Degrees of architecture

I sometimes find that there is a bit of confusion or fustration when it comes to Enterprise Architecture or Architecture over all.

It is often that this comes out in “The business doesn’t let me in to do analysis on their business”, altough this may be true in some cases the fundimendal question I often find myself asking is, What kind of architecture is the Entreprise needing?

There is sometime appropriate to only do Technology Architecture rather then the full stack, sometimes you need to do it backwards, start with projects /opportunities. There isn’t one structure way of doing this.

The biggest thing we as Architects (Enterprise, business, information, technology) will need to realise is the business will be in diffrent places within the maturity when it arrives on our desk, the project will be in a diffrent maturity places along with the team members.

So what do we do? 

We aim to deliver value as soon as possible, avoid analysis paralysis and do just in time deliverables. While doing this we go the extra mile and demonstrate value in other domains.

As  an example:
Working on a fairly major migration program I realised that the business was looking to me only for guidence and roadmaps on the technology landscape including risks, gaps and issues. While delivering this I used the information gleemed from the technology analysis to also produce a Business Services impact Information and this demonstrate how this technology landscape will impact the services the business offers to internal and external customers.

This allowed me to open the door to have a more meaningful conversation with the business itself on the coming changes and help them build a plan for training, communications and mitigations to the impact the migration project was going to cause.

So in summary: 

Understand what kind degree of architecture the program, business or the deliverable requires, build this value and aim to go over and above this by demonstrating other thought points and remember picture is a worth 1000 words.

Starting Enterprise Architecture

When starting an enterprise architecture, this can be a daunting task.

* How do I get started?
* What framework should I use?
* What templates should I use?
* What should I collect?
* Why isn't the enterprise waiting for me to complete my analysis?

All of these questions are valid and understandable, however, the answer to them may not seem straightforward or the one you may want to hear.

So how do I get started?
As you may have read from may other blog posts or information on the old interweb the key to doing a successful architecture is in small influential chunks or to eat the elephant one bite at a time (apologies to all elephants).

So, start with one of your projects and don't be too strict on following one process, ensure you are agile to adapt to the speed of the project or the change. Avoid going analysis paralysis or to much detail.

The aim of the game is to provide value as soon as possible.

What framework should I use?
In short, all of them. They are all great in their own way, TOGAF, ZACHMAN (as the main ones) are brilliant frameworks, however, they are also all-encompassing and hence they all say "tailor us to your specific use".

So if you understand a process in Zachman more then you understand a process within TOGAF then use the Zachman process, use what is the best of each framework to your advantage.

What templates should I use?
For me, the answer is NONE. Why? Because you may utilise effort in getting information that is not of value to you at that point in time and therefore can drop into analysis paralysis as you gather information to fill a template.

So what do I do then?
Well, start simple and then expand. Start with a spreadsheet with ID, Name, Description as a starter for 10.

Either start from the top down, bottom up or both, whichever provides the best value to the organisation.

As an example I choose to start with both (top and bottom) as I was working with some stakeholders from strategic and planning view point and then services mangement from the technology layer view point.

So I started with simple catalogues (lists) of strategic objectives linking this to IT projects and then roadmap of services.

And again I didn't go to deep into things and just started with ID, Name and Description then expanded it when I needed to.

As an example, if I am talking to senior management they are not overly interested if the Network Services is delivered over PSTN, ADSL, Fiber they only care about is it fit for business need and is it cost effective?

Why isn't the enterprise waiting for me to complete my analysis?
There is a trap here that is easy to fall into here, where do you stop, do I go all the way from strategy to port on a switch.

Most likely reason is that your enterprise wants to get on and deliver the outcome and to do this it needs to move on the project so to gain influence and show value you need to understand what is just enough analysis. Once you have established that and built the influence from your team or management then you can start influencing the analysis phase and how much you capture.

Technical Debts

What is technical debt?

Well for me its the following:

• Any code / product written in a hurry.
• Any code / product written with out tests.
• Any code / product written with out peer review.
• Anything made as a work around.
• Anything made with the attitude of "fix it later / post"

I am not saying that technical debts don't happen, they do and the bigger the project the bigger the technical debt.

Just remember to document it and keep it available as it has to be paid back sooner or later.

