Documentation is important for a lot of people. For example for games where you don’t immediately understand what the goal of the game is. Or you just want more details about a particular feature.

KDE has had a long time policy that no applications can enter the main modules or extragear without appropriate documentation. Unfortunately for some applications that’s also the last time the documentation has been reviewed. That means documentation is often outdated.

On the other hand our great time of translators silently continues to translate the documentation, which is kind of sad if that involves dated docs. So I hereby call everyone to review the documentation of your application and try to make it interesting for the users to read.

Personally I think it isn’t useful to describe the interface step-by-step, but focus more on hints and tips to guide the user in the right direction. Like, if you want to reach A, try changing the settings B and C and fiddle with D.

Actually, I would like the documentation to become more 2.0. Userbase is a great step in that direction. What is lacking is the integration back into the application. The php.net is famous for the documentation, examples and discussions that can be found in the comments. I hope Userbase can be integrated better within our applications.

Anyhow, I wanted to write about docs.kde.org. The site was not updating the last couple of months due to several reasons. We have now put the generation of the site onto a different box, We’ve fixed a couple issues with the generation and made several improvements. Like PDF’s for all languages. Enjoy.

Oh, and if you feel like improving the documentation, just write a mail to kde-doc-english@kde.org and they will help you get started in no-time.

“In the 17th century, Saba was a hideout for Jamaican pirates!”

That left us no other choice then to call our new shiny server Saba. It’s a Intel ® Core ™ i7-920 Quad-Core with 8GB memory. Last week Jamaica went down a couple of times. Load spiking, the only way to resolve it was to reboot. And we did that more and more. Jamaica was a server which was only used for serving websites, mostly the Capacity powered websites from svn and git. The server was donated by Trolltech a couple of years ago and was currently hosted by the Nokia sysadmins.

A few weeks ago I approached Nokia to see if they were interested in replacing that server and renewing their commitment to support KDE with a new server. You can not imagen how happy I was to receive a confirmation from them last week, after the third consecutive day in a row with problems for Jamaica. We ordered the server right away and Hetzner.de made the server available the next day already. We used the weekend to setup LXC, deal with the IP- and IPv6-addresses, setup Apache and Varnish and transfer the sites one by one. It seems the transfer was a success, as we did not receive any complaints. I want to thank Nokia for donating this new server. It’s very much appreciated.

It also gives us the opportunity move some other sites to it, like Planet KDE. That’s part of our current project to relief KTown from its duties. It’s the oldest server in town and time to shut it down. But before we reach that day, we have a lot of work to do.

One piece is already completed and that’s the monitoring software for our servers. We used Nagios up to now and have replaced it with Zabbix. We even made a nice public page for everyone to see the state of the current KDE servers at http://serverstats.kde.org. It will report the detected problems automatically on that page, scream in our #kde-sysadmin irc channel and sent us some mails. The page is now a little boring, but if you wonder if something is wrong, you can always go there to see if there are some problems…

And that is very sad. They are making some classic errors in their release process and even worse, they are not learing about mistakes from the past. They have released a new version which is not compatible. Which means you con not migrate at all. There is a plugin that converts your content, but it will not convert any modules, templates or components. Which basically means all the hard work is still to be done after you have used this helper tool. Some notes about all this:

Release numbering
When they went fom 1.0 to 1.5 they made the error to call it 1.5. It should have been 2.0. The release was completely incompatible with the previous releases. And now they have released 1.6 which is again incompatible, so the major should have been bumped again. 2.0 would have been more appropiate.

The reasoning is also that I need to approach my customers, asking for budget for such a transition. This is easier to explain to my cusomers if the customer actually notices a bigger jump in numbers. Convincing the customer is hard anyway, because they don’t see the direct benefit. Directly after migrating there should be no change of the website for the user of the website, whilst the customer has spent money on it for no benefit. Such a migration budget request has to be combined with some new feature for the end user of the website to be able to get through.

Incompatile
The seond mistake is that the release is incompatible with 1.5. This is not what is promised to us when the transition happened from 1.0 to 1.5. 1.6 should be an easy update, it would not happen again. We are now stuck with customers which we know will not have budget to move to the new version. Ending up with new security issues in a couple years.

It’s not only highly disrespectful of the current users, it’s insulting. You leave them behind. Every marketing pro will explain to you in a just a few lines how important existing customers are, how you should treat them with love and care. Joomla does the opposite.

Market share
They clame that from all websites out there 2.7% is Joomla based. Well, you can scratch that. It’s 0%. You will need to regain all of that percentage. The reason is simple. Whenever i approach the customer for budget, i will have to convince him. And that will give the customer the option to make another choise.

