tail -f findings.out

Handy Excel Tip: Open workbook in new instance

Samuel Huckins — Sun, 29 May 2011 21:04:50 +0000

I’ve written before about using Excel for some remote data collection and analysis tasks. One thing that often annoyed me in such situations was the inability to utilize multiple monitors. Whenever I wanted to break out a sheet, or especially when opening a new workbook, I found there is no way in default Excel to open the item in a new instance. It would always open in the same window, severely limiting potentially useful layout arrangements.

As is unfortunately common with Windows improvements… there’s a registry modification to solve it. All credit goes to OnlineTechTips: this article got me fixed up quickly.

Basically you just download a zip file, extract the contents, and rename the file whose title indicates the behavior you’d like in place. After this you just double click to run the file. This is a very convenient way to provide a fix, much easier than following descriptions of where in the registry tree to find what entries and what values to set them as. One of them even adds the option to open in a new Excel window to the right click context menu.

Charlie Sheen is a long way from philosophy– but the journey is winning

Samuel Huckins — Wed, 25 May 2011 19:52:13 +0000

After reading about an interesting phenomenon witnessed in a cartoon by Randall xkcd I decided to see how true it was that you could get to Philosophy from most any Wikipedia term via a reasonable number of clicks (here the devil waits, no doubt). The principle:

If you take any article, click on the first link in the article text not in parentheses or italics, and then repeat, you will eventually end up at “Philosophy”.

—Book of XKCD, Chapter 903 (See alt tag)

Being a good citizen of the Internet I started my search with the article on Charlie Sheen, hopeful that I was a mere two or three hops from my beloved Ivory Tower description. The truth, it turns out, is far more amusing. Follow the clicktrain:

Step 1: Charlie sheen -> Martin sheen -> Stage name -> Pseudonym -> Name ->
Step 6: Noun -> Linguistics -> Human -> Taxonomically -> Science ->
Step 11: Knowledge -> Facts -> Information -> Finite -> Mathematics ->
Step 16: Quantity -> Property -> Modern philosophy ->
Step 19. Philosophy

Now of course I have to write a script to facilitate exploring distances between terms in Wikipedia. After more important projects…

Show all available documents to new Google Apps – Docs users

Samuel Huckins — Sun, 15 May 2011 01:55:52 +0000

At my job we use Google Docs (via Google Apps) for internal collaborative documents and archives. One big issue with this platform is that when new folks are hired/given accounts and they go to view “All documents”– lo, and behold! it is empty. That’s right, if you go to “All items” and select Visibility -> YOURCOMPANY under More options as the filter, you will still only get “No items matched your selections”. This despite there potentially being thousands of documents that should fit your criterion.

As it happens these views only get populated based on creation or modification dates. In other words, new users only get to see documents that are modified or created after their account was created. By “see” I only mean visibility of the existence of documents in the Docs index interfaces and searches. The user can still view a document directly if they are sent a direct link (assuming they have permission). But the fact that you can’t even see documents you have permission to see in searches until they are modified is just insane. Google support verified this was a “feature”, and their reasoning is that in larger organizations new employees might be overwhelmed by a huge list of documents, things would be too hard to find. So the solution is to make it difficult to find anything!

Until the time when Google realizes this is annoying and silly, here’s a quick and easy (albeit not permanent) hack for this situation. Create a new collection (helpful: name = creation date) and add all current documents to it. This step counts as a modification to all the documents, and thus they will then appear in the list of all docs! The annoying bit is that next time you make a user you have to do the same thing.

This hack might not be feasible if you have high hundreds or especially thousands of documents, I am sorry to say. As far as I can tell you can’t drag more than 40-60 of them at once or it throws an error. If you do it in chunks it works fine though.

Lastly, perhaps you dislike messy lists of things you no longer need and want to remove the previous collection when a new user comes on and you create a new one. If you do that by simply deleting the collection you will also delete all documents in the collection! Really Google? You couldn’t add a damned checkbox to confirm the removal of child documents or just the collection container? Anyway, if you really care that much then there is a way. Click on one document, hold Shift and click on another document slightly past the bottom of your screen (after scrolling). This should be a reasonable number for your chunk size. Then hover over one of the selected documents, click Actions on the right, and click Organize. Here you can select the new collection to add them to it and deselect the old one so they aren’t in it. Once you are done going through all your docs the old collection should be empty and you can remove it safely. My approach is to create them with a name based on the date and just never remove them. If your users heavily leverage collections you might need to reconsider this.

