Assessing persona prevalence empirically February 21, 2009Posted by cnchapman in Blogroll, Market research, Technology, User research.
1 comment so far
I just obtained permission to post our latest paper on Personas. We argued previously that the personas method should not be considered to be scientific, and that a complete persona almost certainly describes few people or no one at all. In the new paper, we present a complete formal model, and evaluate the prevalence of “persona-like descriptions” with both analytical methods and empirical data. Full paper on persona prevalence.
There are two key implications here: (1) if you want to claim that a persona describes real people, you need strong multivariate evidence. (2) Without such evidence, we provide a formula you can use that will give a better estimate than simply assuming something. We show how this formula has a better than chance agreement with 60000 randomly generated persona-like descriptions in real data with up to 10000 respondents.
None of this says that personas are not inspiring or useful. It just says that they cannot be assumed to have verifiable information content, unless that is demonstrated empirically. As for alternatives to answer key design and business questions using empirical data, check out our paper on quantitative methods for product definition.
Personas February 8, 2009Posted by cnchapman in Blogroll, Market research, Technology, User research.
add a comment
One of my papers from 2 years ago is still causing discussion: “The Personas’ New Clothes: Methodological and Practical Arguments against a Popular Method” by me and Russ Milham. Email from researchers I didn’t know led me to look up citations, and the article appears to be commonly cited when people present criticism of the personas method. Google search. The paper itself is here.
There are a few misunderstandings of our position out there. Our basic argument is simple. Persona authors often make two claims: (1) personas present real information about users; and (2) using personas leads to better products. In a nutshell, we argue that neither claim has been supported by empirical evidence; rather, the claims for personas’ utility are based on anecdotes, generally from their own authors or other interested parties (such as consultants selling them).
This does not mean that personas are bad, but they cannot be taken at face value. As researchers, we suggest that persona authors should either provide better evidence (and we suggest how) or make weaker claims.
Some persona users don’t make claims about their personas’ usefulness or correspondence to reality; they simply say that personas might be helpful for inspiration for some people or teams. We take no issue with that, as long as they don’t forget those caveats and reify the persona. Unfortunately it is probably very difficult for people to read a persona and not think that it describes a user group.
We’ve recently published empirical work on (quasi-)persona prevalence using several large datasets, demonstrating that once a description has more than a few attributes it describes few if any actual people. I’ll put that paper up as soon as I get reprint permission. (If you have access to HFES archives, it is “Quantitative Evaluation of Personas as Information”, Christopher N. Chapman, Edwin Love, Russell P. Milham, Paul ElRif, James L. Alford, from HFES conference 2008, New York.)
What should one do instead of personas? I advocate stronger empirical methods that have more demonstrable validity.
New papers on user research February 7, 2009Posted by cnchapman in Market research, Technology, User research.
add a comment
Just uploaded 2 new papers on user research. First is work on a multi-factorial product interest scale, designed to be easily administered in survey format and applicable to consumer products. See the abstract on my “papers” page, or get the file directly: wip337-chapman.pdf
Second is an overview of quantitative methods that are helpful in early evaluation of product needs and strategy. The abstract is on my “papers” page, or the complete file is chapman-love-alford-quantitative-early-phase-ur-reprint.pdf
I’ll be uploading more papers soon.
Printing in landscape from Mac to HP LaserJet 1200 on Linux print server November 27, 2008Posted by cnchapman in Technology.
add a comment
Yes, this is a very specific post but having seen many questions about this online, I wanted to post my solution.
First the setup: HP LaserJet 1200 connected to a Linux machine running CUPS print server, sharing the printer out to my home network. Client machines include a Mac OS 10.3 notebook, Win XP desktop, and Win Vista notebook. All are set to use the HP 1200 Postscript and/or PCL driver that came with OS.
The problem: Printing from the Mac to the HP1200 in landscape mode (in any App: Word, Excel, iCal, etc) prints in portrait mode instead, with the edges of the page truncated. Could not find a driver update, and deleting/reinstalling does not fix it.
Solution that worked for me:
(1) go to system properties | printers and “add new printer”. Add it as a “windows printer”, browsing to the workgroup and picking it. The printer should be detected and show up. (If you’re not using the Linux CUPS server, this step will differ. Browse to the printer in the way that fits your setup.)
(2) Give it a name you’ll remember, such as “HP1200v2”. Now the key part: for printer model/driver, do NOT use the LaserJet 1200 driver. Instead, use the “HP LaserJet 6 gimp-beta” driver. This should be available by default in Mac OS.
(3) Click OK, etc., to finish. Test it. Go back and delete your older HP1200 printer setup, and make the new one the default.
Update on the Toshiba IK-WB 15a August 14, 2006Posted by cnchapman in Technology.
1 comment so far
As promised, I wanted to share some of my experience with the new IKWB 15a. In a nutshell, after using it for 3 days, it seems to be everything that the 11a camera should have been. The 15a is very stable (no reboots yet) and responsive. Images load and are streamed rapidly. Image quality is very good. Low light performance is especially good — with auto B&W on, it switches from color to B&W images in low light. These yield crisp, satisfactory photos outside in the middle of the night with street lights on.
