• Forming - everybody is really polite and gets to know each other;
• Storming - while trying to cooperate, team members are confronted with different ideas, values, beliefs and assumptions they all might have;
• Norming - the ideas and beliefs converge to a shared vision;
• Performing - the team reaches a hyperperforming stage in which team members trust and depend on each other to get the job done.

To go through these phases takes a couple of Sprints, depending on the team members, sprint length, etc… etc… In summary, you need to invest to get a hyperperforming Scrum team.

Some teams include (or even exist entirely of) people who are hired specifically for the project at hand. Is it feasible to introduce Scrum in such a team? Is there a tradeoff between the team forming effort and project length?

Why would I pay the cost of introducing Scrum for short projects (less than, say, 3 sprints)? However attractive this shortcut may seem, I think it’s a pitfall you shouldn’t step into. Remember the environment that Scrum supports best: complex projects, with considerable uncertain technology and requirements. The best solution for complex problems comes from solving them together. To solve a problem together, you need to be a team. And each team goes through Tuckman’s phases. There are no shortcuts.

A team specifically assembled for a short project may never reach hyperproductivity. That’s a pity. But the solution delivered will still be better than it would have been in a command and control environment.

Technorati: scrum, tuckman, team, contractor

Years ago, Bill Wake coined the acronym INVEST to verify if you have the right kind of user stories. He said that user stories should be

• Independent
• Negotiable
• Valuable
• Estimatable
• Small
• Testable

I just read something about the new INVEST model, which changes the meaning of Small to Sized appropriately. Mike Cohn appears to use this new definition and I saw a few references on other blogs as well.

I’m not sure I agree. A product backlog is like an iceberg: user stories on the top should be ready for inclusion in a sprint, and user stories further down the road can become Big Ugly Epics. What is so INVEST about an epic? Nothing, IMHO. You can’t negotiate an epic. You can’t estimate an epic (sure, 100 points, XXXXL or a question mark will always do). You can’t test an epic. It might be independent and the product owner can probably put some business value on it. But that’s about it.

Instead, we should relate the definition of INVEST to the ranking of the user story in the backlog. For all stories that are ready for sprint inclusion, each story should conform to INVEST. No excuses. For stories further down the backlog, INVEST can be more loosely applied. As stories rise the backlog, they should become more and more INVEST until they are ready for a sprint.

I think that this distinction is better than watering down the definition of INVEST.

Technorati: INVEST, Scrum, user stories, product backlog

He argues that the information security officer should be involved in almost any project and he’s probably right about that.

More generally speaking, the information security officer is ‘just another stakeholder’ for any project. And for a project to be successful, it should pay attention to the needs of its stakeholders.

But… while there are many stakeholders, there is usually only one sponsor of the project. Does he consider ‘information security’ important to ‘his’ project? Who would have to convince him that he should reserve some resources to achieve an acceptable level of information security?

Seems like a perfect role for the information security officer. Even better if that officer is backed by a proper information security policy (I guess it isn’t a coincidence that Kees has an opinion about that too ;-)).

Doesn’t it all come down to the fact that we tend to overlook the Ilities in projects?

DVCS’s are ubercool and you can do some nice tricks with them. Nonetheless, not everybody needs these tricks (yet?) and integration with popular IDE’s isn’t nearly as good as with CVS or Subversion.

Most people who develop software regularly have a copy of TortoiseSVN installed and it’s perfectly possible to create a local, single user repository with just TortoiseSVN.

Here’s how:

1. Create an empty directory.
2. Right click it, choose TortoiseSVN - Create repository here.
3. Choose “Native filesystem (FSFS)”. You could use Berkeley DB as well, but FSFS is the default nowadays.
4. TortoiseSVN should report that it successfully created the repository.
5. Create another empty directory.
6. Right click it, choose TortoiseSVN - Checkout.
7. Pick the repository you just created, click OK and there you go!

You now have a version controlled directory that you can use as… any other version controlled directory with TortoiseSVN. It really doesn’t matter whether the repository lives on a remote webserver or on your local hard drive.

Technorati: version control, subversion, tortoisesvn

Did you ever need a file that lived on a Samba-share on your home network while you were out (work, school, customer site, whatever)? I do. Regularly. Maybe it’s just my fault and I should think more thoroughly which files to take with me on my USB stick. Anyway, I found out a way how to use my drive mappings transparantly at home and anywhere else by simply using PuTTY and an SSH connection to the server.

Samba uses TCP/IP port 139. You could open up that port to the internet, but it would mean an extra security risk. Besides, lots of ISPs block this port because of many known attacks (e.g. the Sasser worm). It would be more secure to only expose SSH to the internet and still be able to use your Samba shares at home.

To do that, we need four steps: install a loopback network adapter, configure it, create an LMHOSTS file and create the SSH tunnel in PuTTY. This walkthrough is mostly inspired by an old email message on the Samba mailing list.

Install a loopback network adapter

A loopback network adapter is kind of a “virtual network adapter”: it acts as a normal network adapter, but you don’t have to install any physical hardware to use it. It is only usable on the PC it’s installed on (other PCs can’t connect to it).

