The next installment of the Autumn of Agile screencast series is available now for immediate download from the Autumn of Agile website.
Estimating Our User Stories
In this installment we wrap up our estimates on our User Stories, organize our like Stories into Features, and flesh out some of the actual tasks that we will need to perform to deliver the User Stories’ business value in the system. We also take a realistic look at the amount of effort that our estimates are starting to suggest will be required to deliver the project and start to evaluate our ‘must-have’ User Stories against our ‘nice-to-have’ User Stories.
Creating our Development Environment
We then take some time to setup our project work folders in our source control repository and also work through the configuration of our CruiseControl.NET continuous integration server to begin to familiarize ourselves with some of the infrastructure items upon which we are going to rely during the life of our project. We build only the most simple integration steps into the CI server at this point since we don’t need anything more complicated yet, with the expectation that we will enhance the capabilities of our CI server as our own software solution increases in complexity as well.
Still no Code
Sadly, by the end of this installment we still haven’t written any code, but that’s the whole point of Iteration 0 (which this installment now brings to a close) — providing us the breathing room we need to setup enough of our infrastructure to help ensure the execution of the project is a success when Iteration 1 (and coding) starts next week~!
As always, feedback, constructive criticism, and other input is welcome. Happy viewing.
Why “Vault” more than “Subversion” ? Subversion is not more common in the .NET development ?
PS : have a nice voting American people 🙂
I have nothing at all against Subversion 🙂 In fact I’ve used it myself in the past both in OSS projects and at other companies where I have worked. We just happen to use VAULT as our SCC of choice at Microdesk for a number of reasons.
I completely agree that Subversion might very well be the more ‘common’ SCC system in the world of .NET development (though I haven’t seen any actual ‘adoption numbers’ myself that would prove that ‘feeling’) but since we use VAULT internally, that’s what I’m going to be using in these screencasts (recall that these play a sort of dual role in both being inwardly-focused for our own staff as well as being outwardly-focused for the community).
There isn’t really anything that I am going to be doing with VAULT that someone else couldn’t do with Subversion. For every person who wants me to use Subversion, there will someone saying “why aren’t you using Perforce?”, or “why aren’t you using TFS?” or “why aren’t you using StarTeam?” or “why aren’t you using…!” (it goes on and on) as you can imagine.
With the exception of some edge-cases (and the universe of ‘distributed SCC systems’ like git or mercurial), most modern SCC systems today offer pretty much equivalent features at their core even if their implementation of those features is different.
Except for a very few ‘old-style’ systems that only offer ‘exclusive-check-out’ style operations and don’t provide support for the more common edit-merge-commit style of working, most features are common to most SCC systems (merge, diff, branch, label, rollback, whatever) and so I don’t think that my choice of VAULT for these screencasts is going to prevent anyone from either ‘following along’ or ‘following the guidance’ that these screencasts will be providing in some other SCC system than VAULT.
Recall when I started the Summer of NHibernate series that I said I would be unapologetically using a Microdesk-centric collection of technologies for the screencasts; I think its just that with the Autumn of Agile series you will see more and more of these. But the Summer of NHibernate was just as ‘opinionated’ in its choice of technologies (e.g., I used MbUnit instead of NUnit, xUnit.NET, CSUnit, MSUnit, etc., RhinoMocks in place of Moq, NMock, etc., and NDbUnit in place of an in-memory DB like a number of people like to do using SQLLite and so on).
With this series, you will get to see a wider array of my company’s “dirty laundry” since this project will run the gamut all up and down the technology stack, exposing our choices in many areas (e.g., CruiseControl.NET in place of the increasing-in-popularity TeamCity, for example).
If we reach a point where I’m doing something with any one of our technologies that isn’t in any way achievable using any of the other technologies, I hope you (and everyone else) will let me know so I can be sure that I don’t “lose” people due to that. Keep me advised 🙂
Thanks for another instructive series.
Is it possible for you to make your ccnet.config available as a download?
Hi Steve….It would be nice if we can get hold of cruise control config file.
I will make it available along with the next drop of the series (coming before the end of the year). As of now, the file is VAULT-specific and doesn’t yet run unit tests; once its wired up to use subversion and at least runs the tests it will be worth getting a hold of for viewers.