My in between jobs status won’t last for long as I’ll be starting a new job in a few days. My previous employer was acquired by a large firm (about 5000 people) over two years ago, and even though I had a nice working place (with regards to colleges and projects), I felt I needed to push myself by taking up a challenge and to learn new stuff.

So I’ll be working at Kalooga, a small start-up company involved in gallery search. “We” spider the web for collections of pictures, analyze them (by content, text, links, or anything that’s relevant) and enable publishers to show them in widgets on their own sites. This allows publishers to easily decorate their electronic publications with up-to-date, relevant visual content.

The upcoming year, we’ll work on the quality of search results, deal with the unavoidable scaling issues (half a billion pictures is quite a bit) and make sure our clients are well served with usable back-end applications & API’s. The job combines the domain Artificial Intelligence with some challenging web stuff (scaling with a cloud, Services, etc) and I think my background & experience will come well at hand. I’m looking forward to it!

Many software systems used to be shipped with simple default security settings, like no security at all. This open by default strategy is very usable; you can use every feature without hassle. However, if you want other people to stay out of your business and therefore need some protection, you’re supposed to change these default settings into something that fits your environment (like being connected to the internet or having curious neighbours/coworkers/friends with access to your system). As most people don’t read the manuals of the systems they install (I don’t), this strategy leaves us with a lot of open and insecure systems.

Nowadays, the closed by default strategy seems to be favoured by vendors. Only if you specifically allow features (like services) to be accessible, you can access them. This tends to make the world a little bit more safe.

Having default passwords for a system is slightly related. Without any effort or provisioning, you can have a default admin password in case you’ve locked yourself out. It aleviates the lifes of many help desk workers. You can just read the documentation or google a bit to regain access. You don’t remember the password for your MySQL installation? Just google it! On the other hand, if you’re forced to always define your own password at installation time, or the system generates a random one and informs you about it, it is more secure, but you can lock yourself out if you forget about this password.

Like always, there’s a balance between accessibility and usability.

The reason I write about this is that the vendor of the key-pad locker locks in my gym delivers locks with pre-defined master keys. Nice, because without any effort, the trainers are able to open locks for people that forgot about their user-defined code. However, it just takes one google query to find the manual that tells us that: the factory code is 1,2,3,4,5,6!

Naturally, my gym hasn’t read this manual or isn’t aware the manual is this easy to find. It’s very accessible for the support staff in case a user has lost it’s security code again, but sadly doesn’t provide me with a lot of security for my clothes and stuff either…

Why?

Climbing the Kilimanjaro has been on my wish list for a few years. I’ve loved hiking up mountains in National Parks for years, and when I read Douglas Adams’ description of the Kilimanjaro in The Salmon of Doubt, I was sold. He presents the Kilimanjaro as “the tallest mountain in the world”, and sidesteps Everest and all the others because “it all depends your point of view.” Because “If you were going to climb Everest, you would probably start somewhere in the Himalayas, and anywhere in the Himalayas is pretty damn high to start with!”. But about the Kilimajaro he writes: when you finally see … the highest mountain in the world (from base to apex), your blood runs cold. At last you say; “oh, above the clouds”. Your whole head tilts upwards… , ah well, you get the picture. And in his trip he describes wearing a rhino suit climbing up the mountain in order to raise money to fight the trade in rhino horn. That’s not part of my plan so far.

Preparation?

Only 4 weeks in advance we decided to climb Kili; there were still 6 spots left at www.expeditiekilimanjaro.nl. Our ‘international group’ turned out consisting of us and two students from our hometown, hilarious, but it was good fun anyway. There was little time to physically prepare, but each weekend I did a long hike. The highest hikeable point around here is the Kardinger mountain (artificial heap, 15 meter high), so I climbed it many times .

I did bring good gear however. It’s bloody cold up the mountain; worst case around -20 to -30 degrees celsius. Most of the necessary stuff I had, but I bought some additional thermal wear (the best), a ninja-style baclava, and I borrowed a sleeping bag in addition to my -7 degrees-ready sleeping bag

The route

Our plane landed in Nairobbery (please fasten your wallets) (Kenya), flying there is a lot cheaper than anywhere in Tanzania, so from there we took a bus to Arusha (Tanzania). And it’s not the advertised three hours; with the road conditions we experienced you need a plane to cover it in that time. It’s at least five hours bumpy bumpy, and it’s a miracle the busses don’t break down, not even once on the trip.

In Arusha we stayed at Sakina Camp hostel, quite ok, but neither the advertised 20 minutes walk to the center (maybe if you’re one of the famous Kenyan marathon runners). Ah well, in Africa people are not in a hurry. The owner Siporah helped us very well , so we’d go there again. We visited Arusha National Park (with a nice walking safari, almost touching the buffalo’s ) and acquired about the safari possibilities after we’d return from the mountain, but we were really looking forward to starting our climb.

It takes a few days to get accustomed to Africa again, because it’s so different to our ‘normal’ lives. It’s hot, dirty, and you have to haggle a lot to avoid feeling cheated. It feels so unorganized, all existing infrastructure is built by the western colonialists in previous times, and broken down by now. It takes a few days to see it’s unorganized like an ant colony is unorganized, and then it’s quite doable to navigate through society.

