I was hoping to get some feedback from users who are using RT to support
software development teams when using SCRUM approaches to planning work.
We several project teams and keeping people allocated to project for
several months at a time.
We track work based on several tasks (child) related to single story
(Parent).
We also track impediments.
Use two week long iterations. At the end of iterations there are
sometimes several stories and tasks that need to be moved into the next
iteration. Stories (and related tasks) need to be moved from a backlog
queue to the current iteration.
I would like a project team members to look at stories (sorted by
priority) and related tasks for the current iteration and take ownership
of tasks.
Team would like to quickly answer/update tasks in order to provide
answers for stand-up questions. What did you do yesterday? what are you
doing today?
I would like to be able to report on story, task and impediment
life-cycles for project and iterations.
I see that there is talk of a REST style interface so I assume that one
could write script to extract issue data quite quickly.
I could then create burn down charts etc. as required from the extracted
data.
One feature I REALLY like in our current system is the ability to export
stories/tasks/impediments to excel.
Update values in excel.
Next one uploads stories/tasks/impediments back into the tracker via web
interface.
I see that there is a possibility of creating and updating in bulk via
the web interface.
I'm not sure what you mean when you refer to "SCRUM". However, we
ARE supporting many software applications with technical support teams
and Queues oriented for each application/team. We have 1 queue
specifically designed to act as the initial request Queue where tickets
are reviewed and prioritized, then (if approved) moved to the queue that
supports the application the ticket is asking work for. For example, one
large support group, Financial, supports all the various software
dealing with our financial organizations and they have 12 different
queues for that activity. There are, consequently at least 24 different
groups, where each queue has a user group that WORKS on the tickets and
another that is allowed to CREATE tickets (some that have the same
members). Our structure for privileges is rather tight (only a couple
for “Priviliged” Users. Mostly Queue-oriented)) that allow for keeping
tickets secure (only the OWNER and the Queue Manager can Modify a
Ticket) within their queue. We also have a QA Workflow process that
ensures our standard for QA Approval is followed before a ticket is
allowed to be “QA Approved” and then Resolved (for example, we NEVER
allow the ticket owner to QA Approve their own work). A separate user
from the QA Approval group are the only ones with privileges to modify
the CF that is used to indicate QA Approval.
Anyway, that’s pretty much a summary of how we do things. If the
details of any of this will help you, then I’d be glad to share how we
do it.
Kenn
LBNLOn 3/19/2009 10:45 AM, R�diger Wolf wrote:
Hi All
I was hoping to get some feedback from users who are using RT to support
software development teams when using SCRUM approaches to planning work.
We several project teams and keeping people allocated to project for
several months at a time.
We track work based on several tasks (child) related to single story
(Parent).
We also track impediments.
Use two week long iterations. At the end of iterations there are
sometimes several stories and tasks that need to be moved into the next
iteration. Stories (and related tasks) need to be moved from a backlog
queue to the current iteration.
I would like a project team members to look at stories (sorted by
priority) and related tasks for the current iteration and take ownership
of tasks.
Team would like to quickly answer/update tasks in order to provide
answers for stand-up questions. What did you do yesterday? what are you
doing today?
I would like to be able to report on story, task and impediment
life-cycles for project and iterations.
I see that there is talk of a REST style interface so I assume that one
could write script to extract issue data quite quickly.
I could then create burn down charts etc. as required from the extracted
data.
One feature I REALLY like in our current system is the ability to export
stories/tasks/impediments to excel.
Update values in excel.
Next one uploads stories/tasks/impediments back into the tracker via web
interface.
I see that there is a possibility of creating and updating in bulk via
the web interface.
Thanks for the link. After reading the info, I believe that RT could
/easily/ be used to promote such a process. Some terms would have to be
re-named or their meaning translated, but the overall process design is
VERY similar to what we do now in our organization. I have developed
quite a bit of documentation (guides, flowcharts, rules, definitions,
etc.) that supports our method for getting work done in this manner.
SCRUM is a whole new set of terminology, but the process isn’t much
different than the way I did things 35 years ago in a Forbes 100
company. Interesting how some things just seem to cycle around, but with
newer coats of paint.
Kenn
LBNLOn 3/19/2009 12:38 PM, Tom Lahti wrote:
Ken Crocker wrote:
Rudiger,
I'm not sure what you mean when you refer to "SCRUM".
Based on what you say it seems like the following might work.
Each project/team has a BACKLOG QUEUE of stories(requirements). The
product owner regularly prioritizes them.
During the fortnightly planning meeting the product owner selects some
stories from the BACKLOG and moves them to the CURRENT SPRINT QUEUE
prioritizing them as he/she does so.
