Kayla Eucken’s Blog

On Monday, I started running my informal user study. The first few people found a few bugs with my study but nothing too huge. I got through 6 people and arranged for 3 more for Tuesday.

On Tuesday, I finished the last 3 people for my user study and entered all the data from the study. I worked on analyzing the data and graphing it.

On Wednesday, I worked on writing up my study and the study results for my final report. Kyle, Garret, Evan and I went out to lunch because it was Garret’s last day.

On Thursday, I finished up my paper and emailed it off to Magda and Evan to get their comments. In the afternoon, Kyle was using the DB lab for a user study so I worked down in the UbiComp lab. I also changed Scenic and Notifier to incorporate some of the changes I discovered I needed to make by doing the user study.

On Friday, I worked on polishing up my final paper and finishing up my web site.

On Monday, we worked on practicing our presentations. I helped Leilani figure out some of her technical explanations. We got a free catered lunch thanks to Yahoo! and then did our presentations. Mine went pretty well. The people listening liked my demo.

The rest of the day we did trace collection because the RFID group wants to share their data with other researchers because most everyone has to use made up data. We did three scenarios twice and it took a really long time.

On Tuesday, I updated my website to include the PowerPoint slides from my presentation and remedied a backlog of missing journal entries. I worked on the IRB forms but didn’t get much progress. Evan suggested that we could do some kind of study that he had heard of called something like a software engineering user study but we haven’t really been able to find much information on it. I also updated my paper by adding a bunch of screenshots and rearranging sections to remove redundancy.

On Wednesday, I tried working on my paper but didn’t make much progress so I decided to work on collecting the traces for the individual scenarios. I had to learn how to be my own shadow since I was the only on and had to both act out the script and keep track on my movements (shadow). I ended up collecting eight successful traces. Apparently the traces will be shared with the scientific community so that they can actually use real data instead of fake.

On Thursday, I finished the rest of the individual scenario traces for persons C and D. It took a long time and my left arm is sore from supporting the tablet, but it’s done now. Sometime hopefully we will get enough people together to do the rest of the multiple person traces.

On Monday, I incorporated Evan’s changes and got Scenic to compile again so that I can work on it. It still won’t quite take the grammar sequence and generate the icons from it but it is getting very close. I also started working on my slides for my presentation next Monday. It turns out that Google Documents is very good for making slides. Trace collection got postponed again so I worked on getting Scenic to pull the place and people names from the database. I may have to change which table the people names are pulled from, but it is working.

On Tuesday, I worked on the property list population in Scenic. I changed the places to be populated from a table in the ecosystem DB, and added methods to load the list of people group names. Trace collection got postponed again. Kyle forwarded the Internal Review Board forms that I will need to fill out and submit to get permission to do a user study.

On Wednesday, I worked on the CSS for Notifier and added a refresh button to the table of currently subscribed events. I also got Scenic to partly load templates and events. It had issues loading the pictures for people and things and for some reason it would display the ID as the name. E.g. it would reload my icon as ‘keucken’ instead of ‘Kayla Eucken.’ I also added some screenshots from Scenic and Notifier to my paper.

We also did trace collection in the afternoon. It was really hard to get everyone ready, and have all the keys and cards needed, and have all the antennas working… We ended up getting the big group scenario done so that means we will only need two people and two shadows walking behind them in order to do the other scenarios (instead of four people and four shadows).

On Thursday, Kyle and I brainstormed up some questions to answer with our user studies. I worked on the internal review board forms. I also worked on getting Notifier’s text to speech to save a file of the speech and then load it and play it in an iframe. Using GIMP, I also made a nice loading gif.

On Friday, we had our weekly meeting and I showed Magda how Scenic and Notifier were looking. I changed Scenic to have an all purple color scheme and took a bunch of screen shots for my presentation slides. Kyle, Leilani, and I went out to lunch on the Ave since it was our last chance to have a group lunch. Kyle will be gone next week until Friday afternoon and Leilani’s last day is Friday.

On Saturday, I came in at 1 and worked on my presentation a little bit. Evan still hadn’t had time to finish the loading stuff so I didn’t get much done. I did, however, commit a standalone email class with methods to send email and send email with a specific sent date. Kyle had been needing to send emails in the past in order to set up some fake ones for use in her user study.

On Sunday, I came in at 4 and worked until 6:30. I added a little bit more to my presentation slides but I mostly just worked on getting Scenic to actually load template and stuff correctly. Somewhere it was assigning the actor number (-1, usually) to the actor ID, and assigning the actor ID as the actor name. It took me a long time to figure out which of 94+ classes had the erroneous statements. I ended up having to change some classes from listing the actor ID as an int to a String, but I got it working.

