Dissertation Abstract and Update
A number of wonderful, attentive, concerned friends and family members have asked for a dissertation update, and here it is. Thank you for thinking of me! Now all of you who wanted to know but were afraid to ask can know too. The latest short abstract is:
In the fall of 2005, drivers in a small midwestern city began crossing over an interstate on a new kind of bridge. The bridge beneath them looked like other bridges carrying city streets over the interstate, but this bridge could bend. It couldn't bend like Gumby, but it could bend like steel. Building the deck of this bendable bridge involved a state transportation department, a university research lab, and several private contractors. Given the complexity of construction projects, the challenges in doing innovative construction work, and the potential pitfalls of collaboration projects, the success of the bridge is surprising. This dissertation explores how the team managed to build a bridge with a remarkable new kind of deck.
Existing scholarship provides insight on the problems that plague projects and collaborations and identifies many mechanisms to help meet these challenges. My analysis suggests that the bridge project avoided possible problems common in projects such as (1) loose coupling among actors in a project limiting the information sharing that occurs and (2) procurement processes that encourage builders and clients to see one another as adversaries through (a) social language and its associated attention to others, (b) the flexibility and localized control loose coupling affords, and (c) the motivating influences of affect. This study will combine and extend theories about social capital, creative projects, and loose coupling in order to better understand the nature of collaborative projects involving multiple communities of practice and how those projects can be successful.
I've written at least 68 good pages and probably about 50 not-so-good ones that will eventually work their way, in part, into the good stuff. I have a few (< 10) interviews remaining, and that means more time in analysis. I'm on target for my personal deadline of a spring/summer defense and am actively seeking new opportunities beginning summer or fall of 2009.
New (to me) and Useful: Visual Search
Many of you know that online searching frustrates me. That frustration results, in large part, from the ugliness of search results and their inability to let me describe what kind of thing I'm looking for. Pattern matching, result sorting, pish posh. I want internet finding. Enter visual search. Thanks, machine learning and digital image processing!
I've been hunting for new black shoes that I can wear with both jeans and suits for years. Years! Molly, my incredibly helpful librarian friend, introduced me to Like.com. Information Aesthetics, a fantastic blog, introduced me to Modista.
Like and Modista allow you to search for shoes, watches, sunglasses, handbags, etc. visually. On Like, you enter some search terms like "black boots" and then get to click on ones you like and see many more that are visually similar. You can filter by color, style, brand, site, all kinds of useful categories. Modista starts with a kind of item, such as "eyewear" and then shows you a variety of styles. Clicking on examples refines the search. Modista uses fancy sliders for limits on price and other continuous variables.
I found black boots at Zappos that I'm keeping; I almost love them. I'll use visual search again the next time I need shoes or accessories, and I look forward to the improvements coming to Like's apparel searches. Searching by sight is so much more fun than plowing through those boring text lists!
Why not Time Machine?
A couple of commenters asked why I use the ChronoSync + SuperDuper! combination instead of just Time Machine. The main reason? Time Machine uses too many resources. It's also slow. For awhile I avoided it because I wasn't sure how to make a bootable backup, but Mac OS X hints has instructions.
I don't always have my external hard drives plugged in since I'm rocking a laptop and am pretty mobile. Time Machine complained every hour, on the hour, that it couldn't find the drive it wanted for long enough to annoy me. Eventually it stops complaining about not being able to find the drive it wants.
Even if you leave the drive plugged in while working at your base location, for me it's my home office, Time Machine sucks up resources to do those intermittent backups. Even when I'm working on my dissertation, my data is not so mission-critical that it needs to be backed up every hour. Mac OS X Hints has a solution for changing the backup interval too.
ChronoSync can do in 39 minutes what it takes Time Machine over an hour to do. SuperDuper! beats the initial setup by about 20 minutes. So, the ChronoSync + SuperDuper! setup saves me resources, time, and headache.
One more thing - I have an Airport Extreme router, and I hang a hard drive off it via USB also. That drive is open to anyone on our home network. Apple's not kidding when they say Time Machine does not support network backups except to Time Capsule. When I tried using Time Machine to backup to that USB drive off the Airport Extreme, it would run my CPU up to about 80% and break many of my network connections. You may have better luck there. I didn't troubleshoot or try to fix it; I just gave up.
I ordered a rocstor ROCRAID from mwave last week, and that should be here on Tuesday. I'll try out RAID storage for my stuff and see how that goes. It has FireWire connections too, and I'm interested to see how much faster that can really be. I really don't want to have to give my laptop to Apple for a week. They won't let me keep the hard drive and send it in with a different one, and they won't give me a loaner. So I paid $2500 to have a laptop 98% of the time. Would I get it 100% of the time if I'd spent $3000? Sorry for the minirant, but having to get my MacBook Pro's fan fixed is what prompted this latest round of backup chatter.