Bag of useful Bash aliases and functions

Samuel Huckins — Sun, 15 May 2011 01:10:23 +0000

Here are a number of assorted aliases and functions for Bash that I’ve made in the last few months and found to be useful. Enjoy!

Sysadminish

These two print the current value of the swappiness parameter and tail recent syslog messages:

1 2	alias swappy="sudo /sbin/sysctl vm.swappiness" alias sysl="sudo tail /var/log/syslog"

This converts unix timestamps (e.g. 1305401097) that you might find in logfiles to a more easily readable date format:

1
2
3

function fromunixtime {
perl -e "require 'ctime.pl'; print &ctime($1);"
}

The first is simply a quick way to check internet connectivity. Next are some super short ping aliases to avoid Ctrl+c all the time when you just need to verify that something responds. Pass them a URL to get one or three ping attempts:

1
2
3

alias pingoo="ping -c 3 www.google.com"
alias p1="ping -c 1 "
alias p3="ping -c 3 "

Pass a webpage URL and it will download the source to a file named after the domain (if you’re passing a root page, e.g. http://www.cnn.com/) or the page name (e.g. contact.html):

1
2
3

function getsrc {
wget $1 -O `echo "$1" | sed 's/\/$//' | awk -F/ '{print $NF }'`
}

Sometimes you just need a big file. It doesn’t have to be interesting or contain any cosmological insights, but it does need to be a certain size. Pass that size in megabytes and this will give you just that, via a file full of zeroes:

1
2
3

function fakefile {
perl -e "print '0' x 1024 x 1024 x $1" > $1-MB-fake-file.txt
}

This handy one extracts all email addresses in a file and gives you a unique sorted list:

1
2
3
4

alias findallemails="perl -wne'while(/[\w\.]+@[\w\.]+\w+/g){print \"$&\n\"}' "
function findemails {
findallemails $1 | sort | uniq
}

Listing and searching

When you invoke ack-grep it’s case-sensitive by default. It also ignores certain files and directories, based on their nature in some cases and in all cases where it doesn’t know what the filetype is. This incantation removes those limitations and is shorter to boot:

1	alias ack="ack-grep -i -u"

This series allows you to list only directories or only files. You can pass a directory to list the contents or it will use the current directory as default:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16

# Feeds file and directory filtered listers:
function lister {
ls -l `if [ "$1" == '' ]; then echo '.'; else echo "$1"; fi`
}
# List only directories:
function lsd {
lister $1 | egrep '^d'
}
# More memorable version:
alias lsdirs=lsd
# List only files:
function lsf {
lister $1 | egrep -v '^d'
}
# More memorable version:
alias lsfiles=lsf

Media and Fun

These two require identify be installed, part of ImageMagick. Install the imagemagick package for your system first. Pass the filename of an image and the first will provide a range of information while the second will just provide the resolution:

1 2	alias imginfo="identify -format '-- %f -- \nType: %m\nSize: %b bytes\nResolution: %wpx x %hpx\nColors: %k'" alias imgres="identify -format '%f: %wpx x %hpx\n'"

Examples:

Sometimes you just need to jack in. Or at least fill your screen with impressive looking output:

1
2

alias rmatrix='echo -ne "\e[31m" ; while true ; do echo -ne "\e[$(($RANDOM % 2 + 1))m" ; tr -c "[:print:]" " " < /dev/urandom | dd count=1 bs=50 2> /dev/null ; done'
alias gmatrix='echo -ne "\e[32m" ; while true ; do echo -ne "\e[$(($RANDOM % 2 + 1))m" ; tr -c "[:print:]" " " < /dev/urandom | dd count=1 bs=50 2> /dev/null ; done'

Great iPad Apps

Samuel Huckins — Mon, 09 May 2011 00:44:22 +0000

I purchased an iPad 2 recently and have really been impressed with the versatility of the platform. I keep finding new uses for it, including some that are new roles instead of replacements for more limited options. I don’t have the bandwidth for a full review of the iPad 2 and all its great features. But I did end up going on a bit of an app spree, so I’d like to share some of the more interesting ones I’ve found so far.

