I have a confession to make: I'm not using emacs+org mode to keep my to-do list and appointments. What? The same emacs junkie that used emacs for (almost) everything last December gave up emacs? Yes, the same, but only gave up for this use case. I have some good reasons about why I am using TaskWarrior. Which does not mean that I could not be using emacs for the same, I just wanted to try something new.

You could get over the top and write your own task manager as a lot of people have done. In fact, I found about TW after learning about t, the python based minimalistic task manager, through finding t- in HackerNews, an even more minimalistic version of it.

Adding a task to taskwarrior is a little more simple than using org-mode, but this is just a matter of taste. You can add a task to an org file just by writing a line, whereas to add a task to TaskWarrior you issue task add task. Which I feel better about, it is the way it should be.

At first, as part of my emacs 30 day challenge, I planned to use the Nanonote as an agenda, with emacs and org-mode. But this was hard to do: the keyboard has small keys and I have big fingers. I gave up, and almost stopped using it completely. Then, while on Switzerland I used the tiny beast to do some computations using Pari/GP and started to like it again. Although it lacks (mine, with my Debian installation) a good graphical plotter and visualiser (gnuplot does not work completely), it has yacas for symbolic algebra, Pari/GP for number theory and high precision computation and Octave. It also has C compiler... Forth, Lisp, Python, you name it. It is a nice gadget to keep close, but I wanted to use it for something. Here enters TaskWarrior, which fits like hand in glove with my Ben.

I add whatever task is lingering in my mind as I wake up, like buying something for supper?, with task add task. List my tasks by issuing task (note: add alias t=task in your .bashrc!). This is because the default action (which can be customised) is list. You can change this (and a whole lot of other options), but this is out of the scope of this short introduction. Read the full TaskWarrior manual here.

Adding tasks, setting projects and so on

You can add projects, due dates, priorities and tags to a task with the keywords project, due, priority and tag, followed by a colon. Like task add Do something today due:today priority:H project:Hey +tag1. You add a + to a tag to decide whether adding a tag, +tagname or removing a tag, -tagname. To mark a task as done, you issue task #taskID done. And to find tasks matching some criteria, you should task list criteria. You can be more specific, with task list project:specificproject, or use some custom show function like task list complete. I know that TaskWarrior at first is a little verbose... but you can abbreviate: t add Tasktoadd pro:Aproject due:tom pri:L.

Advanced listing and adding

There is also a lot of possibilities to list tasks, task list limit:2 will show only 2 tasks, task next will show you what to do next (based on priority, due date and last time acted). Even more, you can create your custom reports! I did one that listed the what to work next with a minimal view (to fit the screen of the Ben Nanonote), and it is as simple as adding this to your .taskrc file:

Listing and custom reports

I also did a custom report that showed me 10 tasks marked as fun, i.e. with the tag +Fun added to them. Thus when I wasn't like working on real work, I could do some fun work. To learn how specific reports work, you can ask TaskWarrior to show them, like task show next:

There are a lot more options to tweak, like color themes (TaskWarrior comes with quite a few, and you can write your own or download them from somewhere), adding task dependency, i.e. task 3 should only be done (and showed) after task 5, adding recurrence to tasks and quite a few more things. If you are interested in finding more about TaskWarrior, install it, play with it a little and start browsing the online tutorial.

I wake up, plug it to my MacBook, configure the network (see this post about Debian on the Ben Nanonote to read about networking it) and ssh from the Mac to it.

When I'm ready to go to my office, I unplug my Ben and head for my office. While on commute, I can check my tasks in the Ben or add one as needed just by taking it from my pocket. To save battery life I wrote two scripts named on and off, which just turn the screen on and off (the default timeout is way too long and with fbterm is hard to set up correctly, AFAIK) with echo 0 > /sys/devices/platform/jz4740-fb/graphics/fb0/blank

I get to my office, plug the Ben to my office computer, check for due tasks and work. When the day is over, unplug and head home... This way, it has always battery left, and my task list is always with me in a very browsable way. And I use my Ben for something!

All these reasons and use cases could work as well with org-mode. I could start emacs in my Ben at boot up, and work in the same way, but I wanted to try something different. So far, I'm very happy about how TaskWarrior works, and I was just happy with org-mode, mostly because org-mode works like a super organiser system. And I wanted something more minimalistic, for this specific aspect of my work-life. I.e., I won't be using vi without being forced to do it before hell freezes :).

The "Related posts" method I use involves Javascript, thus it doesn't work in the RSS feed. To view related posts, please refer to the original article. Thanks!

Yacas is an advanced computer algebra system, with its own programming language and a lot of handful operators available.

It comes handy when you need to do that odd symbolic computation which is too big to handle by hand (or you are plain lazy).

