Favorite Chrome extensions

For Chrome/Chromium users, here are my favorite extensions.  What are yours?

Privacy/security related

• AdBlock -- probably the most important one here.  I can't stand browsing without it.
• Disconnect -- blocks tracking web page elements from major providers (Facebook, Google, Twitter, etc.)
• Explode -- expands most URLs created by URL rewriters (e.g. TinyURL)
• KB SSL Enforcer -- This one actually causes a bit of trouble, so you may not actually want to use it. It tries to ensure that you're talking to websites over encrypted channels by detecting whether there's an SSL version of each site.  Even though it has systems to override its decisions, it still ends up screwing up a fair amount of the time (especially with online banking) since many sites don't actually mirror their SSL sites correctly.
• Keep My Opt-Outs -- opts out of personalized Google ads

Web site integration

• BugMeNot Lite -- makes a little widget pop up in the Omnibox when BugMeNot is available
• Delicious Bookmarks -- bookmarks pages quickly in Delicious.  There are many Delicious extension, but I like this one since it shows you tag suggestions.
• Google Docs Viewer (by Google) -- fast access to all your documents, easy to open them in new windows
• Google Voice (by Google) -- shows number of items in your Google Voice inbox, makes all phone numbers click-to-call
• Minimalist for Google Calendar and Minimalist for Gmail -- these two let you disable many individual elements on Google Calendar and Gmail.  On my laptop where there's limited screen real estate, this helps quite a bit.

Browser functionality

• Google Quick Scroll -- when you search for something and follow a link on the search results, this presents the option of jumping to the location of the snippet on the page
• History of these2 -- lets you search history more easily by time and site
• RSS Subscription Extension (by Google) -- adds RSS widgets to the Omnibox

Three things for improving Facebook privacy

We're not gonna take it
Never did and never will
We're not gonna take it
Gonna break it, gonna shake it,
Let's forget it better still

-- We're Not Gonna Take It (The Who)

1. ReclaimPrivacy: everyone should be using this and scan periodically. This is an extremely easy way of detecting which privacy leaks you haven't plugged.
2. Remove Facebook Like buttons: a Greasemonkey script to rid you of these from non-Facebook pages. These buttons are nasty and let Facebook track you as you surf around the web.
3. Facebook Link Rewriter: another Greasemonkey script to stop Facebook from tracking which links you click on.

Sure would be great if I could stop doing these types of posts, eh?

Steal this idea: Off-hold notification

Woke up this morning I was all alone

Saw your picture by the telephone

I was missing you oh so bad

Wish I had you here to hold

All I've got is this touch-tone phone

So I guess I'll give you a call

-- Telephone Song (Stevie Ray Vaughan)

Here's another app idea for smart phone software developers. When I'm put on hold and I don't have a headset, I'd like to put the phone down until I am taken off of hold. I'd like it so that when the phone "hears" a loud signal over the line (it may need to do some noise cancellation in case there is noise on my end) the phone will vibrate or (optionally) switch to speakerphone.

This would mostly be for talking to people but a harder problem is handling customer service centers. (Ideally, customer service telephone systems would take your phone number and have an agent give you a call when they're free rather than wasting time on phone calls, but few companies do this currently) In these cases, it might be able to use a voice/music detector and check for repeated messages. Many of things you hear while you're on hold alternate between music and voice, so this might end up notifying you too much. With luck, it could be trained that some pieces were recordings (by memorizing sound "fingerprints" or something like that like) for use in later phone calls.

Quick hacks: webnumbr for fast scraping (of a single number)

Like the previous quick hack, I had little to do with this -- just demonstrating a tool. In this case, the goal was to graph the national average for certificate of deposit interest rates over time (you can find this information at Bankrate.com). At SHDH37, Drew told me about webnumbr which does exactly this (and was possibly developed at a previous SHDH) -- you pick a webpage and then select a single element from that page. webnumbr will scrape the webpage at an interval of your choosing and graph the data. I'm not wild about their graphing (prefer something interactive like this or Google Charts) but it looks like you can get the raw data in various formats (CSV, etc.) which would let you use other graphing methods, overlay charts, etc.

If you're interested, here are graphs for the average 1-year and 2-year CD rates.

Quick hacks: Yahoo Pipes for RSS feed filtering

In my Copious Free Time(tm), I've been working on some very quick projects. I'm writing them up since, while many of them are quite simple, they highlight some interesting tools.

I used to read the "Marmaduke Explained" blog at this website. It has since moved here but the RSS feed includes other entries as well. Using Yahoo Pipes, I was quickly able to create a new RSS feed which only includes entries that include Marmaduke in the title. See the pipe I made here (output available as RSS, JSON, email, etc.). Yahoo Pipes is, like the name implies, a series of tubes -- nodes in the graph feed perform operations and feed into other nodes. So, my point is that Yahoo Pipes is a cool piece of software and probably underutilized -- the basic filtering that I do is just the tip of the proverbial iceberg for what it can do (for example, there are "translate" and "location extractor" nodes). If they added basic scraping abilities, I'd likely use it even more (but see upcoming posts for different ways to do scraping...).

