lot of people would be reluctant to implement a WiP limit on a &#8216;live&#8217; column because they&#8217;re scared of deployments &#8211; which is indicative of other problems.</p>
<p>This also implies to me that to deliver value its not sufficient to just deploy a feature, but also to document it. This leads me to the question: &#8216;what else is required to provide value?&#8217;</p>
<p>Also, what is your opinion of including the documentation requirement as part of the &#8216;done is done&#8217; contract for a specific column, perhaps dev or test (I&#8217;m not sure). Should certain requirements be part of &#8216;done&#8217; contracts or have their own columns on the board? How do you distinguish between them?</p>
]]></content:encoded>
	</item>
</channel>
</rss>