Note that I use the camera exclusively with Cat 5 cabling, so performance comments are about 10/100 Ethernet access, not Wi-Fi.
There are a number of changes to the firmware, mostly to add a few functions and make the HTML pages easier to navigate. One change that I especially appreciated: the FTP recording option can now use a fixed file name. This option has the camera initiate an FTP upload of images either on a schedule or in case of an alarm such as motion detection. In the 11a camera, the uploaded files were named with a time & date scheme, such as “LV-NWCAM1-20060815-010544.jpg”. In the 15a, that is still an option, but it is also possible to give a fixed name to the file, which is overwritten on each upload. For instance, something like “webcam.jpg”. That is helpful if you want to post the image directly to a web server.
In my case, this FTP recording feature would make it possible to feed the image directly into a directory where “motion” can examine it. In other words, it makes the most essential feature of my self-written recording program (see previous posts) unnecessary! On the other hand, you’d have to set up and run an FTP server, which I prefer not to do (even on my LAN) because of security holes. Might be a future possibility to put a standalone FTP server in my network DMZ, though.
One change on imaging is that there is no longer an option for 800×600 images. Maximum is the same at 1280×960, but the next lower option is 640×480. That is OK with me, since 800×600 doesn’t work well with “motion” anyway. The faster responsiveness of the 15a makes the 1280×960 images stream very nicely, either as mjpgs or static jpgs.
There is also additional attention to detail on the hardware side. The size and appearance of the camera are unchanged, but I noticed two nice details. First, the 15a camera comes with an extension power cord on the DC side. This is handy for exterior mounting in particular — you can put the extension cord through a wall and then if the AC/DC brick ever dies, you won’t have to remove the cord from the wall in order to replace the tranformer. Second, there is a small wrapping strip that fits inside the power/ethernet connector area on the camera, and wraps around the power cord. This helps hold it in place, so the power connector won’t come loose.
Overall, I’m very satisfied with it, and recommend it highly. The combination of great pictures, interior/exterior mounting, improved firmware, and a nice lineup of features make it a great deal at its price point (I found a great deal (now gone) at $479, but anywhere up to $550-600 would be satisfactory).
Problem with Delphi Indy 10 components August 13, 2006Posted by cnchapman in Technology.
I added a new camera to my setup this weekend, and got one of the new Toshiba IK-WB 15a cameras — that’s the new model that replaces the IKWB 11 line. So far, it seems like a very good camera. It’s more responsive and seems more stable than the IKWB 11a series.
However, when I pointed my self-written camera recording program to it, I couldn’t get a picture. I guessed that it might be due to a new authentication scheme in the camera, and perhaps my older Indy HTTP components (used in my program) were not compatible. So I upgraded Indy 10.0.76 to the latest 10.1.5. (http://www.indyproject.org/download/Files/Indy10.html)
When I did that and ran Delphi again, I got this error: “Cannot load package ‘IndySystem50 ‘ . It contains unit ‘filectrl’ which is also contained in package ‘Vclx50’“. Searching the Delphi 5 help boards found no solution to the problem. Lots of people have had this problem, but no one posted a solution. As it turns out, all I needed to do was reinstall Indy 10.0.76. There appears to be a bug in the 10.1.5 compiled packages, at least for Delphi 5. Use 10.0.76 instead.
As it turns out, the authentication problem with the new 15a camera was simply that it requires a user name and password to grab images (which 11a did not). In my program, I just added a user name and password to the “user” and “password” properties of the Indy HTTP component. Works great now!
More on the new Toshiba IKWB 15a camera soon …
Creating a home security camera system, part 3 February 10, 2006Posted by cnchapman in Technology.
Getting the picture. In this part, I talk about how to get the images from the Toshiba camera so they can be used by other programs for archiving and motion detection. This was trickier than I anticipated.
As mentioned earlier, I selected the Toshiba IK-WB11A camera for the visual part of the network. The camera worked well right out of the box. There are various options that can be set on it and are self-explanatory.
My first requirement for this system was to store pictures continually, regardless of whether motion was detected in the images. That way, regardless of whether the motion detection algorithm works or failes, I will always have a real time backup (backed up in two places). To do that, I needed to access the images in real time.
This posed a problem: how to access the Toshiba’s photos without using a web browser or their software. The documentation does not describe how to capture the current image as a simple JPG file. Eventually I found the answer from another user online: access the current image as :http://your.ip.address/__live.jpg?&&&
With that working, I wrote a Windows program using Delphi that would retrieve the images every second (or at any specified interval) and save them to the network locations. This uses the Delphi Indy socket components for the http download, http://www.indyproject.org/. I won’t post my code, as there are good examples online already, such as http://www.swissdelphicenter.ch/torry/showcode.php?id=2391.