By looking at the wikipedia entry, I discovered an startling truth: as of 2009, yacas is no longer maintained.

I should have guessed, because yacas does not compile out of the box (the program needed to generate the manuals does not compile), and this led to some naïve patching (just removing all instances of said program...).

The compilation process for the NanoNote with OpenWRT follows the usual lines, with a twist. Yacas uses a host program to generate code, named mkfastprimes. And when cross-compiling, the resulting executable cannot be used.

To solve this small problem, you need to compile yacas in your host computer, and use the generated mkfastprimes to get the desired files. To do so, in the makefile I download the package, uncompress it, configure and make. Some minor problem is that configure generates several makefiles and folders in the working directory... and looks like OpenWRT's makefile system does not allow to change it. Thus I copy the original Makefile to Makefile2, let configure generate a new one, make yacas and copy back the original Makefile.

And this can get you yacas working in the NanoNote. You can download the Makefile and patches from here.

For additional instructions on how to set up a build environment, please refer to Qi hardware's wiki here.

The "Related posts" method I use involves Javascript, thus it doesn't work in the RSS feed. To view related posts, please refer to the original article. Thanks!

Preface: I have been using Linux since around 1998, when I installed Debian from scratch in my old Pentium II. I am more end-user than power user, but the computer I use most often (my netbook) has Linux in it by default. Also, my office computer is a Linux computer. And I am writing this in my MacBook. Which is not Linux, but at least it is Unix. What comes now is a personal rant, after a fight with my netbook. Probably not completely a Linux fault as an Acer one. But anyway, be warned this is a rant.

---

Linux is a time waster. It can come in two time-wasting fashion:

I have nothing against the good part. I even enjoy it, by learning to use gcal, or a2ps. I even wrote a (guest) post on why I think learning these side tools can be rewarding.

But the bad part... this always gets on my nerves. I don't mean that Windows is better in the bad part... but Mac OS is. Mac OS just works, but they have the best thing to be that way: all Mac computers are Apple controlled. Thus they can test everything and say 'OK'. Every hardware part will work perfectly and smoothly with Mac OS version N.

Linux has to work in almost all strange configurations possible... And this means big hardware fuss. You have a winmodem? You can't use a dial-up connection (that happened in my Pentium II days). More recently, you have an internal SD card reader? You can't hotplug it.

All started with an upgrade from Ubuntu 9.10 to 10.04. I assumed dist-upgrade was a good option, I wanted to upgrade my distribution. Then I learned, and was advised that it was a bad idea... But how could I know it beforehand? It was the first time I had to upgrade, in my office this is automatic, and previously I had so little content that overwriting with a newer version was not a problem. It looked like the best tool for the job. The ~6 hours process began, and finally, ended.

Without keyboard, without trackpad, without USB. I had something similar to a brick, under X11. I could boot in console mode and tinker to solve it. I googled and digged into forums, and after around 2 hours I found some maybe solutions. Tried two (both took a while of editing) and they didn't work at all. Gave up (I remember fighting against X11 back in my pure Debian days, and it not enjoyable) and tried to download the latest Ubuntu to install from a Flash drive. Did you know that Ubuntu no longer gives system boot images you can write in USB flash drives? It is either net-install (but how do you boot it?) or CD. It looks like a good move in these days where netbooks have no CD drive. In case you are interested, you can find a flash drive image here.

This process started around 11 am and ended 11 am the day after. Then add quite a few more hours for installing all programs I use, and configuring a few of them. Luckily, I use fluxbox and emacs where configuration back-up is a dream.

Roughly two days for a system upgrade.

Then, the other day, I had some pictures I wanted to post somewhere. Don't remember exactly, and put my SD card in my right-hand internal reader. Nothing. D'oh! I remembered I had to enable hotplugging when booting, by adding some lines somewhere. Googled for a while, found the solution I used back in my 9.04 days.Which needed writing in a non-existent file. Duh. Ok, look for a solution for Ubuntu 10.04. More googling. Several tries (which mean edit, reboot and try). Nothing, I only got the LH reader working. The right one, which is the easily removable, the LH is an expansion slot, isn't. Just from upgrading from 9.10 to 10.04.

And it took me one other day.

These are just two of my everyday examples, and I guess you could share a few of your own. More if you are a sysadmin, or do something with web servers. Share your story, it will relax you. At least, I feel a little better.

Related posts:
9 programming books I have read and somewhat liked...
My first port to the Ben NanoNote: gnugo
Power to the command line
Two weeks, still loving Fluxbox
Three dee (3-dimensional file system browsers review)
Gcal: the ultra-powerful command line GNU calendar
Acer Aspire One 8.9' + Ubuntu + Fluxbox

The "Related posts" method I use involves Javascript, thus it doesn't work in the RSS feed. To view related posts, please refer to the original article. Thanks!

