Gerd’s Blog

Yesterday I finished the Scrum Master training and I have to say that I am a little bit disappointed. Our trainer went to some other issues of the scrum process itself but as yesterday there was nothing really new to me.
At the end of the course we did the Velocity Game with about 90 participants split into 15 groups. They gave us User Stories for the Product backlog e.g find one missing card in a deck of cards and we as a team had to estimate them first. Afterwards every team had to choose an Productowner which was me in ours. He sorted the product backlog in an way he likes the team to do the work. Every Story had an business value and of course the Product owner has to care to get the highest out of his group. After everything was set up the team was asked to estimate their velocity. Then the the team had to solve the User Stories in an given amount of time like a Sprint in Scrum. When the work was finished every group found their actual velocity, new User Stories where handed out and everything started from the beginning.
I really liked the game because everyone got an feeling how it will work in real life. Special in our group this was not new to most of us but it was fun. In the end there was no test like for other Certifications and we all passed.

If you wonder why I am posting this in the very early morning 5:30 ECT. This is an really funny story (at leas in the end). I left from Aarhus as planed by bus to the Airport. In the bus there where several participants from the JAOO conference so we had the chance to talk a little bit about our experiences. As I arrived at airport check in the nice woman told me that I am not allowed to check my luggages through to the flight to Vienna. I only had 40 minutes to reached the next flight to Vienna and as some of you might know the International Terminal in Copenhagen is about 1km away from the local one. So I already felt that it will be hart to catch it. So I went to the boarding area and surprise surprise - 30 minutes delay. Now it was clear that I will never catch the next flight. I came to Copenhagen Airport with about 1 hour delay. The thin that was nice about it was that Eric Evans was on the same flight and we had time for an retrospective on JAOO 2007. As I arrived on the Airport everything started getting complicated. As I booked the flights separately. There where big troubles getting my flight changed to an other one. Finally they did it, maybe also because I have a Miles & More Card. So my flight is going in about an hour therefor I stayed at the airport the whole night. For me it is really interesting how the airports are designed. There drink automates only take cash which is a little bit tricky because if you just arrive you normally may not have DKK. The Chairs in the Transfer Terminal and over the whole airport are just a little bit to small that you can take a sleep. Internet is just free because the JAOO account is also works here. So there is a lot of stuff Airport Planers may consider if they would have been in an situation like me before they design the airport.

So I have just returned from the first day of the Certified Scrum Master Class. Our Trainer is Gabrielle Benefield. First of course we had to introduce our self and I was impressed because most of the people in the room had already knowledge about Agile Processes and in particular about scrum. She started with some introduction and after that it was up to us which particular part of the process we wanted to dig deeper into. First the participants wanted to know more about estimating so we did some card estimation game and it was really funny and also astonishing that all five teams came up with nearly the same estimates.
After lunch Gabrielle talked about which planing should be done including Vision, Road map and Release Plan and all the roles that are involved in the process and step deeper into how we can Track the Progress.
There was nothing really new to me in the calss until now and I also don’t think there will come some tomorrow. This is not bad the much more interesting stuff are the discussions with the participants and some Retrospectives what worked and what did not for them. I do not think Gabrielle had a class that was already that familiar with the process until now so when I spoke with her after the class she was really amazed about it.

At least the Scrum Master Class reminded me that maybe we stepped a little bit to far from the original scrum process in our company. I will have to think a little bit about this but as it looks to me until now we should take a step back. I think the time for doing so is perfect at the moment as we are starting some kind of bigger refactoring and advancement for the InTime suite.

Today was the last conference day for JAOO 2007 and I will visit the Scrum Master Class for the next to days. No Keynote today so even Track Host Frank Buschmann was a little bit confused and came to late to open his Track. Never the less the first speaker was Wayne Fenton the Director of Architecture form eBay. As I posted I met him already the first day at the Conference Party. He spoke about Operational Scalability and how scalability is on of the big requirements for his company. His talk was quit interested because he gave an overview how eBay reached the goal that their system scales very well. He explained a little bit about their production line and how they are doing deployment which I will post about soon.
The second Talk was given by Robert S. Hanmer about Fault Tolerance once again in the Architecture Quality Panel. Robert is working for Alcatel-Lucent where he developed Software that helps to achieve 99.999% of availability by designing fault tolerant systems. The systems need a possibility to detect faults. If they have detected one they are using different strategies form simple reinitialization up to whole reboots to bring the system as fast as possible back to a working state.
After lunch, which is a little bit to South African like for me this year, Kevlin Henney talked about Performance Art. The main message was that Performance matters in every system but as one person said at the end - It does not matter what Kevlin is speaking about, his talks are just great.
The last Talk I visited this day was by Rebecca Wirfs-Brock about Lessens Learned in architecture. She is an consultant and had the chance to see several different architectures. It was interesting that she spoke not that much about the architectures it self further more how she is dealing with the challenges in projects as consultant.
At the end there was an Panel Discussion. What I take from it is normally the design or architecture of a product is not bad at the beginning. I mean nobody want to build a system he cant extend and maintain in the future. Further more there is a leak in communication so that developers do not understand the architecture and therefore the the code gets to a big ball of mud. There was one quite interesting comment form a guy of the audience. He admitted that they are doing architecture documentation on small movies. Several on the panel and I like the idea. I have to try it out.
The finishing panel was held as Podiums Discussion by Martin Fowler, Diana Larsen, Erik Dörnenburg, Erik Meijer and Eric Evans. They did some retrospective on the conference and what they fascinated them.

