Today while doing some research for my open source momentum project I found myself comparing the various cloud computing platforms (Amazon EC2, Google App Engine, etc).  Cloud computing was all the rage in 2009 and I presume more of the same for 2010.  I am in the market for a platform on which to learn Erlang.  I stumbled across some nice Erlang AMIs at Hypothetical Labs so I decided to try EC2 for a while.  As my shiny new Erlang server booted up for the first time I began to wonder… would it be possible to play games in these clouds? 

Here is what I am thinking:

• Games (wars) are open to any number of teams.
• Each team can choose any cloud platform and any technologies they want with which to participate.
• Each game will have a goal or series of goals agreed upon up front that are somehow relevant to the real world (solve a hard problem, create new technology, etc)
• Each team can strive to achieve the goals using their cloud any way they wish.
• Each team can strive to prevent competitors from achieving the goals anyway they wish (except for DoS attacks of course!).
• All technology employed in the game must be open source.  New technologies created during the game must be open sourced within 6 months of being introducted into the game.
• Each team must fund its platform\technology with money earned using that platform\technology.  So for instance, if my Amazon cloud costs $1 per day per server, then my cloud must earn 1$ per day per server via donations or ad revenue or service offerings, etc. 

We would have to formalize these rules and I am sure there are others but this might be a fun way to play with open source technologies and combine them in different ways while trying to see what  some of these cloud platforms are made of. 

Anybody want to play?

Related posts:

1. Building Momentum with Linux and Erlang

The idea was to use ASP.NET’s nicely provided Membership services to register users and then use that for Authentication filtering.  The only problem my is that my domain model doesn’t include the ASP.NET database mappings, nor did I want it to.  I don’t want it to because if I later want to use Active Directory, I want it to be easy.

The solution I ended up implementing simply did the following:

I added two fields to the Register User View, First Name and Last Name, and then altered the controller to pick up this data.

Here is the controller code.

[AcceptVerbs(HttpVerbs.Post)]
public ActionResult Register(string userName, string firstName,
string lastName, string email, string password, string confirmPassword)
{

ViewData["PasswordLength"] = MembershipService.MinPasswordLength;

if (ValidateRegistration(userName, firstName, lastName, email, password, confirmPassword))
{
// Attempt to register the user
MembershipCreateStatus createStatus = MembershipService.CreateUser(userName, password, email);
if (createStatus == MembershipCreateStatus.Success)
{
FormsAuth.SignIn(userName, false /* createPersistentCookie */);
Person personToCreate = new Person();
personToCreate.UserID= (Guid)Membership.GetUser(userName).ProviderUserKey;
personToCreate.FirstName = firstName;
personToCreate.LastName = lastName;
EntityProjectTrackingRepository entities = new EntityProjectTrackingRepository();
entities.CreatePerson(personToCreate);
return RedirectToAction("Index", "ProjectTracker");
}
else
{
ModelState.AddModelError("_FORM", ErrorCodeToString(createStatus));
}
}

// If we got this far, something failed, redisplay form
return View();
}

So, now I have linked the user to my site by recording the Guid ProviderUserKey from the Membership services and updating my database and entity domain model with a Guid field named UserID.  Notice that I also explicitly have to create an entity model object to add the data to the MVC application database.  Well, the ASP.NET AccountController uses what is called Bastard Injection anti-Pattern for Unit Testing and Reflection purposes.  This signature looks like this:

public AccountController()
: this(null, null)
{
}

The problem here is that if I try to use Dependecy Injection (a subject for another time) to access my Repository, none of the turnkey unit testing can find the controller.  I’m still unsure about how this works and for the sake of time, accessing the entity model explicitly works just fine.

Now, I can track what user creates what resource (person or otherwise), and I more filtering options for the whole project.  Next step… Adding the Authentication Filters and User Roles.

No related posts.

Here are some of my other resolutions for 2010:

• Give up drinking soda entirely.
• Learn a new language (erlang) and use it to create a prototype implementation of my open source project (momentum).
• Do pushups every day.

Happy New Year!!

No related posts.

I am super excited to build the first momentum prototype using Erlang.  Reading about the hot code swapping feature was enough to get me hooked.  For the platform I have chosen linux.  There is a new service called Suse Studio (fantastic web site) that makes building linux images easy even for an old Windoze guy like me!  Joe Freeman put together a nice blog post about using this service to setup an Erlang environment.  At this point I am thinking that with a touch of couchdb, mochiweb, erlyweb, erlware, etc one could create a pretty neat linux live cd to install or host a momentum instance\server\node.  If you are interested drop me a line and I will send you the download link to my current development image.