On Monday, I worked on my Project Milestone paper, mainly incorporating Magda’s comments from over the weekend. Toward the end I actually got my javamail system to work. It was actually scary how little authorization information was needed to send an email. Evan read through my paper and identified some things for me to change.

On Tuesday, I got the email notification classes working so that given an event ID number and a probability, if the probability is above the user-defined thresh hold then an email will be sent to the email they listed with the message that they customized. I also worked on getting Notifier to load the notification details when a notification is being edited. Evan forwarded a script that Julie had created for creating and recording interesting traces.

On Wednesday, I worked on my milestone some more in the morning. We were supposed to be collecting traces that day so Evan showed us around to where out assigned “offices” were. I helped Evan and Garret make tags (person, mug, keys, notebook) for four people. We were supposed to start collecting traces at 2 but we got started a bit late and then found that one of the antennas was down so we did some practice runs and eventually postponed the actual data collection until Friday.

On Thursday, I saved a final (I think) copy of my milestone and uploaded it to my site. I transferred all the files from the milestone project to the final paper project so I can work on that now. Garret and Evan pointed out that you can actually send text messages to phones by email so I figured out how to do that and confirmed that it works at least for AT&T, Sprint, and Verizon. Now my program can send mobile notifications!

On Friday, I added the drop down list of phone providers (who support email to SMS) to the Notifier GUI so that new users will select their provider. We had our weekly meeting with Magda and she suggested a few things concerning getting the program ready to demo on the 18th. As a result, I added a Test the Notification button so that Notifier calls NotificationModule and sends out an email or text as if the event had been detected. I also go the text to speech to save to a file and started working on how to get the program to call people and play a recording.

On Monday, I changed Scenic so that it would make use of the toXML methods that Evan had created over the weekend. Scenic will now save the XML and PeexL versions of the events and the XML for the templates. It still won’t load the XML though because Evan hasn’t had a chance to write those methods.

I also worked on getting the Java text-to-speech to output into a file. Supposedly it is possible, but the documentation that goes with the API is really not very thorough.

On Tuesday, I continued working on my project milestone paper. I also added the ability to view templates created in Scenic and to edit them, similar to how the events are. I went home an hour early, ate, and then came back and worked for an hour and then went shopping with Leilani and Kyle.

On Wednesday, I got the delete buttons for the events and templates to work and actually delete the correct events and templates. I also added some of the real templates that I thought up a while ago.

On Thursday, I copied over what I had from the project milstone into another project to differentiate it from the final paper.  For some reason even though I had the exact same files, the new project wouldn’t acknowledge the bibliography file so it wouldn’t put in the citations. I spent quite a while trying to get that to work.

On Friday, we had our weekly meeting with Magda and the group. I showed her my progress and she said it was good. She said to email her our DMP progress reports and our Project Milestone papers. I spent the rest of the day working on my milestone.

On Saturday, I came in for two hours to finish up my milestone and emailed it to Magda.

The FTC workshop was held in the William Gates law building from 8:30 until 4:45. I missed the first speaker of the day but caught pretty much everyone else. The following are my notes and observations from some of the speakers.

Panel: Consumer Understanding and Acceptance of Contactless Payment Technology

Jodi Golinsky, MasterCard

• PayPass, MasterCard’s contactless cards are processed the same way as normal credit or debit cards.
• PayPass cards have unique, dynamic transaction numbers so even if the data is captured it cannot be reused.

Jean Ann Fox, Coonsumer Federation of America

• If you don’t have a lot of money, if you can just tap your card you may spend more.
• Contactless cards, meant to replace cash, allow tracking of consumer/market behavior that cash doesn’t.
• Contactless cards may not be covered by the same federal laws like the Truth in Lending Act.

Mark MacCarthy, Visa

• The cost of the fraud is not put on the back of the cardholder. It falls to the bank or credit card company so they have good incentive to make it safer.
• Visa also has dynamic, transaction specific numbers and they are adding another way to make the number more unpredictable. — MC already has it.

If you get the card number and expiration date you can use them online unless they require entry of the number on the back of the card which is not stored on the chip.

Old cards will be replaced with new technology by a certain date.

Panel:Contactless Payment Cards

Peter Ho, Wells Fargo Card Services

• They put a silver foil sticker on the cards when sent out in the mail to prevent it from being read until the sticker is removed.
• If you get the card’s information, you can’t really do anything with it. You can’t clone the card as a magnetic stripe card.

Dan Johnson, Tully’s Coffee Corporation

• People who have contactless cards actually use them.
• People without contactless cards may still try tapping them on the reader and be confused why it doesn’t work.