Team members select stories from the CURRENT SPRINT QUEUE and add child
tasks.
When all the tasks are done for a particular story, the story is moved
to a REVIEW QUEUE.
During the sprint review meeting the product owner reviews the features
related to stories in the REVIEW QUEUE and stories are marked as done.
Is it possible to roll up child data into a parent?
The question might be how many hours of work are remaining for this
story?
Hours work for a story(parent) = sum of remaining work estimate for all
its tasks(child).
Is it easy to create child tasks? And navigate through them via the web
interface?
I'm not sure what you mean when you refer to "SCRUM". However, we
ARE supporting many software applications with technical support teams
and Queues oriented for each application/team. We have 1 queue
specifically designed to act as the initial request Queue where tickets
are reviewed and prioritized, then (if approved) moved to the queue that
supports the application the ticket is asking work for. For example, one
large support group, Financial, supports all the various software
dealing with our financial organizations and they have 12 different
queues for that activity. There are, consequently at least 24 different
groups, where each queue has a user group that WORKS on the tickets and
another that is allowed to CREATE tickets (some that have the same
members). Our structure for privileges is rather tight (only a couple
for “Priviliged” Users. Mostly Queue-oriented)) that allow for keeping
tickets secure (only the OWNER and the Queue Manager can Modify a
Ticket) within their queue. We also have a QA Workflow process that
ensures our standard for QA Approval is followed before a ticket is
allowed to be “QA Approved” and then Resolved (for example, we NEVER
allow the ticket owner to QA Approve their own work). A separate user
from the QA Approval group are the only ones with privileges to modify
the CF that is used to indicate QA Approval.
Anyway, that’s pretty much a summary of how we do things. If the
details of any of this will help you, then I’d be glad to share how we
do it.
Sorry for the delay. I've been up to my *&^%$% in alligators so I've
been delayed in draining the swamp.
What you state IS workable. We do not move the finished “Stories” to
a review queue, we just added a Ticket Status" of QA testing. We run our
queries accordingly and review the “QA” results of the stories in that
status.
The answer to your question as to navigating to/from child tickets
in the Web is YES. However, I think that the 3.8.2 makes it MUCH easier
to create them.
The rolling up of child data (like time) into the parent is
something you could do, but you would have to create your own cron job
to do that. I haven’t seen anything like that in delivered RT.
Hope this helps.
Kenn
LBNLOn 3/19/2009 3:07 PM, R�diger Wolf wrote:
Based on what you say it seems like the following might work.
Each project/team has a BACKLOG QUEUE of stories(requirements). The
product owner regularly prioritizes them.
During the fortnightly planning meeting the product owner selects some
stories from the BACKLOG and moves them to the CURRENT SPRINT QUEUE
prioritizing them as he/she does so.
Team members select stories from the CURRENT SPRINT QUEUE and add child
tasks.
When all the tasks are done for a particular story, the story is moved
to a REVIEW QUEUE.
During the sprint review meeting the product owner reviews the features
related to stories in the REVIEW QUEUE and stories are marked as done.
Is it possible to roll up child data into a parent?
The question might be how many hours of work are remaining for this
story?
Hours work for a story(parent) = sum of remaining work estimate for all
its tasks(child).
Is it easy to create child tasks? And navigate through them via the web
interface?
Thanks
Rudiger
On Thu, 19 Mar 2009 12:10 -0700, “Ken Crocker” kfcrocker@lbl.gov wrote:
Rudiger,
I'm not sure what you mean when you refer to "SCRUM". However, we
ARE supporting many software applications with technical support teams
and Queues oriented for each application/team. We have 1 queue
specifically designed to act as the initial request Queue where tickets
are reviewed and prioritized, then (if approved) moved to the queue that
supports the application the ticket is asking work for. For example, one
large support group, Financial, supports all the various software
dealing with our financial organizations and they have 12 different
queues for that activity. There are, consequently at least 24 different
groups, where each queue has a user group that WORKS on the tickets and
another that is allowed to CREATE tickets (some that have the same
members). Our structure for privileges is rather tight (only a couple
for “Priviliged” Users. Mostly Queue-oriented)) that allow for keeping
tickets secure (only the OWNER and the Queue Manager can Modify a
Ticket) within their queue. We also have a QA Workflow process that
ensures our standard for QA Approval is followed before a ticket is
allowed to be “QA Approved” and then Resolved (for example, we NEVER
allow the ticket owner to QA Approve their own work). A separate user
from the QA Approval group are the only ones with privileges to modify
the CF that is used to indicate QA Approval.
Anyway, that’s pretty much a summary of how we do things. If the
details of any of this will help you, then I’d be glad to share how we
do it.