And just like with any debt financial or otherwise try and pay it back as early as possible as the interest will continue to grow and payback will be harder.

With fix it later its been my experience later never comes. And work around...... Well they become the fixes.

Is there away of avoiding technical debt?
Not really. Technical debts have the habit of creeping in when we are in a tight spot in our delivery of the project where quality is diminished in favor of meeting the deadline and it usually is a team unconscious decision.

Continuous intigration best practice.

Continuous integration is a hot topic now a days and is one of the best new tools for the software industry in resent years.

With that I would like to share some of my continuous integration best practices that I employ with in my development environment.

• Don't keep all the build code in one build file. Specially true when using ant. Ant is very powerful but can be very hard to debug when everything is a one big file. Split it into multiple files.
• Keep everything under source control.
• Keep one job doing one thing. It is better to have multiple jobs then one big job. This shortens the feedback cycle to the developer as they don't have to wait for a long job to finish or a very heavy processor job to know the compiler failed to compile the code due to a semicolon (;)
• Build the binary once and then fingerprint and reuse. This will help you in identifying where the files are being used in what jobs and you can then tie them together if needed.
• Tag every successful build with in the source control. This will enable you to get the latest good build for release.
• Automate everything. Not to say people aren't great they are just "unexpected" feature prone.
• Make sure the continuous integration server has a notification mechanism such as lava lamp, extrema feedback panel or email.
• If your business process allows automate release cycle.

Hope this will help someone with their continuous integration and release the awesome power of this practise.

Learning to be a software developer.

Most people think that if one knows how to code if,while,switch, ect. and difference between public and private that makes one a software developer.

Well I disagree, limiting one to that knowledge makes one a coder. I spend alot of time in theories and supporting tools then in actual code.

Don't get me wrong knowing how to code is important but theory of OOP (object orientated programming) or FP (functional programming) is even more important.

Here is my list of things one should aspire to learn to be come a great software developer.

1. Learn about build / project tools like Ant, Maven and many others this will enable you to help your team to achieve automation when delivering good quality software.
2. Learn about support tools like findbugs, code coverage tools, pmd this will help you to write good code and maintain standards.
3. Learn about continuous integration tools like hudson, jenkins.
4. Learn about design patterns good source is http://sourcemaking.com/ this will help you in making good source code and API.
5. Learn about refactoring same source http://sourcemaking.com/
6. Learn about source control cvs, svn, git
7. Lean about agile (xp programming, scrum)
8. Learn about project management, release and deployment. Just because you don't necessarily do it, knowing how is a great asset.
   

The above will make you a great software developer and will be a great asset for you when entering the software developer / engineering market.

Configuring CVS pserver on opensuse or SLE.

I am posting the how to configure the pserver for opensuse or SLE *nix installations so I can find it again with out trawling through many different sites that have very general information about this but also hopefully help someone else with the same problem.

1. Install CVS using yast.

2. Open up a terminal and login as root.

3. vi /etc/xinet.d/cvs

4. Change disable = yes to no

5. Make sure the 'server_args' attribute is pointing to your cvs repository i.e 'server_args = -f --allow-root=/usr/local/cvsroot pserver'

Now start/restart the xinet.d with either rcxinet start / rcxinet restart.

This will now enable your pserver to work and you can see it running by looking at your /var/log/message you should see something like

