[datawander]

The demotivation poster (we'll ask for estimates, but then treat them as deadlines) really strikes a chord because that is one of my major 'gripes' with the agile planning process. Particularly when a manager is heavily involved in that process. It may be no coincidence that the best projects I have been on are the ones where the manager deliberately stepped out of the room or was not a part of those phases of the planning process.

Agile is part of a more disturbing trend I've noticed [0] of companies striving very hard to turn software into a literal sausage-making factory [1] and to make software engineers just another cog in the machine or a replaceable part to fatten the bottom line with a lower salary. This is provably the aim of some of the top companies given news on no-hire agreements. [10].

[0] with Java being the favored "currency" of programming languages being the other disturbing trend-- it's much easier to replace a Java programmer than any other for a reason

[1] you know what they say about how sausages are made

[10] http://gizmodo.com/apple-guilty-of-price-fixing-730018979

[wpietri]

Yeah. 100% agreed.

In some ways I don't blame people. Industrial approaches to organizing people provided a major leap forward for humankind. And they work well with primate power dynamics; modern corporate structures are basically feudalism in suits. It's natural that people would just want to take the top-down, command-and-control structures and replicate them in the new thing they're doing.

But they just don't work well. They don't even work well for industry anymore; there's a reason that Toyota, which has a very different management philosophy, wiped the floor with the US auto companies, which stayed stuck in the early 20th century.

To be fair, Agile started out to be 100% the opposite of that sausage-factory approach. I know a lot of the early players, and they sincerely had a very different vision. It makes me sad to see their work used as just another stick to beat developers. Meet the new boss, same as the old boss.

[jaggederest]

I think the correct way to do it is to do the old 'pick two' method - quality, features, timeline, pick two. (I'd say you never want low quality but... situationally.)

[zimpenfish]

Always double your estimates because management will cut them in half anyway.