We took a bus to Arusha, and our hostel provided the best view of town on the mountain:

After an addional day of preparation, packing all the right gear with some help of our guides, and renting additional walking sticks for the way down, we were ready to go. I plotted our camps along the route on Google Maps:

Camps

Route

Walking

For seven days and six nights I’d be stuck on the mountain. Our group of 4 was supported by 16 men (!); 2 guides, 2 assistent guides, 2 cooks and 10 porters. Apart from all the hiking and primitive toilet facilities, it’s quite a luxurious trip. As soon as our group arrives at any camp, popcorn and tea is served, the tents are pitched up, our luggage has arrived and the cooks are preparing dinner. I don’t think I’ve ever made a trip this well-arranged!

Actually, the hiking is quite easy. It’s really moderate to cope with the height difference. The guides keep telling us to do “polle, polle”, because there’s a fair chance to be bothered by altitude sickness. It’s not the hiking that’s hard; it’s the height (5895m) and the cold (-25 degrees) that stop people from reaching the top. Like said, it gets colder up the mountain, and we take cover during the worst parts of a hail storm. One of our porters is less fortunate and slips, with a huge gass canister in his hand! He has to leave early to get his broken hand plastered…

We were lucky in this case to be assisted by the friendly doctors of a group walking along with us. This group was quite special in it’s own, consisting of people with diabetes and supporting staff. The mission of the Kilimanjaro Challenge 2008 was to show that diabetes doesn’t necessarily limit you in doing things like climbing the Kili. Additionally, the purpose was researching how the diabetic body behaves on great height. Herefore, they (or rather their porters) tagged along a lot of solar-powered research equipment, and the research results will be used to advice future diabetic climbers. The group contained enough doctors to support their people in case of special needs, and to check our porter’s hand…

The top

The most strenuous part of our journey spans the trip from Barafu camp to Uhuru peak. Starting at 23.00 at night at 4600m, we have to climb up in the dark wearing enough clothes to protect us against the freezing cold. I wear two hats, two pairs of gloves, and still I have frozen (blue) fingers. We walk slow, out of breath, and everybody is in some way affected by the height (dizzyness, headache and loss of apetite). After hours and hours of just walking up, when the sun rises behind the clouds, we reach Stella point. This is the point where people bail out because they can’t continue, and I was unsure for a while whether everybody could make it. However, making it this far seems to re-fill our spirits and there’s no way any of us would give up. Only 45 long minutes later, walking alongside a big glacier, we’re at Uhuru Peak. Nothing to see through the clouds, but we’re at our destiny!

In retrospective I wouldn’t recommend trying this exam if you’ve only worked with Java shortly, because it tests a lot of in-depth J2EE knowledge. I think of all Sun Java Certifications, the Web Component Developer Exam is the most relevant for Java Web developers, and I would be really happy if the people working on the Java projects I have to deal with at least have the knowledge tested by this exam. This would save me many refactoring headaches of people ‘re-inventing’ Servlet filters and listeners, and makes sure you’ve at least heard once about ‘Model-View-Controller’…

The Java Ranch is *the* place to go for information about all the Java exams. And I used the following books for study purposes;

Down at the ferry I met this funny dude, with his own website. He and his dog entertained us well while waiting for the ferry.

I’m visiting the JavaOne Conference in San Francisco. It’s the best Java related conference I’ve vistited so far, with many new updates & ideas from the Java world.

I finally got some insight into the big Web Service standards mess. There are so many WS-* and related standards, with overlapping goals, incompatible, with extensive or a complete lack of (reference-)implementation/tool support. This hinders judging the viability of a standard. The reason is that several standards organizations (like W3C and OASIS and vendors (like IBM, Microsoft) push their own standards. One way to deal with this problem is to wait on what the Web Service Interoperability Organization does. This organization (which chose the unfortunate WS-I acronym) tests standards and creates of standards that actually work together. Also, major tool vendors make a specific standard & technology a safer bet.

I noticed that the Swedish language seems somewhere remotely readable and the people can be more outgoing than I expected. As of Växjö, it’s a small and quiet town (it’s already dark at 4pm) and the most interesting picture you can take probably is the

3

4

Just some things that jump into mind. In one of the Panel discussions, I think Alex Wolf took quite a provocative position: “The last fifteen years, we witnessed the golden age for software Architecture. We tend to be really negative about this field, but we have lots of achievements. All main problems have been solved, and from now on we can just roll downhill”. Quite an interesting statement for a research field that’s still in the stage of debating terminology (e.g. how do you define Software Architecture ?). Also, two interesting topics I noticed that got attention were “what is it that software architects do?” and Design decisions.

I didn’t write a lot about research as I wasn’t really happy with how things progressed, but this seems to cheer up these days.

A spring roundup links to a bunch of current ideas on the heavy-weight -not-so-loose-as-intended service implementations standard, and specifically SOAP gets beated. REST seems to be the prefered approach.

Luckily, as an expert states Service Oriented Architectures aren’t dead:
provided Architecture is based on patterns and intent, not technology selection.