Looking for examples of frequent delivery
I'm working on a little bit of writing, and I'd like to talk with someone
with some real experience with frequent delivery - at least every 6 months.
Let me be more specific: By delivery I mean software put into daily use by
people using the product for its intended purpose.
There are lots of Agile environments where frequent delivery is common - but
where delivery has a softer definition. It might be delivered to an
acceptance test environment, or even to a server where end user can see and
test the software, but may not use it.
I was speaking to someone on the phone yesterday who was questioning
iterative development. She said something to the affect of "How do you deal
with the issue where users see early releases of the software that they
can't yet use, and don't give much feedback? Then, when more releases come
out, they just dismiss them as being a slight change of what they already
saw?" I asked: "Are they using the software?" "No she said, it won't
really be ready to use until many more releases are completed."
For me this is an example of iterative development, but NOT incremental
release, in that what was released was not usable by its intended audience
for its intended purpose.
In the example above, if the releases were actually intended to be used,
they were complete failures since they weren't. The business value in this
case is much less than zero, in that the software cost the business money,
but then wasn't used to deliver any return on that investment. Eventually
it may be, but in the example above, the user community is being conditioned
to discount the result from the outset making it even harder to get
So, I'm curious to talk to someone who really does incremental delivery and
can speak to how their process has changed as a result. It's my assertion
that release planning becomes a bit more complex than "highest value
features first" when at release time the software has to be usable. I'd
suspect that testing and validating the software is more sophisticated than
automated acceptance test as well. It also pushes the necessity for an
acceptable level of user experience.
If you work in a company that falls into this frequent release category, and
have time to share your experience, please contact me off list to arrange
for a phone conversation. If you work for a company that fails at
incremental delivery, I'd like to hear about that as well. I'd like to
understand what causes a delivery to fail - assuming we're not talking an
engineering failure here. And, retrospectively, I'd like to hear what your
organization has done or might to to change things.
And, alternatively, if anyone's willing to share experiences on the list,
I'd love to hear/read that as well. If someone takes the time to, readers,
please refrain from armchair-quarterbacking. It's good to ask questions to
understand, but not imply that person or their company should have done X
instead of Y.
Thanks in advance! I'd rely on my own experience here, but I'm too biased.
I see everything through user experience lenses. Asking someone else about
their's usually gives me a better perspective.
Agile usability discussion group:
"There is nothing that saps one's confidence as the knowing how to do a
- WIDEST DISTRIBUTION - PLEASE FORWARD - APOLOGIES FOR MULTIPLE POSTINGS
Carnegie Mellon University and the University of Madeira are proud to
announce a joint professional Master's in Human-Computer Interaction.
This unique 16-month interdisciplinary program is the first of its kind in
Europe. Beginning in September 2007, this joint-degree program offers
students an opportunity to study on both campuses in computer science,
design, psychology and social sciences and to work in teams on an extended
applied HCI project in industry.
More information can be found at http://mhci.dme.uma.pt/
We are proud of this new initiative under the Carnegie Mellon/Portugal ICTI:
Information and Communications Institute.
I would appreciate your help in bringing this to the attention of potential
students. Organizations interested in hosting a student team for an HCI
project should contact Prof. Nuno Nunes (njn@...) or Larry Constantine
--Larry Constantine, IDSA, ACM Distinguished Engineer
Professor, Department of Mathematics & Engineering
Director, Laboratory for Usage-centered Software Engineering (Lab-USE)
University of Madeira, Funchal, Portugal
- 2007/3/30, Jeff Patton <jpatton@...>:
> If you work in a company that falls into this frequent release category, and
I'm not sure if my client companies belong to this category.
Decide yourself if my cases fit, and please contact me if you want
My usual project are used by more than at least tens of thousands of
users. It reaches sometimes up to millions. In these cases, what I
usually do is gradually increasing the number of users who will get to
use the frequent releases.
On one project, we did release after 3 to 4 months from the beginning,
to a dozen of (pre-alpha?) users. They started to use the service from
then, and we started to observe and gather feedbacks, while developing
simultaneously. After one month the user base increased to a few
hundreds, and after a couple of months again, it increased to more
than a thousand. Most of the acculumated users are still using the
system. It's more like a constant upgrade, and once in a couple of
weeks a major upgrade. Users get real values from using the service
since the first release(the amount of the value varies though). We
learned quite a lot thanks to this feedback from real uses, early and
Current project is more extreme. From the day 1(if we ignore the
period for initial user research and planning) of development, we have
released, and we are releasing almost daily, adapting the plan. We are
actually adding values day by day. However, the user base is very
small. Around a dozen. We intend to enlarge the size as we go on.