Managing multiple Gmail accounts

Discovered MailPlane (Mac OS X only) today that lets you easily switch between multiple Gmail accounts and see the number of new emails at a glance. It works by overlaying desktop functionality on top of the embedded WebKit/Safari core. For example, you can drag and drop photos, auto resize photo prior to sending, and copy/paste images directly as attachments.

It's not free or open source but overall feels like a high quality app. The time saved in attaching images/files alone is worth the (current price) of $25, and it's a must have for people who use multiple gmail accounts.

Subversion / Trac / Project Management Hosting Services

Code Spaces, Assembla, and Springloops appear to be the more mature Web 2.0 ones.

Overall: Assembla has the easiest to use UI and integrated project management functionality.
UI interaction: Assembla.
Visual design: Springloops.
Project management: Assembla feels the most integrated. Springloops relies on users' own BaseCamp account.
Web 2.0 feel: Springloops. Code Spaces feels more like Yahoo Mail's definition of web 2.0 interaction than 37 Signals sites.

Assembla also aims to create a market place that easily connect coders to project needs.

More are listed at
http://snook.ca/archives/servers/hosted_subversion/

Geocoding IP addresses with GeoIP City (GeoLite)

First install both the GeoIP C Library and the Python bindings.

Download the GeoLite/GeoIP City binary data file, and install the .dat file::

$ mv GeoLiteCity.dat /usr/local/share/GeoIP/GeoIPCity.dat

In a Python shell:

>>> import GeoIP
>>> geoip_lib = '/usr/local/share/GeoIP/GeoIPCity.dat'
>>> gi = GeoIP.open(geoip_lib, GeoIP.GEOIP_MEMORY_CACHE)

>>> print gi.record_by_name("google.com")
{'city': 'Mountain View', 'region': 'CA', 'area_code': 650, 'longitude': -122.05740356445312, 'country_code3': 'USA', 'latitude': 37.419200897216797, 'postal_code': '94043', 'dma_code': 807, 'country_code': 'US', 'country_name': 'United States'}
>>> print gi.record_by_addr("64.233.167.99")
{'city': 'Mountain View', 'region': 'CA', 'area_code': 650, 'longitude': -122.05740356445312, 'country_code3': 'USA', 'latitude': 37.419200897216797, 'postal_code': '94043', 'dma_code': 807, 'country_code': 'US', 'country_name': 'United States'}

Some quick tests showed that GeoLite City correctly geocoded 4 out of 6 (somewhat random) samples I threw at it, compared to the GeoIP web service, which presumably uses their full GeoIP database. For example, it wasn't able to geocode Microsoft.com (207.46.197.32) and matched my work IP address to LA when it should be in the Bay Area.

Communication Arts: Interactive Annual 13 Winners

Wow! That's my reaction after seeing this year's Interactive Annual winners, which have two mobile-related/inspired sites:

• Motofone: "More documentary than advertisement, Motofone tells the story of a team of designers who traveled to three continents to develop a phone suited to the harsh conditions of emerging markets. Video, audio and animation come to life through an interface inspired by dust storms."
• Urban Silo: "With a mobile-inspired interface, this portfolio site for writer and creative director Jim Elliot serves as a storage facility for the past and an incubator for future ideas. Along with print, Web, TV and radio work, it also features an archive of personal projects."

Geotagging Photos

This is my N-th try at automating the geotagging my photos using GPS. I've tried a variety of utilities and web-based tools and found them having too many steps/clicks and don't deal with timestamps very well.

In order to batch geotag photos using GPS data, the timestamps on the camera must be matched to the GPS. Timezone is an issue here because cameras don't generally record timezones. A more minor issue is clock drift, which can often be ignored unless you're traveling at very high speed.

On the Mac OS X, GPS Photo Linker is a well written tool that supports timezone adjustment, batch stamping, and manual stamping. It looks up city/state/country information and adds those to EXIF along with lat/lon data. It also links to Google Maps so you can view the location for each of photos.

Mobile apps for dating

Match.com updated the mobile web version of their site. When will we see mobile versions of the up and coming Engage and Journey of Yuan?

More Click Analysis

DIY heatmap, Crazy Egg review, and out-sourcing eye tracking studies.

Click Analysis

Not quite clickstream analysis showing you the click paths, but a few company offers heatmaps: ClickeDensity, ClickTale, and Crazy Egg.

Email processing in Ruby without Rails

Connect to Gmail using Ruby without using the Rails framework.

MMS Photo Upload

MMS photo upload is really a multi-part email with photo attachment sent from a phone. To receive and process MMS photo on a server, one can either run procmail locally or use a POP/IMAP fetcher.

"Stress-free Incoming E-Mail Processing with Rails" describes how to process incoming email with Rails. MMS2R can then be used to process MMS, stripping out ads and other junk that a network provider (e.g. Sprint) may insert into the message. Sprint even goes a step further and stores the photo on their servers instead of in the message itself.

"Experimental Mobile Gateway" talks about SMS processing in Python and MMS processing to post blogs to LiveJournal.