Lean Coffee Notes

24 views
Skip to first unread message

Matt Heusser

unread,
Jan 7, 2015, 9:14:45 AM1/7/15
to code...@googlegroups.com

A few early risers met, did introductions, learned the method, built the backlog, and even got a few stories kicked out.

I'll put my notes below my signature; you are welcome to join us tomorrow at 7:00AM in Rosewood.

regards,

--heusser

Cool Books

Thinking Fast and Slow, Kahneman

You are not so smart, David McRaney - Biases and what to do about it

You are now less dumb, David McRaney - Biases and what to do about it

Pragmatic Thinking and Learning, Andrew Hunt

Scrum: Twice in half the time, Jeff Sutherland (not great)

The peopel’s Scrum, Tobias Mayer

Quiet: The power of introverts in a world that can’t stop talking, Susan Kane

Zero to One, Peter Theil

Being Agile, Mario Moreira

The paradox of Choice, Barry Swartz  

http://www.ted.com/talks/barry_schwartz_on_the_paradox_of_choice


Freakonomics, Steven Levitt

Mindset: The New Psychology of Success by Carol Dweck

As you Wish: Inconceivable Tales from Making of the Princes Bride, Cary Elwes


Beginning with Test Automation

Large enterprise, tried tools before, just restart

Ability to run tests in parallel is important

One obvious overhead is people ‘losing a vm’

If you search through your code for domains and what machines are being used sometimes you get kind of a shock.

Better to do one test all the way through CI end-to-end than slowly 

We worked with one LeanDog client that did end to end overnight but headless and automated. Typical tools we used included fitnesse, cucumber, hudson. 

We had 7,500 cucumber scenarios that ran over 550 hours. Crazy. Our approach was to go through the user interface. I’d suggest design and architecture. Better to test deliberately at all layers - you need a seat at the design table with the dev teams.

(Host) Devs want input from us on how to make things more automatable, including something like TDD perhaps with different words.

Devs want input about testability - code planning. Not just functions but how will you test security, performance, etc. We actually meet to see how far down we can push the tests.

At some level test automation is the original agile. We’ve always written tools to do some form of testing but only to the extent that it helps us. We abandon it when it doesn’t, which is sort of the idea of minimum viable product. I think testing infrastructure and architecture are under-explored. When developers write code they usually have some design principles; testers-turn-coders often don’t, so what they write is pretty brittle. As an industry I think we need to put more thought into architecture/design of the test frameworks.




Chris Woodruff

unread,
Jan 7, 2015, 9:18:41 AM1/7/15
to code...@googlegroups.com
See you all tomorrow morning. Change of plans gets me in later this afternoon.

Chris Woodruff



Date: Wed, 7 Jan 2015 06:14:45 -0800
From: matt.h...@gmail.com
To: code...@googlegroups.com
Subject: [CodeMash] Lean Coffee Notes
--
You received this message because you are subscribed to the Google Groups "CodeMash" group.
To unsubscribe from this group and stop receiving emails from it, send an email to codemash+u...@googlegroups.com.
To post to this group, send email to code...@googlegroups.com.
Visit this group at http://groups.google.com/group/codemash.
For more options, visit https://groups.google.com/d/optout.

Matt Heusser

unread,
Jan 7, 2015, 11:39:56 AM1/7/15
to code...@googlegroups.com

During lean coffee someone pointed out that automating all the things at the gUI level is a fool's errand.

It reminded me of a good google tech talk that includes the true story how the folks at GOOGLE wrote far to many end-to-end selenium tests for gmail and ended throwing a bunch away:


You can criticize the folks at google for a lot of things, but the raw IQ of the team working on that project was probably higher than what most of us are doing. The motivation and dedication was probably high.

And. they. failed.

Maybe that just doesn't work? :-)

regards,


--heusser
Reply all
Reply to author
Forward
0 new messages