I discovered this tool by accident, when a colleague asked me why I printed a .txt file straight without using a2ps first. My first reaction of course was thinking What? and promptly asking google.

Google answered with this page, and it was interesting enough to deserve an apt-get install. And indeed, it is great! Usually, when I have some straight text file I need to print, I use emacs old postscript-print-buffer, which is nice, but not as nice as all options a2ps has.

As default, a2ps prints in landscape format, 2 pages per sheet. Moreover, it adds a black border around the page and a header with the filename, time and who printed this. Below an example of straight a2ps:

Back a few days ago I had to print my flight details (it was too early to print the boarding pass), and you know, they look like a nicely formatted text file. But when Gmail tries to print it, it removes all white spacing, rendering it unreadable.

A solution? Creating a flight.txt file and pasting from the mail. Then

a2ps flight.txt -B --portrait --columns=1 -o flight.ps

What does all this mean?

And the same example with these settings:

The "Related posts" method I use involves Javascript, thus it doesn't work in the RSS feed. To view related posts, please refer to the original article. Thanks!

Last saturday I was in a porting mood, and tackled two interesting packages to have in the NanoNote: yacas and 4th.

Yacas (Yet Another Computer Algebra System) is a very interesting application to have in such a small device. The best calculator I have used is my old faithful HP49g. But in these days, it is bulky, heavy and slow. It was superseded by m48 (a HP48+ emulator) inside my iPod Touch. Quite nice: the HP48 is programmable, has graphing capabilities and quite a good computer algebra system.

However, the NanoNote is a computer. It has a fully featured (fully as in you can put whatever you want in it, if you are able to do so) operating system, and as such, can hold far more than a pocket calculator. And it has a real keyboard, not a fake touch keyboard as in the iTouch or a (horrible) plastic keyboard like the HP49g.

It turned out to be easier than I expected. The process needs a previous step, like Gnugo did. You have to compile it in your host system, and then use an intermediately generated file to cross-compile. I am almost done in writing a completely automatic Makefile that handles everything for you. As it stands, it is quite messy moving files around, but the program is usable.

The other, 4th is a multi-platform Forth interpreter and compiler. My first try of porting a Forth system was gforth, but it didn't work: gforth compiles itself (more or less), and in a cross-compilation process this is really hard to handle. The developers even acknowledge it in the INSTALL file.

Fortunately, 4th is easy to cross-compile on purpose and indeed, it worked like a charm. I didn't need to do anything more than write the Makefile and copying the libraries to the Nano afterwards. I'm also halfway writing a decent Makefile that copies these libraries where 4th can find them in the Nano.

The "Related posts" method I use involves Javascript, thus it doesn't work in the RSS feed. To view related posts, please refer to the original article. Thanks!

Do you own a Mac? I use both MacOS and Linux everyday, and found myself with a problem. Before I found Shutter taking screenshots in Linux felt like a PITA. In MacOS you can take region screenshots by pressing Cmd-Shift-4. And I missed this feature in Linux, until I cared enough to look for a solution.

Shutter was the answer, an open source screenshot grabber that fills my needs, and probably yours too. Installing is relatively simple, if you have an updated version of Ubuntu. If not, you'll have to manually add the launchpad repositories. All instructions are on Shutter's homepage.

Once you have it running, you can take region specific screenshots, do some adjustments to them before you are ready to use your shot. It is simpler than opening The GIMP and editing. It was heaven when you write tutorials with screenshots. When you tap the cross for selection, the shutter screen automatically disappears and you can select the location you want to shot.

As a bonus, I found Screenie, which is a quick 2.5D compositor à la Apple KeyNote. I have used it for my post about porting pMars to the Ben NanoNote. It is easy to use, drag and drop and some sliders. Afterwards by right-clicking you can then export the image to PNG. You can see an example below.

The "Related posts" method I use involves Javascript, thus it doesn't work in the RSS feed. To view related posts, please refer to the original article. Thanks!

Yes, you can! Adding colors to terminal output is possible. You already know it, from ls --color In this post I show you a script that does it, in a simple way. I don't have a full range of colors implemented, but you can find all here.

The "Related posts" method I use involves Javascript, thus it doesn't work in the RSS feed. To view related posts, please refer to the original article. Thanks!

The tutorial is quite good, but somewhat long, and lacks a few specific examples, so I decided to write just what I read in that tutorial, mixed with the uses I am putting it to, so it is more a Gcal use cases than a full blown tutorial like that.

I discovered commandlinefu.com a few years ago, while looking for a way to do... something. I don't even enter that often, although it is a brilliant place to discover how to do X in Linux/UNIX.

Among its all time greats I found some gems, and some others I discovered elsewhere, or even I made up.