A 3D Animation of Kernel Source Development

fixit! writes "Have a look at this cool 3D animation of the Linux kernel source. This is how a source browser should look like!"

Comments

Errors

[BenjaminHall]

It would be nice to see major errors identified and then fixed in the movies as well.

Re:Rather Interesting Concept

[njdj]

I think that since Linux is very clean, streamlined code

Perhaps you didn't actually read the page referred to in the story?:
---start quote---
The following code demonstrates exciting
features of GNU C used in Linux:

int a, b;
typedef int t, u;
void f1() { a * b; }
void f2() { t * u; }
void f3() { t * b; }
void f4() { int t; t * b; }
void f5(t u, unsigned t) {
switch ( t ) {
case 0: if ( u )
default: return;
}
}
---end quote---

This kind of code is CRAP. I don't know who wrote it, I don't care if he/she is a genius kernel guru. Hard to read, hard to maintain.

Uses?

[shivianzealot]

A very specific niche comment/query...

I'm not a programmer (BASIC doesn't count, right? :p ), but I have observed the development of a few open source projects and have seen the effects of code being introduced by programmers who have valuable contributions, but interact poorly with the rest of the source (usually novices). So, veterens, could this type of map, applied to the project in question, drive home the point and help mold the newbie into better practices, or are we better off oohing and ahhhing now and moving on to the next article?

visualizing complex data

[fiiz]

Yes, that is interesting indeed.

I think in general there may be interesting research to be done in the area of mapping/visualization of complex data: for instance this project of mapping the internet.

Does this really help in general? Are there many cases where such visual maps would help understanding of complex data?
Think for example, it may be interesting to produce such a map of everything2, which is a sort of hyperlinked online encyclopedia, to see where the clustering is.

In astrophysics, 3D maps of the universe have been produced for some time, and the human-eye understanding of large-scale structure was at first more direct than statistical analysis--for instance, people would see the famous filaments, but stats wouldn't.

A post above quoted the possible use in spotting "usefulness" of code contributions, by looking at their interdependencies for example.

watching the bits on an Atari ST

[Anonymous Coward]

My first computer was an Atari ST. The MWC (Mark Williams C) compiler that I bought with it came with an amazing little C demo program. It must have been 15 lines of C code tops. The Atari ST has an 680x0 cpu chip, and a linear memory model. A chunk of that memory was set aside for video, and a separate chip pumped that video memory out to the monitor. With the C demo program, you could change the base video memory pointer to point anywhere in memory, including low memory, where the operating system (TOS/GEM) resided. By doing this, you could actually WATCH the operating system in action, because each pixel on the video monitor represented one bit! You could see counters counting up, flag bits flip-flopping on and off, chunks of bits being read in from the floppy disk, etc. It is, by far, the coolest thing I've ever seen done with a computer. =) Wish I could figure out how to do it on my linux box.

Use GCC?

[hughk]

GCC is a compiler. It shoulldn't be a probllem to construct a local cross reference from the symbol information that it produces, especially if debugging is enabled. The advantage is that GCC would be used in the same way that it is to compile the kernel.

I can't remember if GCC assigns attributes to symbols so it is possible to keep track of code references but to forget the data references, but that would mean chasing through the debug symbol format.