Posts Tagged ‘cheating’

Versioning generally and in zkanji

April 19, 2012 3 comments

Nowadays every software out there has an identification number called version number. Talking with average people who are not programmers (and sometimes even programmers) I found that they often do not know what the version number indicates, and how it is designated to programs and other data. There is the myth out there that the version number actually means something. But where did they get that idea? More often than not the version number is just a random number (though usually larger than the previous one), and its only role is to differentiate between some other, slightly different… version.

Another myth closely related is that big companies have a well defined system which lets them create meaningful numbers for the versions of things they produce, so the number can tell the initiated something. Let’s take one of the world’s largest software companies for example. μ(that’s a micro)$oft is known for releasing some operating system which I can’t recall the name of at the moment. I don’t want to mention the numbering of its OSes but rather components called DLLs, which are basically parts of other programs but in separate files. These can be parts of multiple programs, but as they can also change with newer releases, programs could stop working with newer or older ones. These parts also have version numbering. For example program A uses DLL v1.0 which paints the screen blue and program B uses DLL v2.0 which paints the screen red. Both programs need the screen to be a particular color or they crash. To avoid that, the programs simply check the version of the DLL and use the correct one. (This is highly simplified because the OS nowadays uses ugly tricks to give programs the right DLL (or the wrong one).) This sounds all right, but what if the two different DLL versions had the same version number? This is not just a theoretical question, it happened many times.

One other fine example is the numbering of browsers. The so called “browser war” is raging and so companies came up with clever schemes to get some advantage over others. For example when Goggles (with a small typo) changed a text on a button, they released v19.9999 instead of their previous v18.9999. I don’t know if this worked or not, but their cowpetition moo-zilla copied this trick and now we are at v1X (which is still less than G’s so it must be inferior (or not)). I don’t want to get into which one is better, and there are also other browsers, this was simply an example of how much version numbers can be trusted.

(I feel I must mention that these companies usually have guidelines on version numbering, but I’m not alone with the opinion that those are only used to trick the gullible users into believing that a bigger number means better software.)

zkanji also has version numbers, because every other program has one, and it makes it easier for everyone to know if there is some difference between the releases. Just don’t believe it means anything. Besides it has happened a few times in the past that I found a bug the day after a release and re-released a fixed program with the same version number. (But it was long ago.)

The only rule I have for versioning is that I try not to reach v1.0 until zkanji has everything I have in mind for it. If I were μ$, G or M, it would probably be at v9.2 already.

Categories: Rant Tags: ,