Why I dislike text messages

This article succinctly explains the main reason I dislike text messages: (emphasis not in original)

"...Wireless data service is AT&T's only bright spot, up a whopping 26% per customer. How so? As any parent of teenagers knows, text messages are 20 cents each, or $5,000 per megabyte. After the first month and a $320 bill, we all pony up $10 a month for unlimited texting plans. ..."

I have been waiting for someone to shake things up and get cell phones using instant messaging and/or email rather than text messaging which, in my mind, consists entirely of disadvantages (expensive coupled with payment on receipt, limited message length, impossible to know if messages are even received, ...). With luck, this will help as well.

Steal this idea: Google, optimize those links!

You know, those links at the top of Gmail and related services. Mine say:

Gmail Calendar Documents Photos Reader Sites Web more ▼

Yours do too? How strange. I would think these links should be personalized if not customizable. Google knows which ones I click on since I'm logged in all the time. Why not get rid of the ones I never click on (Sites, Photos, and Web -- if I need to do a web search, I'll use the address bar in my web browser) and show the ones that I always have to find in the ▼ like Scholar. Some people would probably object to a fluid UI (i.e. if Google adjusted these automatically) and I think the preferred methods these days are to have computers suggest changes but not perform them without user intervention. Maybe there's some plugin/extension/religion that lets you change these? (not that this is really a high priority...)

Steal this idea: Phone scammer/spammer warnings

Hear my phone ringin',
sound like a long distance call
-- Muddy Waters, "Long Distance Call"

Why telemarketing is an opt-out rather than opt-in process is beyond me. Nevertheless, if you do get a call from a number you don't recognize, you may want to run it by whocalled.us. I recently got a call from a telemarketer (see my report here*) and had an idea for a program to run on the phones of the future. When I say phones of the future, I mean roughly tomorrow since this idea is so simple it should be very easily doable on an iPhone, Android phone, or something else easily programmable. At least, I would hope this is easy to write (I haven't ever used these phones' APIs).

The idea is this: When I get a call from an unknown number, my phone should look it up on whocalled.us. Possibly there would be a reputation system and it could automatically report good calls to the site (though there are some potential privacy issues here which would need to be addressed) and give you the option to mark a previous call as a scam. Integration with donotcall.gov would be nice too.

* Incidentally, this phone number appears to be running multiple scams. The number is used to connect people with various scams so that the companies actually doing the scams are harder to trace. That's why the representative on the other line said that I had called her.

Technical: Getting Skype working with Bluetooth on Ubuntu with AMD64

If you want to use Skype with Bluetooth headsets on recent versions of Ubuntu Jaunty with an AMD64 architecture, this article might be of interest. If those words didn't make any sense, this article is definitely skippable.

When using Skype with a Bluetooth headset, I was getting error messages of the following form:

ALSA lib ../../../src/pcm/pcm.c:2165:(snd_pcm_open_conf) Cannot open shared library /usr/lib32/alsa-lib/libasound_module_pcm_bluetooth.so

Since Skype only exists in 32-bit mode currently, you can only run it in AMD64 using the linux32 command. However, this requires 32-bit versions of all the libraries that Skype would need, including bluez-alsa if you want to use a Bluetooth headset. Ubuntu currently attempts to handle this with their ia32-libs package but the current version of this package is missing its bluez-alsa libraries (as indicated in the error message). This is the bug that has been reported here.

I have a horrible hack to fix this problem, mostly because I can't find where ia32-libs development takes place nor the Ubuntu development process. I downloaded the 32-bit version of bluez-alsa and used alien -t to convert it into a tar.gz file. I then extracted its contents and made an extremely simple Makefile with this content:

install:
cp ./usr/lib/alsa-lib/libasound_module_ctl_bluetooth.so /usr/lib32/alsa-lib/
cp ./usr/lib/alsa-lib/libasound_module_pcm_bluetooth.so /usr/lib32/alsa-lib/

Next, I created a Debian package with checkinstall which you can find here. To my knowledge, these files are missing in Ubuntu Karmic too, so this might help those users as well.

Takeaway message: If you want to try my hack, download my new package (which I admitedly haven't been able to test on any other machines, so any feedback is welcome) and use your favorite package installer (the graphical one is probably the easiest) to install it.

