csant/log

As of today BBC News serves ads "when viewed from outside the UK." You don't like ads on a serious news site? But you are a regular BBC News reader? A fix is simple: create a UserCSS file for http://news.bbc.co.uk/ with the following content:

#ad1, #ad2, #ad3, #ad4, #ad5, #ad6, #ad7, #ad8, #ad9 {
  display: none !important;
}

And go back to reading your news undisturbed.

Ever since I moved from the standard SUSE 9.3 kernel 2.6.11 to a custom built 2.6.14, upgrading was a joy: the first step on an upgrade to a minor version bump takes a while: it is a minor version increase, yet support for tons of new drivers and configuration options are added, you walk your way through the configuration, make and then make modules_install && make install - and ready you are.

When we talk versioning number, it is interesting to note that with the A.B.C[.D] scheme, or <version>.<major>.<minor>[.<patch>] scheme, a D number first occurred when a grave error, which required immediate fixing, was encountered in 2.6.8's NFS code - today D numbers are released every other day…

It was a joy, I said - until the upgrade came from 2.6.19 to 2.6.20. It took me a few days to fix some issues with proper loading of some modules, certainly due to an older udev I am using. I solved it by finally deciding to compile those modules into the kernel (fair enough, why didn't I do that earlier?).

And then came 2.6.21 - somebody even saw some connection to the disaster of the Chernobyl Nuclear Power Plant… It was followed by a patch release a few days later with two bugfixes for networking issues that should cause all users of ipv4 or ipv6 to upgrade to this version. Right. IPV4: Fix OOPS'er added to netlink fib. 2.6.21 died every half an hour on my machine.

Following this much less pleasant release, there was a long and interesting debate on the LKML about releasing with known regressions.

0 regressions is never realistic (especially since many regressions
might not be reported during -rc), but IMHO we could do much better than
what happened in 2.6.20 and 2.6.21.

No kidding. The quality is really going down the drain - and this, notabene, is said by somebody who is certainly not stress-testing or cornering the kernel into odd cases. I am kernel's Joe Average who is just using his machine to accomplish some other tasks. I am fine with putting up with a few inconveniences, I know that can happen when you run "the latest" - but please, this is not the bleeding edge: I am not fond of running the bleeding edge development kernel! This is the latest stable version of the Linux kernel… Or rather: wannabe stable.

The new security measures for hand baggage on EU flights are utter nonsense.

The main rule is that no liquids may be brought through the security check, except for containers smaller than 100ml fitting into a 1ml plastic bag. The amount is determined by the size of the container, not the amount of liquid in it - pray tell: what kind of sense does that make? Note that you can bring an empty bottle on board, so all you'd need for your half-empty bottle is a smaller container to temporarily fill the liquid into during security check: afterwards you can fill it back. "Obviously" you cannot bring sealed bottles on board (say e.g. wine bottles with a numbered governmental seal), since they may have been tampered with (i.e. we don't trust governmental seals anymore), but "obviously" you can buy bottles to bring on board once inside the security check: it all smells a bit too much of boosting the security area shops' business, doesn't it?

Well, to tell you the truth, I am mighty annoyed. Have we ever seriously considered how pathetic these security measures are? Has anybody ever heard of solid explosives? (or have we decided that the liquid ones are the only dangerous ones?) What would the difference between five 100ml containers vs one 500ml container be? What does make the five small containers more "secure"?

I have seen a lady forced to either throw her 250ml water bottle away, or drink it there, before the check! And this in Berlin Schönefeld, where you are given "boarding passes" without your name on it! I did check in, went for a stroll, and half an hour later went to board the plane: anybody, absolutely freakin' anybody could have boarded that plane instead of me, since there was no check whatsoever at the gate: the "boarding pass" got marked as "seen" manually, and I did not even have to show a passport!

P.S.: Yes, they did confiscate a 400ml jar of jam I had with me as a present for a friend. The 100ml jam was "secure" enough to be allowed on board.

Today there are only two popular methods of user interaction: command-line and GUI. Command-line interfaces are often favorable to the experienced user, if only because graphical interfaces are so poorly designed. They are also severely limited in their input and output capabilities. Graphical interfaces are all some slight variation of a system developed at Xerox PARC long ago, which was designed to emulate troublesome, tangible objects such as paper and desks. As anyone knows, this is a tiresome paradigm; why should computers emulate it?

WIMP interfaces have been dominant since the late eighties - and it is shocking to see how little we have moved away from that model. The only progress we seem to be able to have made is in adding 3D, drop shadows, transparency, a shiny look and, most recently, the ability to slant your windows. As long as we are not able to project 3D objects in the air, 3D for GUI is just a silly gimmick. The user interaction is on the screen, and the screen is flat. 2D.

Instead of eye-candy, some evolution in the way we interact is badly needed, some substantial re-thinking of the UI experience instead of polishing up the surface. WIMP interfaces seem here to stay in the near future - that was 1997, which is now almost 10 years ago! Time for them to be replaced by something more adeguate. Why would we, in an age in which many people do not even know anymore how a desktop (a desk top, not a PC) with tangible objects looks like, care about emulating it? Wouldn't it make much more sense to think about a UI paradigm based on the possibilities the medium (the screen, in our case) has to offer? Designing a GUI emulating your desk top is somehow equivalent to designing a mobile phone based on the paradigm of your fountain pen - nobody would even dream about something like that, and yet we are still stuck in windows and icons mimicking your desk, menus and pointing devices.

The command line interface has not undergone any substantial development either, and ended up being somehow the antagonist of the graphical user interface. For several tasks I find the CLI more appropriate, but I don't really understand why it needs to be an either-or scenario - I would wish for a less radical distinction between the text based nature of the CLI and the GUI, allowing for text based "queries", or dialogs with the application from within the graphical environment, allowing me to interact from the command line with the graphical environment, and allowing for graphical elements on the command line when they'd help the user understanding or finding some orientation. I wish for a better user interface.