Flipboard: This has become my favorite news application on any platform. You provide RSS feeds (or select from pre-entered sites) and Flipboard generates a digital magazine style view of the articles, complete with images and cleaned up typography. You can slide the item up to view the original article, email articles and share with various social networks from within the app as well. I’ve been surprised how much faster I can get through my daily news. One big complaint: get with it guys, add Read It Later support!
Read It Later: Lovely interface to read through your Read It Later list. There’s a flipboardesque style view available for an additional cost, which was a little disappointing. I do love RIL though, so I’ll probably fork over another few bucks to support them as well. [Update, 2011-05-09: I highly recommend buying the Digest view, and anything I said about it disparagingly can be safely ignored. Aside from turning your list into a great Flipboard view, it figures out categories for you, and does a fine job. Good form, Peter.]
~~Terra~~: An alternate web browser. I started using this because of the very intuitive and useful gesture navigation. Two fingers swiped left or right takes you back and forward in your browsing history and three fingers is used to switch between tabs. This greatly speeds up browser use. There are plenty of other cool features as well! [UPDATE, 2011-05-11: Apple has apparently told Readdle to remove Terra from the AppStore. Sure, there's some duplication with Safari, but everyone I've shown it to thinks it's a clearly improved experience. Curmudgeons.]
Adobe Ideas: A drawing app that makes your art a little more… artistic. It smoothes out lines and curves, so your shaky squiggles look a bit more intentional. Not bursting at the seams with features, but great for some quick brainstorming and illustrations.
TED: Awesome free app to view all current TED talks. Easy to search and get summaries at a glance. I’m watching a lot more of these now!
tChess Pro: Very full-featured and well-made chess application. You can use it to learn via the decent AI, move analysis, gameplay instructions, hints and other features. There are a variety of views, angles and appearances so you can view the board however you see fit. You can even email games back and forth and move the board around with gestures!
FlickrFrame: You provide your Flickr credentials and select a set or other criteria and FlickrFrame pulls up a full-screen slideshow of your images. Rotation speed and animation are configurable, so it makes a great photo frame from your iPad.
Netflix: Not as full-featured as you might wish, a cause of some complaints, but it still works wonderfully to stream from your instant queue, search for items, etc. And the iPad form factor definitely works out very well for a mobile movie center.
Speed Test: Just what you might think: easy and quick way to view your current download and upload rates. Good for checking on the go or verifying home network speeds.
TWC Max+: Was just looking for a simple weather app, but this was free so I gave it a shot. Turns out it’s pretty great! It provides current radar and other usual weather fare for as many locations as you’d like to configure. Alerts and some (rather useless) social media stuff is available as well.
CoderCheat: Free app providing a number of helpful and commonly used programming reference/cheat sheets. It doesn’t have a huge variety of languages, but even with just CSS, HTML, Javascript, MySQL, PHP, Python and Regex it was sure good enough for free!

Fast and easy bluetooth management on Ubuntu

Samuel Huckins — Sun, 27 Feb 2011 22:16:21 +0000

I wanted to connect the conveniently svelte Apple wireless keyboard and mouse to my Ubuntu media server. I don’t often have issues connecting peripherals to my Ubuntu systems these days, so I naively assumed this would be a walk in park. Well, it could have been, but I did not find the royal road on my first attempt. Here’s the right way.

Install blueman (assumes you are on 10.04):

1
2
3

sudo bash -c "echo 'deb http://ppa.launchpad.net/blueman/ppa/ubuntu lucid main' >> /etc/apt/sources.list"
sudo apt-key adv --keyserver keyserver.ubuntu.com --recv-keys 951DC1E2
sudo apt-get update && sudo apt-get install blueman

Now go to System -> Preferences -> Bluetooth Manager. This should bring up a handy utility:

For the mouse:

Right click on entry
Select Pair, passcode by default is “0000″
Once paired, right click again and select Connect to Input Service

And keyboard:

Right click on entry
Select Pair, didn’t ask for a passcode
Once paired, right click again and select Connect to Input Service

After these easy steps both devices started working flawlessly. As it should be.

Spruce up your desktop with the National Geographic Photo of the Day

Samuel Huckins — Sat, 19 Feb 2011 21:54:45 +0000

[Update, 2011-02-24] Reader Jason Coombs kindly made the necessary changes to get this working on Windows 7! Now I might be shamed into getting it working for Mac OS X as well His changes are here. I’ve updated my script linked below as well. I’ll add some nicer instructions for Windows users soon .[/Update]

I love checking the National Geographic Photo of the Day page to see what wonderful slice of nature or mankind some intrepid soul has captured and shared. Often a wallpaper-sized image download is provided. I wanted to automate the process of checking the page for a wallpaper link, downloading the file, putting it somewhere appropriate, and setting it as my background. The result is this script, which does all that and even:

Verifies you have at least 25% free disk space before continuing
Renames the image to the title shown on the page
Creates a dir to hold these images if it doesn’t exist

Windows users

Download that script and double click assuming Python is installed.