Mar 1 17:43:18 022656 xinetd[24151]: Reading included configuration file: /etc/xinetd.d/chargen [file=/etc/xinetd.conf] [line=26]
Mar 1 17:43:18 022656 xinetd[24151]: Reading included configuration file: /etc/xinetd.d/chargen-udp [file=/etc/xinetd.d/chargen-udp] [line=14]
Mar 1 17:43:18 022656 xinetd[24151]: Reading included configuration file: /etc/xinetd.d/cups-lpd [file=/etc/xinetd.d/cups-lpd] [line=15]
Mar 1 17:43:18 022656 xinetd[24151]: Reading included configuration file: /etc/xinetd.d/cvs [file=/etc/xinetd.d/cvs] [line=11]
Mar 1 17:43:18 022656 xinetd[24151]: Reading included configuration file: /etc/xinetd.d/daytime [file=/etc/xinetd.d/daytime] [line=16]
Mar 1 17:43:18 022656 xinetd[24151]: Reading included configuration file: /etc/xinetd.d/daytime-udp [file=/etc/xinetd.d/daytime-udp] [line=14]
Mar 1 17:43:18 022656 xinetd[24151]: Reading included configuration file: /etc/xinetd.d/discard [file=/etc/xinetd.d/discard] [line=15]
Mar 1 17:43:18 022656 xinetd[24151]: Reading included configuration file: /etc/xinetd.d/discard-udp [file=/etc/xinetd.d/discard-udp] [line=14]
Mar 1 17:43:18 022656 xinetd[24151]: Reading included configuration file: /etc/xinetd.d/echo [file=/etc/xinetd.d/echo] [line=15]
Mar 1 17:43:18 022656 xinetd[24151]: Reading included configuration file: /etc/xinetd.d/echo-udp [file=/etc/xinetd.d/echo-udp] [line=14]
Mar 1 17:43:18 022656 xinetd[24151]: Reading included configuration file: /etc/xinetd.d/netstat [file=/etc/xinetd.d/netstat] [line=15]
Mar 1 17:43:18 022656 xinetd[24151]: Reading included configuration file: /etc/xinetd.d/rsync [file=/etc/xinetd.d/rsync] [line=16]
Mar 1 17:43:18 022656 xinetd[24151]: Reading included configuration file: /etc/xinetd.d/sane-port [file=/etc/xinetd.d/sane-port] [line=12]
Mar 1 17:43:18 022656 xinetd[24151]: Reading included configuration file: /etc/xinetd.d/servers [file=/etc/xinetd.d/servers] [line=14]
Mar 1 17:43:18 022656 xinetd[24151]: Reading included configuration file: /etc/xinetd.d/services [file=/etc/xinetd.d/services] [line=14]
Mar 1 17:43:18 022656 xinetd[24151]: Reading included configuration file: /etc/xinetd.d/systat [file=/etc/xinetd.d/systat] [line=14]
Mar 1 17:43:18 022656 xinetd[24151]: Reading included configuration file: /etc/xinetd.d/time [file=/etc/xinetd.d/time] [line=17]
Mar 1 17:43:18 022656 xinetd[24151]: Reading included configuration file: /etc/xinetd.d/time-udp [file=/etc/xinetd.d/time-udp] [line=15]
Mar 1 17:43:18 022656 xinetd[24151]: Reading included configuration file: /etc/xinetd.d/vnc [file=/etc/xinetd.d/vnc] [line=15]
Mar 1 17:43:18 022656 xinetd[24151]: xinetd Version 2.3.14 started with libwrap loadavg options compiled in.
Mar 1 17:43:18 022656 xinetd[24151]: Started working: 1 available service

Next to consider is authentication. There are two options for this and that is either system authentication and that will be users on the box you are using or cvs's own passwd file authentication.

If you would like to use the system authentication then just add systemAuth = yes into your /etc/xinet.d/cvs file as we did above and login to the cvs server using cvs -d :pserver:username@host:/cvsroot login.

If you would like to use the passwd option then we need to make sure some requirements are availible.

1. Perl is installed on your system if not install it with yast
2. A user that has access to manipulate the cvs repository (from here on will be called cvs)
3. Logged in as root

Next navigation to your CVSROOT and create a new file named passwd and crypt.pl with 'touch' command.

Next make sure you have perl installed on your system and copy the crypt.pl code from this site (thank you Marty Phelan.) and store in your crypt.pl file. Make the file executeable with 'chmod +x crypt.pl'.

Next run the crypt.pl file with 'perl crypt.pl ' change to the password you want. This will encrypt your password to be used with in passwd file.

Copy the out put of the encryption and open the passwd file i.e vi passwd.

In the passwd file you determain the username password and the user that can access and alter cvs repository i.e
einarhreindal:asd032kasdf2:cvs this translates into username:password:user that can acces and alter cvs repository.

Save the file and check you can login to the pserver with cvs -d :pserver:username@host:/cvsroot login if you have any problems with login in then check the /var/log/message for any debug info.

Great site on how to create a repository can be found here.