Matt Stine

Greg-

This article represents more of a thought experiment than an attempt to draw a definitive conclusion in the matter - something to generate discussion (which it has). If I failed to convey that message then I apologize. At any rate, there's barely enough information out there to speculate...hence the framing of the article's title as a question.

-Matt

Greg-

This article represents more of a thought experiment than an attempt to draw a definitive conclusion in the matter - something to generate discussion (which it has). If I failed to convey that message then I apologize. At any rate, there's barely enough information out there to speculate...hence the framing of the article's title as a question.

-Matt

Greg-

This article represents more of a thought experiment than an attempt to draw a definitive conclusion in the matter - something to generate discussion (which it has). If I failed to convey that message then I apologize. At any rate, there's barely enough information out there to speculate...hence the framing of the article's title as a question.

-Matt

Greg-

This article represents more of a thought experiment than an attempt to draw a definitive conclusion in the matter - something to generate discussion (which it has). If I failed to convey that message then I apologize. At any rate, there's barely enough information out there to speculate...hence the framing of the article's title as a question.

-Matt

Greg-

This article represents more of a thought experiment than an attempt to draw a definitive conclusion in the matter - something to generate discussion (which it has). If I failed to convey that message then I apologize. At any rate, there's barely enough information out there to speculate...hence the framing of the article's title as a question.

-Matt

Greg-

This article represents more of a thought experiment than an attempt to draw a definitive conclusion in the matter - something to generate discussion (which it has). If I failed to convey that message then I apologize. At any rate, there's barely enough information out there to speculate...hence the framing of the article's title as a question.

-Matt

Greg-

This article represents more of a thought experiment than an attempt to draw a definitive conclusion in the matter - something to generate discussion (which it has). If I failed to convey that message then I apologize. At any rate, there's barely enough information out there to speculate...hence the framing of the article's title as a question.

-Matt

Greg-

This article represents more of a thought experiment than an attempt to draw a definitive conclusion in the matter - something to generate discussion (which it has). If I failed to convey that message then I apologize. At any rate, there's barely enough information out there to speculate...hence the framing of the article's title as a question.

-Matt

Greg-

This article represents more of a thought experiment than an attempt to draw a definitive conclusion in the matter - something to generate discussion (which it has). If I failed to convey that message then I apologize. At any rate, there's barely enough information out there to speculate...hence the framing of the article's title as a question.

-Matt

Greg-

This article represents more of a thought experiment than an attempt to draw a definitive conclusion in the matter - something to generate discussion (which it has). If I failed to convey that message then I apologize. At any rate, there's barely enough information out there to speculate...hence the framing of the article's title as a question.

-Matt

Excellent. I've touched a nerve (as I expected).

I do want to clear up one point. In no way am I asserting that "Java on the Mac" is not relevant. It is extremely relevant to me and the team I work with daily. We are 100% Mac developing 100% JVM server-side applications. The vast majority are Eclipse users, so we depend on a reliable client-side JVM to develop our software. With that said, multiple folks have said that Eclipse happens to run wonderfully on Soy Latte. I don't have enough information to know if the same can be said regarding Netbeans or IntelliJ IDEA.

Furthermore, should the same "economic analysis" be performed by Oracle, Google, or other companies, the results may differ. It might make sense for one of these companies to develop a Mac OS X JVM that integrates with the native Aqua UI. Unfortunately, I just don't have enough information to predict what any of these companies are going to do.

I can say that I have "heard" that Oracle isn't the friendliest company when it comes to using the Mac platform as an employee of the company, and it's incredibly hard to get Apple hardware paid for. Take that for what it's worth.

While it's not a PDF, the printer-friendly view (you'l find a link very close to the bottom of each article) provided by our site is fairly effective. Give it a try and let us know if it meets your needs.

Cheers,

Matt

While it's not a PDF, the printer-friendly view (you'l find a link very close to the bottom of each article) provided by our site is fairly effective. Give it a try and let us know if it meets your needs.

Cheers,

Matt

While it's not a PDF, the printer-friendly view (you'l find a link very close to the bottom of each article) provided by our site is fairly effective. Give it a try and let us know if it meets your needs.