Unfortunately, if you want to archive the images indefinitely, you will need to invest in a program to read the image files and save them. There are a lot of ways to solve that, ranging from writing your own (as I did) to buying one online. The Toshiba camera offers the onboard ability to detect motion and send email. However, I did not want to bog it down sending mail and the like — I wanted a fast, responsive camera, not a camera trying to be a server.
Now the second problem arose: I wanted to do motion detection on the live images. If there was motion detected from one frame to the next, I wanted the frame with motion to be stored on an external machine (not in my house) in real time and also backed up to various local locations. I set up “motion” on my linux server to monitor the images. It likes to look at a live stream, not at sequentially saved files with different names. So I pointed it to the camera URL as above — but it didn’t work. It seemed that it could not authenticate successfully against the camera, and thus could not get the image.
I solved this with a two-step approach. First, I modified my file-saving program to save the latest image for a camera to a specific, unchanging file name (in addition to the incrementing file names for the archived images). Second, I installed the Abyss web server http://www.aprelium.com/ (free for personal use) and configured it to serve up the current camera image files on my network. Now Motion could look at the file for a given camera and run motion detection perfectly.
Here’s a screenshot from my application, showing how it archives the images to multiple places and to the web server location. (Note that the picture quality is low because the hard drive box is very close to the lens.)
This also solved one of my other requirements: to be able to monitor cameras. I can simply point a web browser to the appropriate file on the web site. Of course that only runs locally, since I don’t want to serve up images on the Internet. (That would be easy to configure with Abyss, just not necessary for me since any with motion detected are stored externally anyway).
Creating a home security camera system, part 2 January 31, 2006Posted by cnchapman in Technology.
add a comment
To monitor my home, I had several requirements:
- Premises to be monitored both inside and outside
- Detect motion on the cameras, and also to record continuously (as a backup in case motion detection failed)
- Images to be high resolution, better than standard webcam resolution
- Images to be archived both onsite and offsite
- To be alerted automatically when motion detection occurred, and to archive those images specially
- To do this at a relatively low cost, using as many standard, reusable, and modular components as possible
- To have high reliability and uptime of the system
- To be secure
- To learn something while doing this
Probably the easy way to put all of this together would be to buy a commercial IP camera monitoring suite. However, there were several prohibitions there, primarily the cost (just the recording software alone is usually $200 and up) and questions about the reusability of the components should I rework the system in the future.
Ultimately, this is what I came up with:
- Premises to be monitored both inside and outside: selected the Toshiba IK-WB11A, which works both indoors and outdoors. http://www.toshiba.com/taisisd/security/products/prod_camera_detail_ikwb11a.jsp
- Detect motion on the cameras, and also to record continuously (as a backup in case motion detection failed): I configured the Linux “motion” software, which performs motion detection (see a future post for details on the complex software setup) http://www.lavrsen.dk/twiki/bin/view/Motion/WebHome
- Images to be high resolution, better than standard webcam resolution: the Toshiba camera has resolution up to 1280×960, which produces a very high quality image.
- Images to be archived both onsite and offsite: see my future post on archive and software setup. I save images to multiple local locations and have live email of images offsite whenever motion is detected. I wrote my own monitoring program (details and maybe even the EXE to be posted in the future).
- To be alerted automatically when motion detection occurred, and to archive those images specially: same as above.
- To do this at a relatively low cost, using as many standard, reusable, and modular components as possible: I used standard IP network cameras (Toshiba); wrote my own Windows application to record images; store images on Windows and Linux boxes, and a NAS device; email images to a gmail account; and use free motion detection software. Any of those pieces could be easily replaced with little need to reconfigure the others.
- To have high reliability and uptime of the system: using network cameras (instead of PC attached) ensures higher reliability of the units and no downtime from the PC. Battery backup keeps systems up. I wrote a Linux script to monitor and restart the motion detection as necessary (see upcoming posts on software)
- To be secure: my network is configured with double NAT and two firewalls, with stateful packet inspection and no open incoming ports. External storage is configured with strong passwords and is only known to me.
- To learn something while doing this — I definitely exceeded this goal!
What I did NOT do. I did not use standard, PC-attached web cameras. Their resolution is low, and I was concerned about their uptime reliability (because of USB drivers and the like). I did not use 802.11 wireless networking, because I found it to be too slow to serve high-resolution images at a moderate framerate (1-2 FPS). Instead, I used 100mbps CAT-5 (already wired through most of my house).
Creating a home security camera system, part 1 January 27, 2006Posted by cnchapman in Technology.
1 comment so far
After my house was burgled, I wanted to establish a home security system with 3 goals: (1) to discourage future burglary; (2) to continuously monitor my home; and (3) to store images offsite, so any future intruders could be caught. After a lot of work, I now have a system that works and meets all of those goals. However, it wasn’t easy, and I wanted to document some of the technology, tips, and tricks here.
This will take quite a while to document, so I’ll be posting pieces of the solution over time. Stay tuned …