Kevin Fu, University of Massachusetts Amherst

• You can disable the chip in the microwave but you get sparks. You can also disable it with a hammer.
• When they say the chip is encrypted, it really isn’t.
• They have created readers where you can put a reader in a briefcase, walk up to someone, and it will read their card inside their wallet, inside their pocket and come up with their name, credit card number and expiration date.
• Even technically educated people may have no idea that their cards may have chips in them.

Tom McAndrew, Coalfire Systems

• As long as you still have backwards compatibility it lowers your security.

Etona Uedo, Nomura Research Institute, Ltd. (from Japan)

• Almost half of people in Japan have E-money, but only in big cities.
• Using E-money as money alone is very costly. Coupling it with some other service is better, such as for mass transportation ticketing.
• They don’t track your name and don’t really care what it is, but they may track your movement and behavior.

After lunch:

Panel: Mobile Payment Devices

Susan Grant, Consumer Federation of America

• Consumers may not be aware that the personal information on their phones can be compromised.
• Children may not carry credit cards but they do carry mobile devices, so if cellphones act as credit cards, there may be social changes needed.

Peter Wakim, Nokia Inc.

• NFC (Near Field Communication) is developed to be purposely close range so a tap is required.
• Still only in trials (since 2001)
• They have used PINs to increase security. You have to enter the PIN each time before it will function as a credit card.
• If you lose your phone it can be deactivated over the air where credit cards cannot be.

Siva Norendra, Tyfone USA

• They have investigated putting RFID info and antenna on memory cards instead of SIM card to allow you to keep the phone provider from being a middleman.
• It is important to allow consumer choice to have any phone, any service provider, and any bank.

Andras Vilmos, SafePay Systems Ltd. (from Hungary)

• They should allow customizable selection of mobile services.
• In Europe, mobile ticketing was what pushed the technology.
• They have a bunch of trials throguht Europe but they are not connected and still not ready to be commercial.
• People who haven’t tried it really don’t care but people who have tried it tend to like it though they still want security features like PINs.
• The readers for contactless payment cannot tell the difference between a card and a phone.

There is already antivirus software for cellphones and viruses, though there isn’t much point yet since little important information is stored on them so far.

Panel: Meeting the Challenges: Strategies and Approaches

Alissa Cooper, Center for Democracy and Technology

• Even though the bankc and merchants who came today are “best actors,” not everyone tries to do the best for the customer.

David Moorman, PCMS Group

• “Something that is readable from two inches today will be readable from 2 miles in the future”
• Google: “PCI and the Circle of Blame”
• One-off transactions add up if undetected - Hanford breach was 4.5 million one-offs

Kathryn Ratte, Federal Trade Commission

• The FTC does not govern banks, but they do regulate most other commerce.

John Carlson, BITS/Financial Services Roundtable

• You can’t be so afraid of how a technology could be abused that you don’t look into it at all.

Other observations that I took away from the workshop:

• Both Leilani and I, and my landlord have WaMu debit cards which apparently have the PayPass chips in them. We had no idea. My landlord had never heard of the tap and go technology. The documentation that came with my card only mentioned that it had PayPass and suggested stores it could be used in. It made no mention of what technology was behind it or any possible dangers. I was a bit unnerved to find out that I was actually carrying around a little RFID chip that may have my name, credit card number, and expiration date and no one bothered to tell me that. I’m interested to know whether my name is masked on the chip.
• You really can’t stop people from reading the cards. You just have to make it so that the information they would get is not usable.
• As long as there are online merchants who do not require the added security features, the backwards compatibility means that stolen credit card numbers can be used there.

On Monday, I made the tables of events in Notifier able to be refreshed. I also started working on getting the properties lists in Scenic be populated dynamically instead of being hardcoded. I had to get Garret to help me because the dropdown lists would be rendered before the data got back from the DB. I ended up writing a listener interface and having a method fire a ThingListChanged event when it was done populating from the DB. That approach involves quite a bit of code but it appears to work reliably, which is the important thing.

On Tuesday, I realized that there was no purpose in having Recurring or Active properties associated with events in Scenic because within Scenic it changes nothing whether something is active or recurring or not so I moved those fields and buttons over to Notifier.

On Wednesday, I got my program to save the event details for events to be notified by email or SMS. I also figured out how to cause cascading deletes in MS Server 2005. It actually only requires checking one box when assigning a foreign key constraint.

Evan was back and I showed him my progress.  Karl was there and suggested that I use WebAnywhere which is some program that is being developed here that helps blind people view web pages without needing to install anything, not even browser plugins.

We also got accounts for the WordPress blogging software and I spent a while getting that set up and transferring over my previously written journal entries.

On Thursday, I spent almost the whole day at an FTC (Federal Trade Commission) workshop called “Pay on the Go: Consumers & Contactless Payment.” I will be putting up my notes and observations from listening to the panels.