At the end I have to say the three conference days of JAOO 2007 where grate. I had a lot of fun and learned several new things. I am just wondering why we got no Conference T-Shirt this year but at leas everyone got a bag. So, now I will go out with Rupert for some beer to discuss the last two days a little bit. I am really thinking if I will come to this conference next year again. Definitely it is the best i have ever been to, but lets see maybe we will be able to do something like the JAOO in Vienna for the next Year.

Yesterday there was day to at the JAOO 2007. The Keynote was held by Eric Maier who presented his vision of programming in the future. The presentation was about bringing all the bloody code away from the Programmer. With bloody code he means code that has nothing to do for solving the current project.
The first talk of this day was by  Jean Bezivin who talked about model transformation At the beginning he started explaining Models Meta-Models and al the theoretical stuff that in my opinion most of us already know. This part was a little bit boring. In the second Part he was switching to the Eclipse Model Discovery project they are doing the transformation with. The presented stuff was quit interested, the presentation technique was like a professor at the university (see one of my last postings).
After lunch we went to  Eric Evans talk about  strategic design. His conclusion was that you have to find your main model and build your business around it. I was thinking about the main model of the system we are currently developing - maybe I should spend a little bit more effort on this in the future.
Next talk was once again by Klaus Marquardt, this time more technical then the last one. He was talking about Complexity Management in projects. His conclusion was that you have to be very careful with KPI’s Management are requesting.
Joseph Yoder spoke about The Adaptive Object-Model which was quite interesting because we use this technique for our Event Processing Engine too. Our Event Object Types are definitions of the Event Objects that are the runtime container for Events - the objects send through the system.
The last Talk at this day was a more practical one held by Jimmy Nilsson & Kim Harding Christensen about LINQ for Domain Driven Design. I know LINQ for some time now it was just nice to see how other people solve problems with it. At the end one of the Developers of LINQ had a small discussion with the two speakers why some stuff was build in the way it is now and which trade offs they had to agree to.
After all tat there was a so called “Sun-Keynote” which was more or less an marketing event form Matt Thompson and one Sun developer I even don’t remember his name. They showed some examples with JAVA Me and Sun SPOTS. I didn’t like tat kind of marketing events titled as Keynote.
In the Evening there was an other new Event called Jam Session. There where some pizza, beer and wine and several speaker and visitors played music. We met some smart people including Jimmy Nilsson. We were speaking about the situation in Software Development in Austria and that there is no conference like the JAOO in Central Europe. He asked if we are planing to do something against this. I am really thinking about this since last year I have been to JAOO. Lets see if we can do this.

Yesterday was Day one of the conference and was really interesting. Everything started as last year we got a nice little bag with all the stuff you need at an conference. Just one thing was much better this year, I had no problems getting my batch.
The morning keynote was held by Robert C. Martin and he is a really good speaker. At the beginning he was rising the question if software engineering is already a profession and appealed to every developer to write better and cleaner code. He is wearing a green band on his left arm for about 3 years now because it reminds him to do so.
The second talk we where attending was by Charles Simonyi about “Democratizing Software Creation” which was really fascinating. His company is developing an language to describe other languages which you can imagine is not that easy. The talk was given in combination with one person form Capgemini who showed an demonstration of an DSL generated within their language. I am still a little bit sceptical if this is really working the way they showed it and wanted us to believe it does.
Organically Agile was the Talk by Klaus Marquardt who is working for Dräger Medical. The purpose of his talk was that even life critical systems can be developed within agile methods.
Lex Spoon did an talk on Scala an language that is very interesting. The language combines Functional and Object oriented concepts and I would like to take a closer look on it in the near future.
Gilad Bracha is working on a language called Strongtalk which is used to build an Executable grammar. The exciting part here is that Lexer and Parser can be easily extended because of flexible concepts like mixins that Smaltak provieds.
The last talk before the Party Keynote was by Andreas Zeller and titled Beautiful Debugging. He presented his Ideas written in the Book Beautiful Code. The book as the talk I can only recommend you.

Party Keynote was given by Robert C. Martin again . He was speaking about the Space flight he did this year. The Pictures he showed where just amazing special the ones taken at the ISS. Take a look at www.charlesinspace.com to get some of the expressions we where able to see.
After the Keynote there was the Conference Party as last year. We met some people there e.g Joe Hummel, Wayne Fenton and Roy Osherove, all three are speakers at the Jaoo. We had some nice conversation special Wayne Fenton spoke a little bit about who Ebay does roll-out. I will make an other posting about that later.

