Live data from Hacker News

IBM Is Going Agile

blogs.wsj.com

41–49 of 49 posts

Re: IBM Is Going Agile

#41

"The mission is to have innovation and the speed of small companies..." If they keep laying off people, eventually they will be a small company. Problem solved!

> If they keep laying off people, eventually they will be a small company.

They'll be a small company, but they'll have the innovation and speed of a large company :)

Re: IBM Is Going Agile

#42
#1) Nothing says Agile like top down process design!

#2) I didn't quite believe Michael O Church's anti-Agile essay [1], but nothing justifies it more than this quote:

> The system is designed to foster accountability. And with that, finally, comes speed.

[1] https://michaelochurch.wordpress.com/2014/06/20/on-programme...

Re: IBM Is Going Agile

#43
post #24
post #20

Enterprise companies are all going agile this year, I bet someone from Gartner is suggesting this shift... and the result is clearly not that effective.

Seriously, who instigates these fads? It's like every year there's some new business model that everyone follows like a bunch of lemmings. Last year (or the year before?) it was everyone working in the office (as opposed to at home) because someone at Yahoo! decreed that was the way to go.

If it isn't obvious, announcements like these are purely PR. Once the general public starts reading the stories like "Facebook is agile and uses open offices", they begin to associate the experience of their Facebook iPhone app with the result using agile and open offices. They then ask their boss why their CRUD accounting/finance/analytics app can't look and feel like FB and then those bosses (the CRUD vendor clients) put pressure on their vendors (such as IBM) to change or they're going elsewhere.

Re: IBM Is Going Agile

#45

I worked at IBM for 5 years, and some of the teams I was on used "agile" development methods. In that they used the names, but our "scrums" were an hour long, in a meeting room, with everyone bringing their laptop. Management more than 1 level up expected waterfall-like development, so it all just meshed badly. I really don't hold out much hope that this is going to work out that well without a whole bunch of trainin…

I spent a long time at a boutique consultancy that specialized in rescuing at-risk software projects for fortune 100 firms. Mainly, that meant we got very, very good at building up waterfall adapters for our internal processes.

I won't say we were following agile, exactly, but the concept remains approximately the same. Big company management thrives on the mistaken notion that they can plan out the world years in advance. The key insights that made our adapters work were 1) no one has enough of a memory to hold anyone to those plans even six months down the line and 2) no one expects anything to work anyway, since they're all used to their plans falling over.

Which isn't to say this is going to work. I suspect that, like all things recent IBM, any success they find will be more attributable to inertia than execution excellence.

edit: clarification

Re: IBM Is Going Agile

#46
post #11

That seems unlikely. Big companies like big ships have a lot of inertia. They may adapt "Agile", but almost certainly in name only.

This was the case in my few months there. We had "sprints" and "stories" but 75% of all the stories in every sprint would just get pushed to the next sprint with no thought behind it. Part of that was because our stories themselves were so broad and were worked in a long term waterfall manner. Our standups were just everyone falling asleep while each person took their turn saying what they did the previous day. If so…

Sounds like pro-forma agile - phagile - if you will.

I was taught agile by a certified scrum master, but the dude was completely serious about 'this just lets you fail faster, and you will fail, and often, as this process starts' which really helped.

He was right. It took a while to learn how to keep people honest and create succinct stories with a real scope. Particularly the guys who want to be architects or engineers perceive their jobs to start with the abstract and work down to specificity.

They can struggle with "What should this screen to today. Right now, today?" and want to talk about database design, message passing, data structures or anything else that doesn't answer the question.

At least for me and my planning it was far easier to make stories as small as possible and let the complexity build into a release which was roughly plotted out after velocity stabilized (even if into an upward slope).

I was probably doing it all wrong, but the hardest part was always keeping anyone from saying "well 6 more 2 weeks sprints means your release date is x, right?". Everyone wanted to just take the backlog depth, divide by velocity and project a date.

Re: IBM Is Going Agile

#47

I'm surprised that there is no mention of Rational Team Concert: http://www-03.ibm.com/software/products/en/rtc Used in a big way to implement Agile within IBM and provides a nice coherent system for task management.

I'd rather unclog blocked toilets with my bare hands than touch anything with "Rational" on its name

As someone who regularly uses Git (and prefers it), RTC really isn't all that bad. I'd even go so far as to recommend it for some teams/companies.

Re: IBM Is Going Agile

#48

Earlier quoted context omitted.

I'd rather unclog blocked toilets with my bare hands than touch anything with "Rational" on its name

As someone who regularly uses Git (and prefers it), RTC really isn't all that bad. I'd even go so far as to recommend it for some teams/companies.

As someone who wasted time and sanity with that ginourmous piece of crap called Clearcase, I'd rather flush money down the toilet than ever give money to IBM/Rational again

Re: IBM Is Going Agile

#49

I'm surprised that there is no mention of Rational Team Concert: http://www-03.ibm.com/software/products/en/rtc Used in a big way to implement Agile within IBM and provides a nice coherent system for task management.

Took a look at the white papers and this actually looks interesting. Why would I use this over something like Jira with Greenhopper? I'm trying to keep my mind open about tools like this.
Post reply on HN