Steal this idea: Musical intersections

Listening to you,
I get the music.
Gazing at you,
I get the heat.
Following you,
I climb the mountains.
I get excitement at your feet.
-- The Who, "We're Not Gonna Take It"

There are a couple times when two or more people want to listen to some form of streaming music (road/train trip, shared office space, etc.). Streaming radio sites like Pandora should have the option of combining multiple peoples' musical taste and come up with music that both people would like. Being able to generate a road trip playlist for portable music players would be nice too, though there are the obvious copyright issues.

I imagine coming up with good shared playlists is trickier than just taking the intersection of the songs/musical features. Ultimately, it would depend on the model of listeners tastes. Of course, in many cases, these intersections can be small (my freshman year college roommate and I both liked Stevie Wonder but had little else in common musically).

Wait, what?

Marge: I don't understand why we have to build a ray gun to aim at a planet I never even heard of.

Homer: Don't blame me, I voted for Kodos.
-- "Treehouse of Horror VII" [snpp]

McCain and Obama are looking more and more similar each day! Their automatic image picker seems to have some problems telling the two candidates apart... (seen on huffingtonpost.com, emphasis not in original)

Steal this idea: Scratch paper printer

Somewhat counterintuitively, the Computer Science department at Brown produces a ton of paper (literally a ton if you choose the Right time scale...). A good portion of our printouts only use one side of the paper (though people are generally good about printing things duplex, 2-up, etc.). Much of this scratch paper is recycled immediately while some is put in a scratch paper pile. After trying to reduce our printouts, what we can do to lower our paper usage? Last year, I made a "notebook" out of scratch paper. a folder, and some clips which worked out very nicely for a course. While I like this method and would like to see it done on a large scale, it seems unlikely to reuse much of our scratch paper supply.

One idea that I'd like to investigate is a "scratch paper printer." This would be a printer stocked entirely with scratch paper. There are couple potential issues. The main issue is that I've been told that printers jam more frequently on scratch paper[1]. I want to find out if this is true for all or just some printers. Ideally, we can find a (potentially slower) printer which doesn't jam as much on scratch paper. A lesser issue is keeping the printer stocked since scratch paper doesn't come as easily as store-brought reams, but since this can be done on demand by users, I'm less concerned. Please contact me if you have any comments on this proposal, since I'll put pitching it to the department in the coming future.

[1] It might be interesting to try to design a printer which is more robust to paper jams (steal this idea, please) but beyond what I'm looking for at this point.

Copying links from Google search results

You may have noticed that when you copy links from Google search results, you don't get what you expected -- when you hover over them, your browser says one thing, but when you right-click on them, it says another. For example, searching for snappletronics, Google gives you the code for a link like this: (emphasis not in original, obviously)

<a href="http://snappletronics.blogspot.com/"
class="l"
onmousedown="return rwt(this,'','','res','1','AFQjCNGxo5T82OPvPMIGpTmHTE3JzTBwzg','&sig2=vMD0Ey969KDVrwhFXwVgBw')">

What does this mean? It means that when you right-click on the link, some Javascript function gets to run and change the link, so the link turns from Snappletronics to Snappletronics (watch the tooltip or your browser's status bar). The second link tells Google that I clicked the link and maybe other identifying information. I believe this surreptitious link-changing is a pretty strong violation of "Don't be evil."

Fortunately, there's help if you use Firefox. If you have Greasemonkey, this script will remove their hooks and prevent them from changing the link on you.

Update (11.12.2010): The above script is now outdated. Depending on when you read this, this script might be a decent replacement.

Free the Phones

This article, Free My Phone, does a good job of explaining many of the things that are wrong in the telecommunications world. I hope that Google or (much less likely, unfortunately) OpenMoko can break things open (I intend to buy an OpenMoko phone in a year or two, provided they have reasonable stability). Also, the problem is slightly worse than he describes for Sprint -- they couldn't even transfer numbers from my old phone to the new phone, let alone the rest of my data. Also, they're responsible for the br0x0red Bluetooth implementation which prevents my phone from synchronizing contact information (or anything else) with my computer.

Gmail mobile app 1.5

I just deleted Gmail mobile 1.5 from my cell phone, keeping version 1.1.1. The new version (1.5) is way slower, as mentioned here. Startup and shutdown times are noticeably longer. Two possible factors:

1. The file was about 3x larger and might take longer to load into memory (though I don't really know much about the memory hierarchy of phones, so this may be bogus). The new version does have a bunch of new features (draft saving, etc.) so the new bloat might be warranted.
   
2. When starting up the new version, it displayed a list of emails immediately. The old version displayed an empty list and fetched the emails, so it looks like the new version might have cached emails locally. This might explain the slow startup and shutdown times. However, considering that usually there is at least one new email when I check Gmail from my phone, caching the titles of emails from last run is not really a savings.

The weirdest thing is that Google claims here:

New! Our new Gmail version is even faster and uses less data than before.

So maybe it's just my/some phones?