Cheers,

Matt

@Scott

Idealistic? Absolutely. The example is intentionally oversimplified to demonstrate the fact that even in an ideal scenario with no task dependencies, deadlocks, etc., task switching has a real cost.

Software tools can help, but cannot eliminate the cost, as it is a cognitive issue. Even if I'm immediately presented with all of the relevant information/artifacts necessary to complete my task following a task switch (something that Mylyn does extremely well), I still have to "switch my brain on" to that task. It still takes me real time to get into flow. And every time that I have to restart that process, it's costing my company/customer real money.

So while we can only eliminate this problem in a dream world, we can take actions to minimize it as much as possible. This article is an attempt to point folks in the right direction.

Cheers!

Thanks (I think?)!

I certainly agree with your statement about the pain of merge going up with multiple projects, repositories, etc. I think the key here is to use good judgment about your repository structure. If such a complex structure is mandatory for some reason and cannot be adjusted, feature branching is probably not a good idea. That said, I think you'd be hard pressed to deliver on some of the things that I mentioned with any strategy coupled to such a complex repository structure.

Just blitzing through Martin's article (an excellent one by the way), I found that I probably hadn't highlighted an important point. I'm in no way advocating feature branches to enable development in isolation. Continuous communication in daily standup meetings, etc. are still critical. Also, keeping feature sizes small to prevent long gaps between merges are also critical. If you want to develop in isolation, you need to be explicit about your modularity - use something like Tracer Bullet Development.

Thanks (I think?)!

I certainly agree with your statement about the pain of merge going up with multiple projects, repositories, etc. I think the key here is to use good judgment about your repository structure. If such a complex structure is mandatory for some reason and cannot be adjusted, feature branching is probably not a good idea. That said, I think you'd be hard pressed to deliver on some of the things that I mentioned with any strategy coupled to such a complex repository structure.

Just blitzing through Martin's article (an excellent one by the way), I found that I probably hadn't highlighted an important point. I'm in no way advocating feature branches to enable development in isolation. Continuous communication in daily standup meetings, etc. are still critical. Also, keeping feature sizes small to prevent long gaps between merges are also critical. If you want to develop in isolation, you need to be explicit about your modularity - use something like Tracer Bullet Development.

Thanks (I think?)!

I certainly agree with your statement about the pain of merge going up with multiple projects, repositories, etc. I think the key here is to use good judgment about your repository structure. If such a complex structure is mandatory for some reason and cannot be adjusted, feature branching is probably not a good idea. That said, I think you'd be hard pressed to deliver on some of the things that I mentioned with any strategy coupled to such a complex repository structure.

Just blitzing through Martin's article (an excellent one by the way), I found that I probably hadn't highlighted an important point. I'm in no way advocating feature branches to enable development in isolation. Continuous communication in daily standup meetings, etc. are still critical. Also, keeping feature sizes small to prevent long gaps between merges are also critical. If you want to develop in isolation, you need to be explicit about your modularity - use something like Tracer Bullet Development.

Thanks (I think?)!

I certainly agree with your statement about the pain of merge going up with multiple projects, repositories, etc. I think the key here is to use good judgment about your repository structure. If such a complex structure is mandatory for some reason and cannot be adjusted, feature branching is probably not a good idea. That said, I think you'd be hard pressed to deliver on some of the things that I mentioned with any strategy coupled to such a complex repository structure.

Just blitzing through Martin's article (an excellent one by the way), I found that I probably hadn't highlighted an important point. I'm in no way advocating feature branches to enable development in isolation. Continuous communication in daily standup meetings, etc. are still critical. Also, keeping feature sizes small to prevent long gaps between merges are also critical. If you want to develop in isolation, you need to be explicit about your modularity - use something like Tracer Bullet Development.

Amazing how simple principles keep popping up all over the place - "There's no such thing as a free lunch!"

To add to the book recommendations, try out stuff from Alan Shalloway (particularly Lean-Agile Software Development) and anything by Mary and Tom Poppendieck. Also I still really like Ken Auer's XP Applied.

Jared,

OK, great. Now I want you to compare this to the iteration-less approach (or do you want me to do that? :-).

Matt