If a simple upgrade possibility would have been in place, i can migrate the customer ‘silently’ and it would remain a customer for sure. Just look how awesome WordPress does upgrades, that is the way to not loose customers.

Reputation
I can surely understand that there can be a big disruption sometimes. Sometimes you can not avoid big changes to code. But now Joomla has passed the acceptable line. Making it impossible to migrate smoothly between 1.5 and 1.6, just like the migration towards 1.5, hurts Joomla’s reputation. How on earth do you expect me to believe you when you are now claiming that this won’t happen for 1.7? Which means selling services around Joomla just got a lot harder.

Real CMS?
The core idea of a CMS is to have nice separation between content, layout and code. Making it possible to build a website which you only have to write the content once and make it accessible in diffent layouts. I tell my custmers that whenever they are fed up with their layout, or when the company changes their style, they can order a new layout with us and the content will remain.

I will now stop telling customers that, because it is now clear that the chances are big that whenever such a layout change is needed, their site is dated and incompatible with the current version. Which means we can not deliver a new template for which can safely say that he can use it for the next 3 years. It’s become clear that you can not count on any modules and components being availale for that version at that point in time.

KDE
Before anyone claims that we did the same between KDE 3 and 4, please consider that that was needed because Qt’s code changed massively, it was a one time thing and we learned some lessons from that too.

Advise
Who am I to give advise, but I would like to point out to the Joomla team that they are on the wrong track and they should quickly fix this and provide a smooth upgrade path between 1.5 and 1.7 or 1.8. If you care at all about your users.

Due to a failed disk, one of our anonsvn/anongit servers is down. while other servers have taken over most services, the repository browser on projects.kde.org and quickgit.kde.org depend on that local mirror. That is why they are currently unavailale….

Yes, I bought one. Well, a used one, it was cheap now version 2 is out, the version 1 is dumped for crazy prices. And v1 is what I wanted. First, I seem to have to explain why I bought it. I prefer Open Source devices of course, but those don’t exists. Meego tablets are rare, and all currently available devices have their problems. I hope next year I have a choice in this area. There are dozens of rather cheap Android tables. Everyone who knows me, knows I won’t buy a Google device as first choice if there are alternatives. This means, the only reasonable tablet for me was the iPad.

I’ve been playing with the iPad for a few days now, and I’m impressed with it. It’s a fine device. The software is very complete and integrated in each other. The App store contains very nice applications. For example Flipboard. It is an app that aggregates news from different sources and presents you the news in a magazine like style. Each page with a new layout, based on the length of the articles. For example, Planet KDE looks like this:

Pretty neat. If you click on the item, you get the full story. There are also beautiful apps for Google Reader, Facebook, Formula1 Live Timing. There are also apps that are made for the iPhone that run in a tiny window on the iPad, those include things like an app for the Dutch Railways, Spotify and more. What they all have in common is the level of completion. Most apps are very complete and well integrated. For example some apps open a browser to view an internet page. That comes with a ‘Done’ button. When pressed takes you right back to where you were before. Although there seems to be a lack of real multitasking on the device, it opens apps exactly where you left. We know that from browsers which reopen all the tabs from the last time, but on the iPad that’s everywhere. For example the news reader opens exactly where you left last time, so you can continue reading new news immediately.

There are definitely lessons which can be learned from the user interface. I hope KDE people can grab an iPad and look at how things are done there. Compare the App Store, which combines games, e-books, apps, music and stuff to KPackageKit and you know we have a lonngggg way to go. Having such an example is good, it can lift us to the next level.

This device fills the gap between my N900 and my desktop computer. Sometimes I don’t feel like booting my computer, but I do want to be online for my Mail (yes, I admit I’m adicted), useless surfing and what not, without actually having a full desktop which makes me actually do stuff. As a bonus I configured my Myth server to be a Media Server via uPnP and my desktop to be publish my Photo’s. Now I can watch my photo’s and recorded shows from the comfort of my bed. Something I did not buy the device for, but which is a very nice extra feature of a tablet.

But be realistic, not everything is perfect. You can see that it is a commercial platform. There are lots of apps that cost money. Often that’s not a lot, only a few dollar, but if you don’t want that, you are stuck with lower quality, reduced functionality or nagging ads. Also the level of settings is very, very low. Ever seen a mail app where you can count the amount of settings on two hands? I know the amount of settings runs in parallel with the amount of features.

If we on one hand can avoid these disadvantages (which we easily can) and on the other hand learn from the good things from an iPad, I’ld be very happy. Yay for Peter Penz, Martin Gräßlin and Aurelien Gateau‘s blogs.