1. Go to Control Panel - Add Hardware
2. Click Next, then wait a while, choose “Yes, I have already connected to the hardware” and Next again.
3. Choose the last option, “Add a new hardware device”, click Next.
4. Choose “Install the hardware that I manually select from a list (Advanced)”, click Next.
5. Choose “Network adapters”, click Next.
6. Choose Microsoft, “Microsoft Loopback Adapter”, click Next and Next again.
7. Click Finish.

Configure the Microsoft Loopback Adapter

Since the loopback adapter won’t have a DHCP server to ask for an IP address, we need to configure an IP address manually.

1. Right click on “My Network Places” (on your desktop or in your Start Menu), click Properties.
2. Find the network connection that is associated with the “Microsoft Loopback Adapter”. Usually it’s called “Local Area Connection 2″. For clarity, I renamed mine to “Loopback” (what’s in a name?). Right click the connection name, click Rename and change it.
3. Right click the Loopback connection, click Properties.
4. On the General tab: disable “Client for Microsoft Networks” and “File and Printer Sharing for Microsoft Networks” (only untick the checkboxes, DO NOT uninstall!).
5. Click “Internet Protocol (TCP/IP)” and click Properties.
6. On the General tab: click “Use the following IP address” and pick a private IP address on an unused subnet (I used 192.168.100.100 with subnet mask 255.255.255.0).
7. Click Advanced.
8. On the WINS tab: click “Disable NetBIOS over TCP/IP”, then click OK and again OK.
9. Click Close.

Create or edit the LMHOSTS file

1. Open the file “C:\WINDOWS\system32\drivers\etc\lmhosts” (create it if it doesn’t exist - note that the file should have no file extension).
2. Add the following line to the end of the file:
   192.168.100.100 servernamehere
3. Save and close.

Configure PuTTY

At last, we have to tell PuTTY to create an SSH tunnel from the loopback connection to our home server. I assume you already have a PuTTY saved session to connect to your home server.

1. Open PuTTY, click your saved session and click Load.
2. Go to Connection - SSH - Tunnels.
3. In Source port, type “192.168.100.100:139″ (I know, the field is quite small, but it will accept the value anyway).
4. In Destination, type “localhost:139″.
5. Click Add, and do not forget to resave your saved connection!

To confirm that everything is working, connect to your home server with the saved session in PuTTY. Click Start - Run and enter “\\servernamehere\sharenamehere”. Depending on your Samba setup, you may have to enter a username and password. Done!

Background information a.k.a. Frequently Asked Questions

Why do I have to install an extra loopback adapter? Can’t I use the adapter that’s already in my PC?

Yes, in fact you could use the adapter that’s already in your PC. But… you would still have to disable the Client for Microsoft Networks and File and Printer Sharing. That means you won’t be able to use any other shares besides your home shares. Not very practical. An extra adapter allows you to have Microsoft Networking enabled on your normal network connection and use your home shares simultaneously with other (work, school, …) shares.

Do I really need the LMHOSTS file?

No, you don’t. You could refer to the IP address as well (e.g. connect to “\\192.168.100.100\sharenamehere”). But I wanted the solution to be transparant, whether I’m at home or not. If I use “\\servernamehere\sharenamehere” at home, then I want to use it anywhere else as well (or else I have to remember too many things). That’s what the LMHOSTS file is for.

Happy Samba-PuTTY-tunneling!

Technorati: samba, putty, ssh, tunnel, windows xp, lmhosts, port forwarding

Leontien, Martin and Emmy

This time I gave a presentation on Scrum and why I think Scrum is a major improvement as project management framework for software development projects. The slides of this presentation are available for download (dutch only).

After two hours, many questions and an extra cup of coffee everybody was quite enthusiastic about the whole concept of short iterations, potentially shippable product increments, team commitment and the like.

Technorati: scrum, project management, agile, whitehorses

On a personal level, I have seen Emmy grow from a little baby to a small girl who already walks around the house (a little shaky, but still). Leontien found another employer, Capgemini, where she will be very busy consulting public organizations about their business processes.

Professionally, I became a Prince2 Practitioner. Prince2 is the de-facto standard methodology for project management (founded in the UK, used worldwide). It was a 3-hour essay based exam, but I made it. I also certified myself as a ScrumMaster, which basically means I attended a two day training.

Scrum is a framework for iterative, incremental development in cross-functional, self-managing teams. I’m impressed and kept reading on the subject since (Agile Project Management with Scrum, The Enterprise and Scrum, Agile Estimating and Planning, Lean Software Development). As far as I’m concerned, this is the way to make software. Who needs waterfall anymore?

In the near future, I’ll definitely keep extending my knowledge and skills on Scrum and project management and try to share my findings here every once in a while.

Technorati: agile, prince2, scrum, waterfall, project management

All in all, a fresh and clean layout, suitable for personal and professional rants.

But… where are the rants? Shame on me! Let’s try to fill this blog with at least one post a month, starting now. Sometimes I ask myself if I’m not a writer, don’t have the time, don’t make the time or happen to be a victim of coincidences.

We’ll see.

Technorati: redesign, Anthony Baggett, wordpress, random quote