jakub.dziwisz.org

I’ve just read a great article on pair programming by Jeff Langr on http://developer.com/. Being one of the most “spectacular” agile technics, pair programming is also one of most often abused (and in such degenerated form - criticized). The reason is simple - it’s very easy to do it wrong and don’t even notice that. Therefore I recommend reading this article even if you aren’t completely new to the topic. Have fun .

A few days ago I run to the following problem - in order to prepare test case for one of my classes, I needed to build a mock object. However, due to design issues I needed to mock not an interface, but a class. I quickly tried EasyMock and jMock but both of them failed, complaining that they cannot mock classes. I very briefly googled for a solution, but didn’t see any and was almost ready to start writing My Own Ultimate Mocking Library… Unfortunately, a second later I was brought back to earth. Both those libraries offer extensions capable of dealing with this kind of problems (i.e. ClassExtension for the first one and CGLIB for the other) . Well, tomorrow will be better!

I’ve just uploaded yesterday’s presentation, it’s available in Download section. It’s a bit similar to the one given at CONFidence Conference, however I removed an introduction to grid computing and heavily extended part devoted to Shibboleth, GridShib and alike. Additionally, I added a few more slides on my ideas how could Virtual Organizations for Open Grids be built. I hope you’ll like it!

Tomorrow I’m going to give another presentation on security in grid computing, it will be held on Cracow Grid Seminar (its abstract is available on IT w Krakowie web page). Unfortunately I didn’t manage to prepare the slides earlier and I’m working on them now. Power Point crashed three times in last two hours. This makes me think that Open Office is not that much worse . Maybe I should get back to making presentations in LaTeX?

For the first time I heard about agile software development in 2002. It was obvious for me that many of these ideas are right. I adopted unit testing, refactoring, iterative development and a few other techniques. However, I could not imagine a project where, in my opinion, the most important concepts are obeyed - no fixed releases (in terms of provided functionality) and no fixed scope of project or no fixed deadline for the entire project.

Of course, for me as the developer it is a very good situation. If it takes more time than I thought initially to solve a problem, it can always be done carefully. Nevertheless, I could not imagine a customer that would like to go for such a deal. This would require a large amount of trust between two companies.

Recently I read Jay Packlick’s presentation on Agile Customer Collaboration (page is in Polish, but presentation in English; and here is a direct link to the presentation). Jay presents here a few very interesting ideas on Agile Contracts. For example Multi-Stage Contract, where the entire project is divided into a set of small releases. After each release companies choose whether to continue, re-negotiate or walk away. This gives both parties much control over the project’s course. Who knows, maybe it even works?

Recently I fell in love. With Ruby, a programming language. It has all the features I have been looking for since I started to learn programming. It has an elegant yet powerful syntax, a number of APIs and… is as simple as possible, but not simpler.

Yesterday I discovered Google SOAP Search API. Without any hesitation I decided to write a simple spell checker based on Google API. Of course, in Ruby. The script is only four lines long:

require 'soap/rpc/driver'
stub = SOAP::RPC::Driver.new("http://api.google.com/search/beta2", "urn:GoogleSearch")
stub.add_method('doSpellingSuggestion', 'key', 'phrase')
puts stub.doSpellingSuggestion(ENV['GOOGLE_KEY'], ARGV[0])

It assumes that in the $GOOGLE_KEY environment variable your Google key is stored. Additionally, the first program argument is used as a word (or a sentence) that is to be checked. Certainly, the script is too stupid to be really useful. But I think that it demonstrates Ruby’s power quite well.

Two weeks ago I saw a very interesting phone. Probably it is one of the first models where the rotary dial was replaced by a key pad. However, buttons placement is very clever!

What is so nice about that? Completely new functionality does not have to lead to completely new interface. It is even better to change the UI gradually, so that a user has time to get used to it. Then she has no problems with using the new product. Moreover, she is instantly able to take advantage of the new features. It is crucial to remember about that, especially when dealing with the bleeding edge technologies, like grid.

I just discovered that ERCIM News are available on-line, you may find it at http://www.ercim.org/publication/Ercim_News/. ECRIM stands for European Research Consortium for Informatics and Mathematics. It aims to foster collaborative work within the European research community and to increase co-operation with European industry.

Number 63 was particularly interesting for me as it is entirely devoted to security and trust management issues. One of the most interesting statements that I read there says that building trust and security in information society will play key role in 7th Framework Programme.

This sounds realistic especially after listening to what Wolfgang Gentzsch said during his todays lecture - most areas of grid computing are mature enough to support building real applications on top of them. We do not need to build new infrastructure, we just have to build applications. I could not agree with him more [well, it’s silly to argue with Wolfgang Gentzsch ], but I do also think, that we need more research on the topic of trust.

Because the systems available now are really secure. But is is not enough when dealing with a dynamic and distrubuted society, like grid users. In my opinion to satisfy this requirement we need authorisation frameworks based on dynamic and distrubuted reputation systems.

Sleeping while “driving” in a traffic jam is a very challenging task. Only a bit easier is reading a news paper. But finally I discovered a method how to save some time during my daily commuting.

Podcast is a fantastic alterntaive to traditional radio. I may easily choose program I’d like to listen to. And there are thousands of podcasts available. Of course, most of them is rubbish. But still there are many interesting. For example ITC: Security or Security Now! So now I’m not scared of traffic jams anymore .