uk-netmarketing Archive (2011-2015)

Hi Ben,



Yep, easy to be cynical with advancing years!



And yes,  different developers different speed -   but  velocity is the
great leveller, and a good scrum mater will manage that.   Obviously
collaborative planning and estimating helps,  hence planning poker on the
more complex development tasks.   But you are only as good as the
consistency of your estimates.   The more you do that the better it gets
though, as the team buys in and sees that it takes pressure off the more
you do it.



Also I would add different types of development impact that – belt and
braces process not required for a simple website implementation with
standard feature set, for example.   But on-going software (product)
development or large, complex implementation a different matter.

We introduced both Agile and TDD (and bits of XP,  predominantly pairing)
a few years back.  The pairing (and regular team code reviews) offset the
silliness,  both in coding and determining what to test,  and whilst it
slows you down short term the gain overall in quality and productivity mid
to  long term is good.   Across the company and multiple projects the
evidence is anecdotal  (easier to measure on a single project),   but we
are working on ‘throughput’  accounting to measure that -  found this book
useful for the throughput measurement
http://www.amazon.co.uk/gp/product/0131424602  -  hopefully soon to be put
into practice.



We’ve backed off on 100% test coverage upfront,  as I think you are right,
it can be too onerous.   Diminishing returns, I guess.   We tend to test
core system components,  build out making judgment calls on what to test
within each project,  and accept that there will be a bit of a build-up of
technical debt and refactoring as a consequence.     But that is no bad
thing.   As with anything there’s a balance to be found.



It’s nice to be able to deliver systems that are more usable and used and
less buggy and more predictable though,   which is the outcome of agile and
test driven process.



angus











*From:* uk-netmarketing-bounces at mm.chinwag.com
[mailto:uk-netmarketing-bounces at mm.chinwag.com] *On Behalf Of *Ben Thompson
*Sent:* 01 December 2011 11:45
*To:* uk-netmarketing
*Subject:* [uk-netmarketing] Developers and development [was Sports -
Community Website Platform]



The problem here is that not all developers are equal. Some can be 20 times
faster than others simply because they have better knowledge or tools to do
the same job.



My viewpoint on Agile is that it appeared at the same time as Test Driven
Development to offset the additional time TDD consumed. It works if you can
split delivery into separate releases but how many projects actually work
like that.



I'm also starting to query the joy of TDD as many people seem very capable
of creating the tests they expect to pass and miss all the ones regarding
Joe Hacker trying to break the site for fun.



It may be that I'm getting more cynical in my old age. Those of you who
know me from days gone by will probably think that wasn't possible but I
can assure you it is.



Regards,



Ben Thompson

On 30 November 2011 17:28, John Braithwaite <john at ergodigital.com> wrote:

Hah! Did you just use the phrase 'paradigm shift' without a grinning a
little at the cleverness of it all.