Tweeteorology.com
tweeteorology.com just launched. It's the first project from my new company - Magical Pork. Tweeteorology shows tweets (posts to Twitter) about the weather. You can limit the search by location. I am actively developing Tweeteorology and welcome your suggestions.
Thanks, Molly, for the tweeteorology idea!
Backing up My Mac
My posts about cloning at Boot Camp drive and swapping hard drives in a MacBook are the most popular, according to Google Analytics. Today's special treat, therefore, is a quick overview of an easy, successful, inexpensive backup routine for all your Mac owners.
Our goals:
- Create a bootable backup of a Mac,
- Update that backup occasionally, updating only the stuff that's changed rather than copying the whole drive again, and
- Not spending much money.
What you'll need (besides a working Mac):
- SuperDuper! - the free version
- ChronoSync - latest version; $30 until v4.0 comes out, then price goes up to $40. v4.0 is already late, so the price could change any day.
- An external hard drive
* - something bigger than your internal drive. I use a 3:1 ratio of external:internal storage
Steps to make bootable and differential backups:
- Make a bootable backup using the free version of SuperDuper! (see the SuperDuper! user's guide for compete instructions; it comes with your download)
- Schedule syncing with ChronoSync; tell ChronoSync you're syncing two Macs (see the ChronoSync help menu for complete instructions)
You'll see that "cannot locate target" message in the ChronoSynce window if you forget to plug in the external drive. - Remember to leave your computer running when your backups are scheduled. I set a reminder in Google Calendar. It reminds me every week to leave my laptop running overnight so it'll backup.
Making the bootable backup with SuperDuper! took just over 3 hours on my MacBook Pro with 192GB of files. ChronoSync usually takes about 2 hours to sync - only about 45 minutes if I exclude my VMWare virtual machine from the backup. My virtual machine is over 40GB now, and since it always appears as "changed" in ChronoSync's analysis, it gets backed up every time my sync task runs.
I'm experimenting with backing up over a network using rsync and scp. When I get that worked out, I'll post the instructions here. I'm obsessed with backup now that my dissertation is coming along. I don't want anything to happen to those files (or my pictures), so I have them (both) in quadruplicate and on 2 continents. I suggest you set up something similar for the files that are really important to you.
*Notes on shopping for hard drives
Prices on external hard drives are dropping pretty fast, and you should be able to get a good deal. You can build your own external drive by buying an enclosure and an internal drive, or you can buy a ready-made external drive. Sometimes internal drives go on such a huge sale that building your own is cheaper, but ready-made externals are now competitively priced. Building your own has advantages like making the drive swappable and maybe helping you get a FireWire port for less money, and it's not hard. If you find a great deal on an enclosure and internal drive, go for it!
Where are these deals? I recommend checking the forums at Fat Wallet first to see if any site is having a big sale. I often buy drives from NewEgg because they have great prices, really fast shipping, and reasonable return policies. I use only Western Digital and Seagate drives. PC Mag has a good review site for hard drives.
Storing, sharing, editing data
I'm starting a new research project on which I collaborate with two other people, one in Ann Arbor, MI and one in Phoenix, AZ. We use Macs and PCs. We have no budget for software. We're likely to have a bunch of qualitative data to keep track of and share.
My first task is to gather some information on potential participants for our study. I've spent enough time working with databases to prefer them to files and folder structures for these purposes. Databases have advantages in that they can store relational information, can sort data, are easy to search, can be viewed and edited by more than one person at a time, the list goes on. Bottom line, I want a database. I starting by building a Drupal site to keep track of the data about those potential participants - who they are, where they are, how to contact them, how they're related to each other, what my thoughts are about them. I ended up abandoning Drupal to build my own MySQL/PHP website that stores and displays the data; I also built insert and edit pages to ease those data functions.
This is my third try-and-abandon with Drupal. I can understand how Drupal might make maintenance easier for non-technical users, but as a quasi-technical user, Drupal gets in my way at every turn. My MySQL database has 9 tables with 2 - 10 fields. I would have to add all <90 of those fields by hand in Drupal. Then, I'd have to create views to see them and views to edit and views to add/insert. How is that better than just building a PHP site myself? If most of the work is during setup, how does Drupal save any time or frustration at all? It seems like Drupal has taken the command line, coding aspects of building my own database-driven site and replaced them with a convoluted web-based GUI. I'm not sure I see the point or the cost savings there.
The real problem here is not that I still haven't found Drupal useful or advantageous. The problem is that I still don't have a great, easy, usable way to enter, edit, store, and share data with my colleagues. Even after I build this mySQL/PHP site by hand, I'll still have to figure out a way to get the data back out so we can analyze it. That opens a whole host of problems whose current solutions such as NVivo and Atlas.ti are expensive ($240 and $119 for students, respectively). Opportunities abound for helping qualitative researchers capture, store, share, and analyze their data collaboratively. What price point would be appropriate here? If you're a researcher, how much would you be willing to pay for a personalized, secure, web-based data sharing solution? Would anyone else even want such a thing?