Continue reading →Entries Tagged 'design' ↓
SFTP from Nautilus
March 31st, 2010 — software, tinkering
Adding Runlevel 3 to Karmic Koala
March 31st, 2010 — software, tinkering
I’ve got a brand spanking new install on an Ubuntu machine that I would like to use as a home media server. Because it is ancient, I’d like to run it without windows. Theoretically, because Ubuntu is a Linux distribution and Linux is a Unix clone and the core of Unix was laid out in SysV, I ought to be able to use runlevel 3.
| # | network | multiuser | GUI | servers |
|---|---|---|---|---|
| 1 | ✘ | ✘ | ✘ | ✘ |
| 2 | ✔ | ✔ | ✘ | ✘ |
| 3 | ✔ | ✔ | ✘ | ✔ |
| 4 | ✔ | ✔ | ✔ | ✘ |
| 5 | ✔ | ✔ | ✔ | ✔ |
| 6 | reboot | |||
Google Feedback
March 11th, 2010 — bug
The direct click path from ?couchdb tutorial is:
Which is the page:
This is half a screen of non-scrolling adverizing covering a page from IBM. This is indisputably a lower quality result than sending me to the IBM page directly.
Valid JSON
February 21st, 2010 — bug
FYI, for the geeks out there, even though:
{ foo: 'bar' }
Is a valid way to initialize a javascript object, it is not valid JSON. In JSON, the key and string values have to be double quoted:
{ "foo": "bar" }
This matters because jQuery 1.4 uses native parsing and the getJSON method will silently fail on malformed files.
Organizing Information
January 25th, 2010 — design
Every time I write this down, I dislike how it sounds, so I’m going to brainstorm for a while.
Templates build composite documents from constituent documents. The example that I mentioned previously was:
<html> <head><title>Greeting</title></head> <body><p>Hello <?php print $name ?></p></body> </html>
Which can be represented as a tree by:
What I would like to explore is a templating system that is drawing from a content tree which organizes binary strings.
Versioning Basics
January 21st, 2010 — software
Versioining systems are pieces of software that track changes in files. They exist in a variey of forms, and one of the more popular currently is git which was designed for maintaining the Linux kernel.
The basic building blocks of a git repository are “blobs.” In git, a blob is simply a file:
Continue reading →Templating Basics
January 21st, 2010 — software
We'll start with the natural building block of any digital system: the bit. A bit is just a binary digit, a one or a zero. Bits are used to encode information. For example, imagine I have a bunch of red cards and a bunch of green cards. I give a red card to everyone in the country except for the President to whom I give a green one. Since the cards only have two possible states, they are a form of bit. The information associated with that bit, however, is highly complex.
Bits are combined into series to represent larger sets of possibilities. Whereas the President bit could only represent two possibilities: someone is the President (red card) or not the President (green card), a string of n bits can represent 2n possibilities. Almost all digital information in the world today is stored as groups of eight bits called "bytes."
Continue reading →Stupid Find
November 9th, 2009 — bug
This command should work, imho:
$file = ~/personal/private/.htaccess;
find ~/ -name .htaccess -type f -exec diff $file {}
\|\| ( echo 'Replacing: ' {} \&\& rm {} \&\& ln -s $file {} )\;
It should hunt down all the copies of my .htaccess files that are only there to password protect a directory and make them symlinks to an authoritative version.
There are many ways to accomplish the task and ⟰ is apparently not one of them. ☺
A Project Root
November 8th, 2009 — software
In the last post I mentioned the idea of adding a user root special directory at '...'. My other new directory is the repository (or project) root.
For example, I designed the site mr.pcvs.org for the volunteers in Mauritania. Part of the site is a cookbook, both of those sites though are a part of my web repository.
Continue reading →System Usage
November 8th, 2009 — software
I want a web service where I can easily put a url to the output of a system command. I’d like to be able to stick a tag in this post and have the output of:
df | awk '/^\//{ used += $3; total += $2 } \
END { print (used / 1024 / 1024) "g" \
" / " (total / 1024 / 1024) "g" \
" = " (used / total) }'
Piped into it whenever the page is loaded (with some caching to handle). It would be a network-level /proc filesystem. The systems could key the values on URIs. Putting it at the network level also lets you run a genetic algorithm on the programs that and being used to configure routing.
Adding artificial intelligence to routing does sound like a pretty good idea, now that I mention it. ☺ I could let remote programs run commands that I don’t care about being public like df, free. I could even expose a version of ps.
The network really would be the computer in that situation as individual routing agents learned from each other the optimum distribution patterns.