Today I have arrived at the JAOO conference in Aarhus. I am visiting the conference with a good friend of my Rupert. On the bus form the airport to the town we met with Rajeev Dayal one of the Google GWT Developers. He will held a Talk about “Building Large Applications with GWT 1.4″ on Thursday. After we arrived at the hotel we went to lunch together and spoke about working for Google developing the GWT, Rich client applications for the Browser and several other things very interesting things. He gave us some information how development is done by Google and I have to say I now understand much better why things like the blog search issue I mentioned in one of my last postings can happen there.

Right now I am sitting in my hotel room and looking at the schedule for tomorrow. The opening Keynote will be held by Robert C. Martin and the Party Keynote by  Charles Simonyi. In the evening there is the Conference Party so hopefully I will find time to post about the talks before the it starts.

Today I want to post about something different than software development. Yesterday I found the Video about Rand Pausch’s last lecture and was really impressed. It was not about a person that still has fun living even knowing that he will die, further more about a university professor that dedicated his live teaching students to reach their child hood dreams.

A full version is also available.As some of you might know I am still studying Computer Science at the TU Vienna. I started in 2000 and no I was not just studying the last 7 years. In September 2002 I decided to work full time and this is what I am doing until now. Therefore my progress at the University is not as fast as expected.

There are several reasons for starting to work full time but one of the biggest was that studying was more a burden than a pleasure and I do not mean it in an hey lets go out and have fun way. After nearly seven years at the TU including several years working as Tutor, I have to say there is and was no Professor who fascinated me in any lecture he gave to aim for the stars in a way the talk above does. There was neither a person I would

There was one sentence in the full video that fascinated me most:
When he got his PhD his Professor told to him that he is an excellent seller and he should use this technique to sell something more valuable than any product of the world - he should sell education.

We are complaining that skilled, enthusiastic developers are so hard to find but forget that they have not learn it. Not in school an not at the university. Where are the Mentors the Rand Pauschs that can captivate students with their words?

Probably most of you know I am currently working for SENACTIVE and we are developing an Complex Event Processing System Called InTime in C#. Yesterday I had a meeting with an potential customer and he was asking what our platform target is. I said it is Windows as we are developed in C# and until now Mono does not support everything we need from the CLR. He was complaining about C# because he gave it a try a few years ago and was not satisfied and all their applications are running on Linux/UNIX. I found an interesting analogy that at least everyone that likes Winter sports can understand.

The opinions of JAVA and .net are similar to them of skiers and snowboarders. If you just try one of them you don’t like the others because they are different and you just see the bad things. If you are trying both of them you know about their advantages and disadvantages find both attractive.

It is really interesting that languages and Runtime environments still matter today or at least people are concerned about it.

In one of Alister Cockburns last posts he was writing about Using RUP to fix Scrum. As i read his small article one sentence attracted my attention. He wrote: The pendulum has swung too far from “too much planning” to “not enough understanding”. I have the same feeling for some time now. The picture on the left shows a typical iterative scrum process where the backlogs on the left side are used to add features to a new iteration. What’s interesting is that more and more people and companies try to hold the Product Backlog as small as possible and push new features directly into sprints. So what is bad with this approach you would ask?

The backlog was intended as a place where ideas can grow. I never have seen Stories/Features that can be developed as they are written in the Backlog. In my Opinion this happens because most of the people involved don’t have or take the the time to think about the feature that should be implemented. A short cycle to develop new features should be aimed but not for every price. There are some other aspects that need to be considered and one of is that people must understand the product they are developing.

The second issue I have seen with short Product Backlogs is that Release Planing suffers. If you do not have enough Stories in your backlog to plan for the next release you will ship everything breaks down. In this moment everybody in the Team looses the goal for the Product. I would compare it with a scene from Forrest Gump where he starts to run without any goal. This is not working in reality.

There are three lessons I have learned in the last three years in reference to the issues I mentioned above:

1. Never start any sprint without a goal
2. Never start a product development without an Release Plan
3. Try to discuss Stories/Features as soon as possible with a bigger group of people

As most of you will know Google Reader finally got a search function. I proposed such function also in an last post about Meta RSS feeds so at least for me it was no surprise. After so long time the feature I waited for was finally available so I started to do some demo queries. Searching for some words delivered really good search results. Then I started to try some of the more complex queries like searching for links or using the “site:” operator. None of the advanced operators worked. I really have not expected that and it looked to me they have done a new search engine just for the reader instead of using the old one that does already search for blog entires.

Today they posted an article in the goolgesystem blog with a few tips how to search for stuff in the reader which really makes me think. Why do we need a How to for a search if all the other different content searches (web, news, blogs) can be done with the same query syntax? Why they have invented the wheel again and done an new search engine just for the reader? Is Google getting to big? I do not know until now but it reminds me about several other big companies that are making this mistake over and over again.