Agility - SDN Hot Topic?

[ttrapp]

Hi there,

I would like to read more about agile methods in SDN, think of Scrum or Extreme Programming. Perhaps this would be even a topic for SDN day or an SDN Topic of the month.

At the moment I'm writing a blog entry about my thoughts while listening to an inspiring lecture about SCRUM @ SAP.

Cheers,

Tobias

[ttrapp]

SAP implemented Scrum and othert techniques successfully. More and more customers do it, too.

[Former Member]

Hi Tobias,

Do you have any documentation on the same?

How would you implement SAP using Scrum?

Thanks,

Mohammed.

[Former Member]

hi tobias,

nice topic you brought up and thx for the link to the SAP presentation.

what often comes to my mind when I read such agile development abstracts is this: agile development like the scrum variant is always very much about intensive communication amongst the involved players aswell as tight team building. I understand this is very crucial to the successof the methodology. Now, I wonder how this can be achieved, given that in this industry often times the stakeholders are physically widely distributed. even with the best supportive tools (videoconferencing, collaborative tools, etc.) available I do not believe that this is enough (setting aside the practical problems arising when you really want to set up an optimal infrastructure accessible by developers, product managers, user groups, customers, ...).

Anyway, on my way home from work I grabbed a copy of <a href="http://www.amazon.com/Agile-Software-Development-Cooperative-Game/dp/0321482751/ref=pd_bbs_sr_1/103-0535208-8967836?ie=UTF8&s=books&qid=1172855265&sr=8-1">Agile Software Development - The Cooperative Game</a> for the weekend to read. I hope it's worth it.

I hope we get some interesting discussion going here in the Coffee Corner,

anton

[Former Member]

Hi Tobias -

Glad to see you at CoffeeCorner.

I've only seen E(xtreme) P(rogramming) actually used once - and it struck me as a logical offshoot of RAD/JAD approaches that have been around for many many years. (As we all know, when heavy-hitter consultants can't come up with anything new to bill out on, they recycle old stuff with new words and bill out on that.)

EP definitely makes for an exciting work-place - the only problem is the usual one - controlling the "everybody talking at once" syndrome.

Regards

djh

[ttrapp]

Hi David,

yes, it took some time until I found the SDN CoffeeCorner. In real life I'm much faster there

You are absolutly right with your comment about recycling and that "everybody talking at once" syndrome. But to me there are even more problems:

*) How can we integrate that kind of development model with existing processes/quality standards?

*) Do we need organizational changes if we want to apply agile methods?

*) What does the management think about agile methods?

*) And what kind of problems/challenges occur in bigger development projects?

In fact SAP started some agile pilots. You can find a really interesting slides at http://www.andrena.de/ObjektForum/Veranstaltungen/61/Agile_Development_%40_SAP.pdf .

Cheers

Tobias

[Former Member]

Hi Tobias - this is a topic that really interests me and I will try to reply further over the weekend.

But I'd like to say this real quick about rethinking procedures.

When I was at UBS (then PaineWebber) Financial Services, the IT department had a really slick procedure (wasn't an SAP shop, but it doesn't matter):

1) if you moved your stuff to acceptance testing on time, you got credits toward your yearly bonus.

2) but - if you tried to "game" the system by pushing too early with bugs, and QA sent your stuff back to DEV, credits were subtracted.

They really used this protocol diligently and it really worked.

But it obviously won't work in a RAD/XP process.

RAD/XP is more like the way I type - I never bothered to become an accurate touch-typist in the standard sense - I developed my own very fast 80/20 method and I can correct so fast that I wind up completing perfect output about as fast as a true touch-typist.

But this kind of iterative error-correction process will NOT work in a shop that wants certain temporal checkpoints ...

Hope I've made myself clear here - it's a complicated issue ...

Best regards

djh