Extreme Programming - Trends in yesterday's weather

Author

Trends in yesterday's weather

Frosty Madness

2005-09-20, 9:57 pm

If we don't estimate all user stories upfront (BRUP) how do we spot trends
in yesterday's weather?

For example, if a system is becoming increasingly difficult to change, you
would expect velocity to drop. But if we gradually experience this change,
we might just gradually alter our estimates so we're still implementing the
same number of story points per iteration. Whilst this might be an
unconcious shift, our estimates are being coloured by the viscosity of the
system. If all stories are estimated upfront, the drop in velocity becomes
more obvious.

Anyone know any good techniques for not investing in big requirements up
front *and* spotting trends in yesterday's weather? Or are weather patterns
restricted to the current release?

F

Ron Jeffries

2005-09-22, 3:56 am

On Tue, 20 Sep 2005 14:06:25 GMT, "Frosty Madness" <frosty.madness@gmail.com>
wrote:

>If we don't estimate all user stories upfront (BRUP) how do we spot trends
>in yesterday's weather?
>
>For example, if a system is becoming increasingly difficult to change, you
>would expect velocity to drop. But if we gradually experience this change,
>we might just gradually alter our estimates so we're still implementing the
>same number of story points per iteration. Whilst this might be an
>unconcious shift, our estimates are being coloured by the viscosity of the
>system. If all stories are estimated upfront, the drop in velocity becomes
>more obvious.
>
>Anyone know any good techniques for not investing in big requirements up
>front *and* spotting trends in yesterday's weather? Or are weather patterns
>restricted to the current release?

Are you presently experiencing something that seems to you might be a variation
in your use of yesterday's weather?

--
Ron Jeffries
www.XProgramming.com
I'm giving the best advice I have. You get to decide if it's true for you.

Frosty Madness

2005-09-22, 3:56 am

"Ron Jeffries" <ronjeffries@acm.org> wrote in message
news:pma4j1dn1ks7fdsrv34p1b6q9ekdgalt7p@
4ax.com...
patterns[color=darkred]
>
> Are you presently experiencing something that seems to you might be a
variation
> in your use of yesterday's weather?

Yes. I believe the software is becoming more resistant to change. We aren't
seeing any change in velocity though. I can't help thinking that a story
that would have been rated a 2 or a 3 earlier in the project is being rated
as a 5 at this point in the project. This could just mean that we're getting
more accurate in the planning game. But it might not.

If our requirements were quite static, I would feel OK about estimating
stories for a release rather than an iteration. But they're not. What could
I use to show up the viscosity?

F

Laurent Bossavit

2005-09-22, 6:57 pm

Frosty,

> If our requirements were quite static, I would feel OK about estimating
> stories for a release rather than an iteration. But they're not. What could
> I use to show up the viscosity?

What is your 'planning horizon' ? That is, how many iteration's worth of
prioritized, estimated stories are sitting in your backlog ?

The longer your planning horizon, the higher the "viscosity".

(And the longer the horizon, the more uncertainty in your planning. The
correct trade-off will depend to a large extent on why you're planning.
I think everybody should plan, but that there are vast differences in
teams' *purpose* for planning.)

Laurent

Naoya

2005-09-22, 9:57 pm

If your software resisting for change, it is time for refactoring. You need
to predict force coming changes and think better design.

naoya

"Frosty Madness" <frosty.madness@gmail.com> wrote in message
news:SWrYe.167104$R83.48888@fe04.news.easynews.com...
>
> "Ron Jeffries" <ronjeffries@acm.org> wrote in message
> news:pma4j1dn1ks7fdsrv34p1b6q9ekdgalt7p@
4ax.com...
up[color=darkred]
> patterns
> variation
>
> Yes. I believe the software is becoming more resistant to change. We
aren't
> seeing any change in velocity though. I can't help thinking that a story
> that would have been rated a 2 or a 3 earlier in the project is being
rated
> as a 5 at this point in the project. This could just mean that we're
getting
> more accurate in the planning game. But it might not.
>
> If our requirements were quite static, I would feel OK about estimating
> stories for a release rather than an iteration. But they're not. What
could
> I use to show up the viscosity?
>
> F
>
>

Laurent Bossavit

2005-09-23, 3:58 am

> If your software resisting for change, it is time for refactoring. You need
> to predict force coming changes and think better design.

