The Story of Maths. Episode 1. The Language of the Universe
First episode at the BBC iPlayer : The Story of Maths. Episode 1. The Language of the Universe
Laurie Santos: A monkey economy as irrational as ours
Google Kills Wave
Now, Google has just announced that they are killing Wave!. Quoting them: (...) But despite these wins, and numerous loyal fans, Wave has not seen the user adoption we would have liked. We don’t plan to continue developing Wave as a standalone product.
I think I can now live in peace knowing that I am not a total idiot, that there is order in the Universe, and that even a company as powerful as Google cannot bend basis human nature. In other words, keep it simple stupid...
Amen.
Helpdesk call on Pacman day
Falcon is dead, introducing Nyx and Alcyone
While thinking of what I had learnt doing Falcon, I suddenly had an idea (actually the idea came while I was sleeping last night and I have been frenetically coding since I woke up this morning). To make a long story short, turns out that few days ago I did some changes on my website pascal.alseyn.net and removed the computing and mathematics sections. A lot of files, courses and documentation that I had accumulated over the years suddenly disappeared. This was a bit of a pity, but the main reason for me to do that was that I had an overhead keeping them in sync with the original versions scattered all over my hard drive (original versions that I am updating or rewriting in this moment). Eventually, one thing leading to another, Nyx and Alcyone were born.
Alcyone, located here http://alcyone.alseyn.net [dead link] is a simple looking repository of documents. It contains some stuff that for some reason I want to keep online. This is how it all works: when I want a document to appear on Alcyone (in other words to be public), I have nothing else to do than putting a little .nyx file next to it wherever it is located on my hard drive. This .nyx file is a small json document containing some metadata. An example is given below.

Soon after, Nyx which has jurisdiction over Atlas, the root of my file system, will discover the .nyx file, register the target file, and will send it to the remote website database. From that moment the file will be available online. The nice thing is that if anything happens to the original target file on my laptop (for instance being updated), then the new version will be automatically uploaded replacing the old one. Of course, my computer is not always online, so the synchronisation with the remote database is done at first opportunity. Then, later on, if I want to remove this file from being visible online, all I have to do is to delete the .nyx file :-)
The online public version of Alcyone, is pretty basic, just a repository of documents, but the local version, the one I have on my local web server on the laptop, has got all the nice Falcon-ish user interface and search capabilities I had already before :-)
I am so happy ! It's like having the best of both worlds (^.^), actually the best of three worlds: my files are where they should be, I have some nice falcon-ish web browsing capabilities which highlights some structures that a tree-like file system would not be able to express, while having an automatic-unobtrusive-no-headache-it-just-works publishing mechanism :-)
Last, Nyx and Alcyone are named after, respectively, the Greek primordial goddess of the night and one of the daughters of Atlas and Pleiones.
The sausage conjecture
In three dimensions, if you have 56 balls of fewer, then the solution is to arrange the balls perfectly aligned; their collective envelop then looks like a sausage. On the other hand, if you have more than 56 balls, then the solution is to put them tightly packed in a bag (like potatoes).
And what about in 4 dimensions ? Not many people can visualize in four dimensions but the question still makes sense, because we know what it means for four-dimensional objects to touch each other without penetrating each other, and we can compute envelops and their volumes (*). The fourth dimensional space behave likes the third dimensional space: there is a number below which the (four dimensional) balls should be aligned (in the four dimensional space) and above which they should be packed in a (four dimensional) bag. We know that this number is between fifty thousand and one hundred thousand but we do not know what is it exactly.
For dimensions more than 4, i.e., dimension 5, 6, 7 etc, the conjecture is that the sausage is always the optimal way to place the balls, regardless of the number of balls. We have proven that this is true for any space of dimension more than or equal to 42, but we still don't know if the sausage arrangement is always the optimal arrangement for spaces of dimension 5 to 41.
(*) We are still talking about "surfaces" here, but in a four dimensional space a surface is a three dimensional space, so that's why I said "volume" (because to me that's what they are), but beings living and perceiving in a real four dimensional space would probably call them "surfaces".
Benoit Mandelbrot: Fractals and the art of roughness
From Scratch
"Natural" integers
... only to realise that actually I started to count at zero.
Special purpose nodes in Falcon
By doing this I would have what I call "special purpose nodes". Such special purpose nodes would not be structurally (from the point de vue of the graph or the code or the database) any different than any other node, but they would have been (most of the time automatically) created, for the purpose of adding structure to the graph. If I do that I would, in fact, put on the graph (or more exactly within it) an additional semantic to mirror my existing [Lucille:Genetic Timeline].
At this point, why not push the envelop and, after all, use other special purpose nodes, related to each other in a tree like structure, to mirror in Falcon all the other organisation schemes I already use to manage my data on the hard drive. Such additional structure would not be harmful, and in fact it may even be useful if one day I really cannot semantically find a node but only remember the day or month I entered it, or any other detail which can be interpreted within one of the supported organization schemes.
This said, it's even deeper than that. This simple way to add additional structures to the graph might solve a problem I sometimes have: when I exposed my file management system in The Fourth One, I presented it as if it was all so perfect. This said, I occasionally I do hesitate between storying a file in the PerSubject tree, also known as [Lucile:Atlas], or the Generic Timeline. Most of the time I end up finding a good reason why it should go to one place rather than the other, but this extra step is only the result that something is fundamentally wrong. If I have already in Falcon the Generic Timeline (alignment of date nodes) as well as a reproduction of the [Lucille:Atlas] then I can just add the new data node to the graph where it will be automatically linked to the good date node (and month node) and then manually link it to the good node of the Atlas tree (if this was relevant). This way I will have tree ways to find the information: by date, by subject or by related concept, all this in case I additionally really do not remember any of the words of the text contents of the node (in this latter case the text search engine is still the fastest way to go).
The nice thing with the idea of adding structure using special purpose nodes is that you can add, and have peacefully coexisting, as many structures as you want, including some really exotics ones :-)
