10 Years Later, What’s Next For Agile?

I just got back from Agile 2011 and I have to admit, I was skeptical. It’s been 10 years since we signed the Agile Manifesto and I just had to wonder, “it’s been 10 years – why do we still need a conference on this? We get it already!”

But I was also excited because they had a new stage called New Horizons about What’s Next for Agile? And I had the extreme honor of getting to present on this stage to share what’s been happening in the Startup community with the Agile community. To talk about how I believe Lean Startup is what’s next for agile. Not just startup projects, but all development.

And in the end – as it always seems to be – meeting all of the awesome people, having all of the enlightening (and sometimes slightly drunken) discussions, and remembering just how freakin’ amazing the Agile community truly is – I left feeling incredibly inspired, energized, and with this unshakeable feeling that there is just so much potential – so much more we can all be doing. So, even if 10 years has gone by, that doesn’t take away from how much more we can accomplish.

And with that in mind, here’s a quick screencast on how I believe Lean Startup pushes agile development to the next level…

You can find my complete slide deck on slideshare and I’d love to hear your thoughts on what’s next for agile.

What do YOU think is next for agile? Or do you think all this agile stuff is just crazy talk?


15 responses to “10 Years Later, What’s Next For Agile?”

  1. Agile Scout Avatar
    Agile Scout

    Solid. Agile continues to grow – Hopefully in the positive direction

  2. Abby Fichtner Avatar
    Abby Fichtner

    Thanks! We’ll just have to make sure we’re guiding it right 😉

  3. Bob MacNeal Avatar
    Bob MacNeal

    The side by side comparison of Agile and Lean Startup is a concise and helpful representation of the evolution of thinking.

  4. Abby Fichtner Avatar
    Abby Fichtner

    Thank you so much!

  5. William Pietri Avatar
    William Pietri

    Hi, Abby! I admire your enthusiasm, but I request with great feeling that you stop telling people that the Lean Startup approach is good for all development. It’s not. Having been involved in the Agile world since before the term existed, and having been pursuing the Lean Startup methods for the last two and a half years, I think telling people otherwise is dangerous.

    There are certainly things people can learn from it, and as you mention a lot of exciting techniques are being pioneered there. But the business context is radically different. A startup has no stakeholders. No existing customer base, no existing code, and no existing product. It is a high-risk venture, and is very likely to be out of business soon. The people pursuing them are generalists by nature and extreme optimists by temperament. There is no power hierarchy, no management, and nobody who’s not an owner. The Lean Startup approach makes sense in that environment, but only in that environment.

    So by all means learn things from it. But learn them carefully, and call the resulting product something else.

  6. Lisa Davidson Avatar
    Lisa Davidson

    Very Nice post. I am more keen to know about Agile testing methodology and how can a software quality testing company best use it. Something interesting like http://www.qainfotech.com/agile_methodology.html

  7. Abby Fichtner Avatar
    Abby Fichtner

    Thanks, Lisa! The best source for information on Agile Testing is Lisa Crispin: http://lisacrispin.com/
    I’ve also got some testing posts here on HackerChick: http:////tag/testing

    In Lean Startup, the term “testing” takes on a wider meaning referring to testing our assumptions by watching the customer’s behaviors. Internal manual testing is minimal at best. Tests should be automated in order to facilitate continuous deployment, and tests by human beings should be done by customers rather than testers whenever possible.

  8. Adam Nemeth Avatar
    Adam Nemeth

    I hope that in 10 years time, we’ll be more confident with our systems, therefore most of these Lean consequences won’t be needed. I don’t think anyone tests gravity nowadays just to be sure that it still exists, and I hope agilists started to forget testing getters and setters and basic collection handlers (add/remove,etc)

    On the other side, a software is a medicine for an information handling problem. Just like we wouldn’t like to be tested with unvalidated medicines as humans, even if rats don’t resemble us very well in every way, I don’t think it’s as much fun for the users to be the guinea pigs instead of actually having some guinea pigs in the lab.

    Because when you publish something which turns out to be bad, you messed it up for your users. Maybe you call it back in an hour or two, let’s say a day, but you still messed up a lot of people’s life, in case you aren’t in some private testing anymore. These people want to rely on you on solving their daily needs, and if you fail them, they won’t trust you anymore.

    Continous deployment has the same kind of problem: if all your tests are automated, you’re essentially modeling the bureaucrats: your system will be, sorry, inhuman. The main problem with old methods was that if there was an idiot at the top somewhere, his mistakes went down unchanged ’til the end. Agile was about speaking up. You won’t have flexible automated tests, an automated test is a very rigid, stupid bureaucrat compared to humans. Tests won’t speak up this way.

    I’m sure all of you hated when in your neighborhood, in your office building, etc, the bureaucracy of architecture and urban planning made some stupid mistake. Yet in architecture, the workers and testers are still human, while in our case, they’re freakin’ robots.

    I think we, as educated people, should be able to use logical induction and probability theories before we give something out; we should be able to distinguish what we should look at in a certain situation (hence, modeling on paper).

    So, what I hope that 10 years later, we will recognize problems so much that we know the solution as well ( http://www.patternlanguage.com/archive/ieee/ieeetext.htm ), and that we realize, that what’s beautiful shall be beautiful even without its details, like the Taj Mahal ( http://bit.ly/q8lpIU ), even if the careful crafting of each little rose on it makes it one of the most beautiful human structures ever created.

    I hope that 10 years from now we’ll be using much more complex elements than a class, or statement with such confidence, that we will know the problem just as well as the solution, and that, 10 years from now, we’ll be such connected to our users that we don’t dare to use them as guinea pigs anymore, yet we understand their wishes before trying it out in real life. I hope we won’t be fulfilled with our internal problems this much.

    Because that’s what would help humanity with software engineering, that’s what would help us, humans.

Leave a Reply

Your email address will not be published. Required fields are marked *