choosing platform for project / component

[Wass Lisovsky]

How does XP suggest to choose development platform: OS, database,
language for front-end and for business logics etc etc - provided that
the team members are strong enough in several platforms, and there is
enough resources around with a necessary expertise in other platform
that can be added to the team?

Unlike design which can be changed radically several times during the
project, platform can be changed with a seriously greater cost which
can be hardly reduced.

And is it exclusive responsibility of developers or of manager to
choose technology?

Wass

[Uncle Bob (Robert C. Martin)]

liwass@yandex.ru (Wass Lisovsky) might (or might not) have written
this on (or about) 11 Jul 2003 08:39:07 -0700, :
How does XP suggest to choose development platform: OS, database,language for front-end and for business logics etc etc - provided thatthe team members are strong enough in several platforms, and there isenough resources around with a necessary expertise in other platformthat can be added to the team?Unlike design which can be changed radically several times during theproject, platform can be changed with a seriously greater cost whichcan be hardly reduced.And is it exclusive responsibility of developers or of manager tochoose technology?

XP doesn't say anything at all about this. On the other hand, if you
follow XP principles, you will find that this selection happens in a
reasonable way.

An XP team that had not decided on a platform would write "platform
selection" as a story. The customer would write an acceptance test
for that story. The developers would estimate the story. The
customer would schedule that story based upon it's importance and
value.

What if the customer didn't pick that story early?

Possibly the developers would choose a platform independent language
to work in. Then the customer could delay the platform story for a
long, long, time. Indeed, the customer may be able to discard the
story and use any platform at all.

Alternatively, the developers might be unable to estimate the other
stories. (Their estimates would be infinite ;-) When the customer
asks the developers what it will take to shrink their estimates below
infinity, they'll answer that they need to know what the platform is.
Then the customer will pick the platform story.

The acceptance test on the platform story could be very telling. It
might simply be:

assertEquals("Windows", platform);


Robert C. Martin | "Uncle Bob"
Object Mentor Inc.| unclebob @ objectmentor . com
PO Box 5757 | Tel: (800) 338-6716
565 Lakeview Pkwy | Fax: (847) 573-1658 | www.objectmentor.com
Suite 135 | | www.XProgramming.com
Vernon Hills, IL, | Training and Mentoring | www.junit.org
60061 | OO, XP, Java, C++, Python |

[Wass Lisovsky]

Sounds like extempore ;-)

Fits into the framework, but does it work in real life? Or, to be
precise, you suggested the way the task should be planned, not the
process the task itself should be done.

Except trivial cases that are few, platform selection is always
decision over multiple variables: reliability, readable code,
maintenance costs for application in production, cost of expertise
etc.

Developers can estimate some of those variables (provided they have an
idea of the system to be developed) -- but isn't it a complicated
process to make a decision when those estimations are on the plate?
Decision that can't be easily formalized into a single formula or even
algorightm?

Who should make the decision, customer or developers - that's a
question.

Wass


"Uncle Bob (Robert C. Martin)" <u.n.c.l.e.b.o.b@objectmentor.com> wrote in message news:<i8bugvoeogt7su9h4shltbs4c1cql4rjnk@4ax.com>...How does XP suggest to choose development platform: OS, database,language for front-end and for business logics etc etc...?Unlike design which can be changed radically several times during theproject, platform can be changed with a seriously greater cost whichcan be hardly reduced.And is it exclusive responsibility of developers or of manager tochoose technology? XP doesn't say anything at all about this. On the other hand, if you follow XP principles, you will find that this selection happens in a reasonable way. An XP team that had not decided on a platform would write "platform selection" as a story. The customer would write an acceptance test for that story. The developers would estimate the story. The customer would schedule that story based upon it's importance and value. The acceptance test on the platform story could be very telling. It might simply be: assertEquals("Windows", platform);