Kanban Digest #lkfr12
Le meilleur des twitts de la conférence Lean Kanban France 2012!
Vous n’y étiez pas! Voilà une petite synthèse pour voir la richesse des échanges qui ont eu lieu pendant ces deux jours de conférences sur le Kanban.
Day 1
@gmantel#LKFR12 keynote by @agilemanager starting strong: stop trying to apply agile methods by the book, be agile locally in your context
@LeanVoices Progress is better than perfection #lkfr12
@hakanforss « There is no Target Process in Kanban. Kanban is about improving your business » – @agilemanager#LKFR12
@PatrickSteyaert#lkfr12 items in the backlog are options @agilemanager keynote, would love to see the idea of incremental commitment to be more explicit
@dominicad « The kanban system is the flow of work on the board, not the board or cards themselves. » @agilemanager#lkfr12
@cyetain Adjusting explicit policies affects the flow of tickets & impacts the amount of rework. @agilemanager#lkfr12
@cyetain Here is the secret, the board isn’t the kanban. It is just a way to visualize a set of virtual kanban.
@nicolasstampf#lkfr12 only commit when work is pulled in the queue.
@JasperSonnevelt: ”Don’t adopt an agile method from the book. Think for yourself” @agilemanager keynote at #lkfr12
@mdeocto#lkfr12 « problems in IT ops : distributed teams, interrupts, specialization, dependencies » @dominicad
@mdeocto#lkfr12 @dominicad: « The point of Kanban in IT ops : Visualizing the risks. Show them ». Simple. Powerful.
@pierre_fauvel #lkfr12 @agilemanager implement one practice after another
@pierre_fauvel#lkfr12 @agilemanager there is no target process in Kanban
@pierre_fauvel#lkfr12 @agilemanager policies change with the trust into the team
@pierre_fauvel#lkfr12 @agilemanager process improvement after standup #lkfr12
@danielfroding “@hakanforss: To get to a pull based change you may need to start with a small push model. Act your way into new thinking.#LKFR12” agree!
@dominicad »The real bottleneck isn’t the IT system – it’s the change to the IT system. » @PatrickSteyaert#lkfr12
@pierre_fauvel#lkfr12#personnalkanban limit wip to keep track of the Story
@pierre_fauvel#lkfr12 @valtech le premier choc culturel c’est le wip
@dominicad « #kanban helps create capability to change. » @PatrickSteyaert#lkfr12#evolve
@pierre_fauvel#lkfr12 @valtech Kanban serait plus facile a mettre en oeuvre ds les entreprises de type « Controle »
@pierre_fauvel#lkfr12 @siamha la proprete du Kanban reflete la maitrise du processus
@pierre_fauvel#lkfr12 SocGen piloter les sujets, pas les personnes
@stph_p: Limiting WIP enables to focus and not to overload our memory with too many things. It preserves coherence #lkfr12
@ourfounder#lkfr12 @cyetain Doing the right experiment to learn a lot is pretty uncomfortable
@pierre_fauvel#lkfr12#personalkanban we need to complete to compete
@pierre_fauvel#lkfr12#personalkanban the minimum constraint to have things flowing
@pierre_fauvel#lkfr12#personalkanban change occur when somebody sees a problem and wants to fix it
@pierre_fauvel #lkfr12 #personalkanban we pospone the complex tasks until when it is too late, and then we do crap
Day 2
@pierre_fauvel#lkfr12#reinertsen high wip constraint have high effect with low cost, a good way to start
@pierre_fauvel #lkfr12 #reinertsen the cost of a queue is the cost of waiting time x the number of item delayed
@pierre_fauvel#lkfr12#reinertsen science based approach needs thinking and theory
@tlc_labs Non-homogenous flow=multiple work types w/ different costs of delay. EG a heart attack victim vs patient with a cold. ~@DReinertsen#lkfr12
@gmantel#LKFR12 HT @dreinersten you cannot load a process to 95% capacity with variability
@gmantel#LKFR12 @DReinertsen explains, based on maths and probabilty, that you need WIP constraints to reduce variability effects of a process
@pierre_fauvel#lkfr12#reneirtsen start with local wip constraints, less risky@pierre_fauvel#lkfr12#reinertsen don’t wait until buffer is full to manage flow@pierre_fauvel#lkfr12 @alexis8nicolas planifier la transformation -> changement de proche en proche, viral, evolution plutot que revolution@hakanforss small frequent improvement steps are better pierre_fauvel12:28pm via Twitter for Android#lkfr12@hakanforss swimlanes in standup : green:don’t discuss yellow: important but no need to discuss, red: issues we discuss@pierre_fauvel#lkfr12 @alexis8nicolas idee du mgt malmenee: tout le monde doit etre occupe a 100% -> optimisons plutot le flux de travail de l’equipe@pierre_fauvel#lkfr12 @alexis8nicolas planifier la transformation -> changement de proche en proche, viral, evolution plutot que revolution@pierre_fauvel#lkfr12 @hakanforss record each step, expectation, measure, learning@pierre_fauvel #lkfr12 @hakanforss it is fine if 50% of your experiments fail@o_olelan: « All improvements are experiments » we should have at least one active experiment a week – #lkfr12 @hakanforss