And while you're at it, buy stocks low that are going to go up. You'll
make a killing if you do that...

;)

Naoya

2005-09-23, 7:57 am

"Laurent Bossavit" <laurent@dontspambossavit.com> wrote in message
news:MPG.1d9db11fa1b1750f9899b0@news.noos.fr...
need[color=darkred]
>
> And while you're at it, buy stocks low that are going to go up. You'll
> make a killing if you do that...
>
> ;)

So, how many have you killed? My first one was nearly killed, but after
that were all OK.
Always small change and many interations. All up to software design to me.

naoya

Phlip

2005-09-24, 6:56 pm

Frosty Madness wrote:

> If we don't estimate all user stories upfront (BRUP) how do we spot trends
> in yesterday's weather?

Last iteration, we had a record-breaking hurricane. This iteration, we are
having a record-breaking hurricane. Next iteration, we will haaave..?

;-)

--
Phlip
[url]http://www.greencheese.org/Z

Land[/url] <-- NOT a blog!!!

JXStern

2005-09-25, 3:58 am

On Thu, 22 Sep 2005 06:12:02 GMT, "Frosty Madness"
<frosty.madness@gmail.com> wrote:
>patterns
>variation
>
>Yes. I believe the software is becoming more resistant to change. We aren't
>seeing any change in velocity though. I can't help thinking that a story
>that would have been rated a 2 or a 3 earlier in the project is being rated
>as a 5 at this point in the project. This could just mean that we're getting
>more accurate in the planning game. But it might not.
>
>If our requirements were quite static, I would feel OK about estimating
>stories for a release rather than an iteration. But they're not. What could
>I use to show up the viscosity?

Hmm.

So you want to compensate for bit rot?

I'm not sure you should, even if you could. We all live in the
moment, and all that.

In my view the best software projects are those in which you do knock
down the big problems first, and then you do process increasingly
smaller issues over time. This is a GOOD THING! Well, at least, it
can be a good thing. I guess you're suggesting you see some phenomena
that are NOT good things. But I still like to see a GOOD project as
following some fractal process of adding in bells and whistles as a
system matures. The objective and subjective value of relatively
small features can be surprisingly large. Many of these are way down
the priority list just because they require touching a lot of areas,
and this may be seen as some kind of viscosity increase. It can be
hard to do well, too. And it's only human to renormalize the metrics
to fit current conditions. If my fractal ideas and comments are
valid, it's hard to see how any fixed scale could be any more useful.

J.

Frosty Madness

2005-09-26, 3:56 am

"Phlip" <phlipcpp@yahoo.com> wrote in message
news:sMkZe.1211$8B6.1150@newssvr24.news.prodigy.net...
> Frosty Madness wrote:
>
trends[color=darkred]
>
> Last iteration, we had a record-breaking hurricane. This iteration, we are
> having a record-breaking hurricane. Next iteration, we will haaave..?

Aren't hurricanes defined by a certain wind speed and a certain swell? If we
had a nasty hurricane last iteration, and we alter our definition of
hurricane to be a higher wind speed and a higher swell because hey, we
survived, right? Now suddenly, this iteration, we just have a stiff breeze.
We stop whinging and go wind surfing :)

F

Frosty Madness

2005-09-26, 3:56 am

Phlip

2005-09-26, 9:56 pm

Frosty Madness wrote:

> Aren't hurricanes defined by a certain wind speed and a certain swell? If we
> had a nasty hurricane last iteration, and we alter our definition of
> hurricane to be a higher wind speed and a higher swell because hey, we
> survived, right? Now suddenly, this iteration, we just have a stiff breeze.
> We stop whinging and go wind surfing :)

My goodness. I think we won't mind a meander into politics that strays
not _too_ far from the concept of Yesterday's Weather...

You have adequately sumarize the Official RNP response to Global
Warming!

;-)

--
Phlip

Sponsored Links