Linux users

Download it, set it to be executable (chmod u+x nat-geo_background-setter.py), put it somewhere like /opt. You could add an alias to run it whenever you fancy:

1 2	# Update path to where you placed the script: alias newbg="/opt/USER/nat-geo_background-setter.py"

If you want it to run every day on Linux you could add a crontab entry:

1
2
3
4

crontab -e
# Inside your crontab:
# Set BG to picture of the day
0 6 * * * /opt/USER/nat-geo_background-setter.py

Enjoy a bit of nature on your desktop!

Apple’s plans for the enterprise

Samuel Huckins — Sun, 30 Jan 2011 23:50:06 +0000

I don’t follow Apple very closely most of the time, but these facts I learned in close proximity recently struck me as pointing to an interesting possibility:

Apple has a giant datacenter with no clear publicly released purpose. It’s one big server farm and it’s online now.
Apple has a cloud OS patent. It’s most focused on the administration of such an OS with cloud-based, network-booted capabilities. The hardware platform isn’t specified, but they applied for this back in 2006, before the iPhone was released and well before iPad rumors took flight.
Apple is quietly and quickly moving into the enterprise space. Not through the server route, or even necessarily through MacBooks and iMacs, but through revolutionary alternative devices such as the iPad.

Perhaps Apple plans to release their own cloud-provided OS, enterprise-branded and customized. Securely connected to their sprawling new datacenter, Apple could provide businesses of all sizes with a productivity-enhancing and compliance-increasing platform that corporate users, IT teams, and management might all manage to get behind. There might be team-specific apps that are developed in-house or generated from a series of general template applications that provide commonly-needed functionality. From knowledge workers’ daily activity to retail scenarios various organizations could stand to benefit from devices, interface patterns and quality that users prefer enough to bring in personally.

Comments welcome!

Counting columns in a CSV from the CLI on Linux

Samuel Huckins — Sun, 09 Jan 2011 15:49:38 +0000

In wrangling some DB exports recently I needed the number of columns in multiple CSVs. It’s easy enough to just open them in OpenOffice and check the final column number, but I wanted a CLI utility. I even entertained some notions of wc having a magical option to provide this, but no luck there.

So here’s a simple function I added to my .bashrc to do the trick:

1
2
3
4
5
6
7
8
9

# Displays number of columns (intended for CSV-type files),
# pass in separator then filename:
function colcnt {
DELIM=$1;
FILE=$2;
echo "# of lines | Columns in $FILE"
echo "(>1 row means there are other separators in the file):"
awk -F"$DELIM" '{print NF}' "${FILE}" | sort | uniq -c
}

It’s in a function because Bash aliases don’t accept parameters. It uses the variable NF from awk which is the number of fields in the current record. The number of fields is sorted then unique entries are counted, providing a way to detect rows with different field separators or rows with missing columns.

Say you have a simple csv file called foo.csv:

1
2
3
4
5
6
7
8

"i am","some cols","yay"
"i am","some cols","yay"
"i am","some cols","yay"
"i am","some cols","yay"
"i am"|"messssssed"|"up"
"i am","some cols","yay"
"i am","some cols","yay"
"i am"|"messssssed"|"up"

After you add the function above to your .bashrc open a new shell (I found that sourcing .bashrc did not always update the function definition) and run this:

1
2
3
4
5

> colcnt , foo.csv
# of lines | Columns in foo.csv
(>1 row means there are other separators in the file):
2 1
6 3

This tells us that six lines had three columns and two lines had one column (as defined by the separator we passed in). If the CSV was consistently formatted (without the pipe-delimited line) you would only see one result row, being the total number of lines followed by the number of columns each line has (8 3).

If there’s a better/built-in way to do this please share! So far it’s been easy to use and helpful.

Reminder for mysql rubygem install on Ubuntu

Samuel Huckins — Tue, 04 Jan 2011 03:09:10 +0000

Quick easy steps to get up and running with ruby, mysql, sqlite, rails, and most especially the mysql gem on a fresh Ubuntu 10.10 64-bit system:

1
2
3
4
5
6
7
8
9

# mysql and dev packages needed for mysql rubygem:
sudo apt-get install mysql-server mysql-client libmysqlclient-dev \
libmysqld-dev
# Needed for ruby-debug:
sudo apt-get install ruby1.8-dev
# Rails package includes lots of goodies you'll likely want:
sudo gem install rails sqlite3 ruby-debug
# And finally the gem:
sudo gem install mysql