2004 Archive

As more and more open source software runs well on windows (something which has not always been that way to my great frustration), becomes more cross platform (e.g. xchat, sodipodi, gaim, openoffice ), i found myself needing an apt-get equivalent on windows.

Usually windows versions of OSS are either file archives to be dumped into their install directory, or wrapped in a (big) windows installer. I searched for a windows version for apt-get, but got not much to work with / from.

I wonder why, can’t be that difficult. Shame i have no credit left on my time account.

Linus on kernel management style

“While it turns out that most people are idiots, the corollary to that is sadly that you are one too, and that while we can all bask in the secure knowledge that we’re better than the average person (let’s face it, nobody ever believes that they’re average or below-average), we should also admit that we’re not the sharpest knife around, and there will be other people that are less of an idiot that you are.”

Amen

From the back yard. Pity it is too cold to sit there with a beer.

Those marketing guys must be rather clueless, lots of software vendors pick a couple of letters and prefix all their software titles with it. I can only gues this is due to lack of inspiration or to create at least the impression that the software is like a suite of packages which work together (which it does not in most cases) It’s not only the professional marketing departments who lack inspiration, Open Source developers, whether professional or not, have a liking for prefabricated names too. Here’s the slate what i could come up with without much research, lots of letters left, so i you have some plans grab one, while they are still in stock:

1. The a is liberally replaced with a @ in both names of classes (eM@il) and names of software packages;
2. The e seems to have grown into a prefix to be used for a class of software (eMail, eLearning, eCommerce) roughly describing doing the classical thing, but then with the internet (whatever that means in practice);
3. gnu – Applications which like to identify themselves (i guess) with the FSF or the GNU software set or the GPL (gnuTella,;
4. g – GNOME stuff;
5. i – Apple (iTunes, iChat);
6. k – KDE applications (kNode, kWrite);
7. moz – Based on mozilla framework (mozBlog, mozEdit) [Note: also the postfix zilla is often used];
8. x – X windows applications (xChat, xEyes);
9. xar – Xaraya modules [ hesitated a bit to put this one in ;-) ] And of course there is microsoft which just puts MS in front of everything. I’m sure there are more examples, the above is by head. So, if you want a letter prefix, get one fast; we only have 26 letters.

I decided to disable comments alltogether on this blog. The amount of time required to manage the ip banning and deleting the spam comments is getting too much in comparison with the small scale blogging i do over here.

I just dont have the time to go to war on this and keep securing the blog, while giving others the opportunity to comment on the postings. My apologies to the serious commenters. Please use the trackback facility to be able to refer to the postings which you want to comment on.

Microsoft granted patent on button click timing
Microsoft Corp. has been granted a patent relating to the use of buttons on hardware devices that form part of a user interface. U.S. patent 6,727,830, granted April 27, is described as a “time-based hardware button for application launch.” The patent abstract goes on to explain that the patent relates to how different functions can be invoked depending on whether a button is clicked once, clicked and held down for a period of time, or double-clicked within a short period of time. [MacCentral]

Mind mapping seemed an attractive concept to bring some order into my chaotic thoughts, so i spend an afternoon looking for a program running on OSX to give it a spin.

The first thing I noticed all these programs assumed (with the exception of PersonalBrain which is not available on OS X) that my thoughts are organized in a tree like manner. At some point someone must have decided that a tree like organization for creative thinking is appropriate.

I am probably not a creative person

The leading mind mapping software on OS X does not even support quick linking between one part of a tree and another part. Yes, you can draw a line between two leaves, but that’s not exactly what i had in mind.

The PersonalBrain mentioned above comes quite close to what i was looking for, taking each idea as a possible starting point for organizing. So, i played with it for an hour, deleting it from my PC right after that. I managed to make a mess out of it immediately. Instead of stimulating navigation and organizing thoughts I found the program chaotic itself. It created links for me which i didnt want with no obvious feedback when I did that. (and i got dizzy from the UI too)

There must be something wrong with my brain.

If i understand all the involved techniques right, i need a very graphically oriented and very fast RDF editor with lots of attention to keyboard shortcuts and screen organization, which hides the horrendous details of RDF graphs and how they are specified. There are some RDF editors out there, mostly java based, but they are exactly that: RDF editors, nothing more, nothing less. Not very imaginative about making RDF being a storage format for semantical relationships and abstracting the interface around it.

I must have misunderstood the semantics.

Distributed and peer 2 peer systems  seem to have found it’s way into several applications:

• instant messaging (icq, jabber, msn etc.)
• source control (bitkeeper)
• file and content sharing (kazaa and the likes)
• personal content (weblogs)
• connectivity (wifi grids)

It is my belief that this trend is more general than in computing only. By its character the principles of distributed production and p2p delivery are just easier to implement in information technology than in other areas. The area in which the principle will have the most impact in the next decades is distributed energy management.

We will, most likely based on a combination of solar/wind energy and fuel cells, produce our energy needs in a distributed manner, delivering surplus energy to micro-grids in our neighborhood which in turn delivers to the macro-grids in our town.

The same advantages which hold in computing are also valid in energy production:

• localisation: create energy based on where it is needed, not at the place we have the largest factory;
• reliability: the effect of a failure is only affecting a small area, not a whole city, like last summer in the US
• choice: if i have a surplus of energy i can deliver it peer-to-peer to someone of my choice;
• customization: energy needs for everyone are different, smaller energy plants controlled by the users, not the producers;

Now that fuel-cells will start to appear in notebooks and cellphones the process has started; we’re still at the mainframe stage though, but the general outline of the things to come isn’t hard to imagine.  Each change in type of energy and type of communication has caused radical changes in the past, i fully expect this one to do the same

By its character the principles of distributed production and p2p delivery are just easier to implement in information technology than in other areas…. We will, most likely based on a combination of solar/wind energy and fuel cells, produce our energy needs in a distributed manner, delivering surplus energy to micro-grids in our neighborhood which in turn delivers to the macro-grids in our town. customization: energy needs for everyone are different, smaller energy plants controlled by the users, not the producers; Now that fuel-cells will start to appear in notebooks and cellphones the process has started; we’re still at the mainframe stage though, but the general outline of the things to come isn’t hard to imagine.

Interesting discussion last night on Xaraya’s IRC channel. I installed An amazingly good looking IRC client for OSX. This client has a feature called a buddy list which shows you the online status of people on IRC servers they are on regularly, much like the IM clients show an online status for people.

The feature saves me visiting tens of irc servers just to find out that the person i need is not there.

To be able to maintain this online status, colloquy sends /whois commands to the registered servers and channels to get the online status of the people in the buddy list. Obviously the sending of some command is necessary for this information.

When i started adding people to the buddy list, one of the serverops asked me why i was sending him /whois commands every 30 seconds. At that time, i didnt know what was happening. A little investigating showed that it was the buddy list feature.

In short, the discussion went fairly quickly into “user experience” versus “irc is not designed for this and stop doing this please”. This is interesting for me, because usually i am on the side telling people the the system is not designed to do this while at this time i was in the end users chair

There’s no escape from this argument, the feature saves a lot of time, but is using a technique not intended for that particular goal. Who should give in? I dont know.