So this is the technology stack for now.  I held a vote with all the project team members at this point (me) and it was unanimous.  Until requirements emerge that lead us away from this platform it is time to have some fun.  Armed with Programming Erlang most of my blog posts for the new year will document my adventures with these tools as well as explore how we can build a system that is capable of calculating the momentum of a software project in a meaningful way.

Related posts:

1. Cloud Wars the Game
2. Building Momentum
3. Another Year in the Books…

Jimmy Nilsson wrote a great post yesterday entitled The big picture of software development in which he says,

If we are in control of the code, then the business people can come with a request for a new feature that will create a new business opportunity and we can achieve it in a matter of, say, a few weeks. A couple of changes in the codebase might translate into millions of dollars. That codebase leads to high business value and the cost for achieving it is low.

This statement captures the ideal for any company.  A few weeks worth of changes could potentially reap millions of dollars in market opportunity.  This is where we all want to be and where few very of us actually are.  This statement is only valid for a project with momentum; a project that is healthy and can be moved from point A to point B along a given vector in a consistent and predictable way.

Over the next few weeks I will delve into slightly more formal definitions of mass and velocity and propose ideas for how these terms can be applied to real world software projects.

Related posts:

1. Software Development: Having Fun vs. Doing My Job

No related posts.

I wonder if it would be possible to formalize this feeling into something that can be modeled, measured, reported and ultimately used to bring transparency to software projects. I started an open source project on github.com today to play around with this idea. Check it out!

Related posts:

1. Another Year in the Books…
2. Building Momentum with Linux and Erlang

Hugh MacLeod suggests in Ignore Everybody: and 39 Other Keys to Creativity that I have created a mental pillar to hide behind which provides  opportunity for me to disengage from a creative and possibly productive activity such as writing.  He is exactly right.

We create pillars in the software industry all the time because software systems are complex and not predictable and our craft is at its core still a largely creative endeavor.  Personal pillars like mine above save us from putting ourselves out there to expose our creative (or lack thereof) inclinations.  Organizational pillars such as a policy preventing the use of open source software exist; well I am not sure *why* they exist but they do.

The complexity inherent within any software project leads to one of the most common pillars which I will refer to as requirement explosion; otherwise recognized as the phenomenon that prevents you from finally starting that open source project that you have been thinking about for months because you first need to come up with a project name and a description with which to create your github.com account (or should it be google code or sourceforge?) but before that or at the same time you should have a project website ready that provides your vision for which you will need to acquire a domain name that ideally matches the github.com name (that you haven’t created yet) as well as the visual studio solution name which by the way you probably need to create ASAP so users have something to download once they stumble upon your project.  Requirement explosion is paralyzing as we fear starting the first task without fully understanding the necessity and order of all the tasks and thus we do not move forward.

The spirit of this example actually occurs in some form in almost every project and team I have worked on.  No matter the level of technical talent, no matter the brute force hours; pillars exist that prevent us from excelling at what we do best; using technology to create value for our users.

Towards this end and in the interest of getting things done I intend to identify pillars created by me and around me and tear them down.

No related posts.

Particularly compelling was an example of the unacceptable cost not of creating smart building technology but of deploying (and supporting and troubleshooting) the network topology required for such technology using our current IP stack. Internet 0 is solution that changes the game.

I took this video as a subtle reminder that we as software developers must continue to hone our craft. Microsoft.NET, Java, Ruby; these are tools that we use to solve problems today but they will not be the tools we use tomorrow. Sometimes I catch myself thinking; how can I solve this problem with a Microsoft.NET application?. I should be thinking; how can I *best* solve this problem? Maybe that means spending a little less time becoming an expert in a particular tool and a little more time becoming an expert in delivering solutions that provide tangible value to customers.

The future looks incredibly promising for our industry; there is so much interesting work to be done that we can’t afford to stop learning for even one day.

Related posts:

1. Software Development: Having Fun vs. Doing My Job

I signed up for an account and created a first cut CV which you can see here:

http://careers.stackoverflow.com/zacharyruiz

This site is definitely one to keep an eye on!

No related posts.