On Friday, I typed up my notes from the FTC conference. Magda wasn’t here for our weekly meeting so we talked to Evan instead. He had us dowload Latex/Miktex plugin for Eclipse that is used to create formal writing papers. It allows multiple people to be working on separate parts (intro, abstract, etc) and the program will automatically put the parts together and format it all nicely. It also handles properly formatting the bibliography and citations, which is nice.

On Monday, Evan isn’t here this week to help me with Scenic so I designed the GUI for my Notifier application. I managed to implement most of the GUI although it is mostly unformatted. I got to reuse my selectable-row table class so that was a lot faster. For lunch, Leilani and I had leftover sandwiches from a convention on the weekend that were on a tray in the refrigerator with a sign that said “Eat Me.”

On Tuesday, I worked on the listeners for the buttons in my Notifier application and made it load the user’s events and subscribed events from the database.

On Wednesday, I continued working on my Notifier GUI functionality. We went to a pizza lunch for undergrad and highschool students who are working at UW for the summer. I started trying to send an email from my program but I needed a mail server so I set that aside and started trying to get java text-to-speech to work.

On Thursday, I spent most of the morning trying to figure out compiler errors I was getting where eclipse would recongize the classes that I was importing but then when I tried compiling it wouldn’t be able to resolve the classes. I eventually started IMing with Garret and figured out that I had to route my call to the class through the servlet because otherwise it would choke when it found that the class wasn’t GWT compatible. At any rate, that meant that I got my text-to-speech program to work. I also worked on making my Scenic GUI prettier and created a Scenic logo.

On Friday, we met with Magda for our weekly meeting. She liked how I had made Scenic look and said that it was good that I had moved to working on the Notifier program instead of being stuck without Evan.

On Monday, I worked for an hour and then went on a hiking trip with the database group. We hiked up to Rattlesnake Ridge. I was the slowest one going up, even slower than the six(?)-year-old with blisters. The view was awesome though and I took a bunch of pictures. The way down was much better but I was tired so I skipped the dinner that everyone else went to at Cafe Flora.

On Tuesday, I managed to get most of the extended Scenic GUI implemented, although I left it unformatted and most of the buttons didn’t do anything. I got the functionality of moving between pages to work though.

On Wednesday, I spent most of the day formatting the GUI and making more of the buttons work. I also got a drop-down list to be populated from the database, which was a major achievement as it meant that my DB connection classes work. I also started making a table of user events from the DB. I got the data to display easily enough but I wanted the ability to select a row in the table so that a particular event can be altered. Making a table row selectable using the Google Web Toolkit ended up being quite difficult to figure out.

On Thursday, I continued formatting my GUI and actually got the table rows to be selectable and to turn colors when you click them and only one row to be selected at a time.

There was a speaker named Uwe Roehm from the University of Sydney. He discussed data management for genome sequencing projects where they generate terabytes of data every week that even the scientists don’t know how to evaluate entirely. He emphasized how they were using the new MS SQL Server 2008 to help.

On Friday, we had our weekly meeting and I showed Magda my GUI. For lunch, we took a group lunch to University Teriaki. Evan, Sam, Andrew, Kyle, Leilani, and I went.

Unrelated to work: Friday night my boyfriend Jason was visiting. We crawled out my window and watched the sunset from the 3rd floor roof. He proposed! My engagement ring is a One Ring replica.

On Monday I continued working on my project (programming) and started trying to get it to use rows with specific columns instead of just a label for each row. There was already a Row class from the RFID Browser they had done previously so I copied that code and modified it. It was somewhat difficult to understand since they had enabled sorting in their table and I won’t need that in mine and that made their class a lot more difficult. I eventually figured out which GWT tables I actually needed to extend and got it to display a row of data.

On Tuesday, Magda emailed back my proposal (at midnight) with comments. I looked at the Cascadia paper like she suggested and it occurred to me that most of the event detection details I had been thinking of doing had either already been done or should be done by someone else who knew how to build it into the PEEX event detection system.

I met with Magda and changed the focus of my project to specifying and saving event templates in Scenic. Scenic is a graphical, drag-and-drop way of specifying an event such as “Bill With Coffee Mug Inside Lounge”. Currently you have to start from scratch but we would like to have it work so that you can load an event such as “[Generic person] With [Generic Item] Inside [Generic Place]” and then specify which person, item, and place.

Evan gave us a quick database tutorial (I already knew SQL though) and I made myself some tables in a test database they had made for us. Then I tried getting my program to have the right classes and details so that the user’s events table could be populated by from the database but somehow on the laptop I am using I didn’t have authorization to connect to the server with the database. Evan emailed the tech people and apparently they are working on it.

On Wednesday since I couldn’t get my program to connect to the database I spent most of the day rewriting/proof-reading my proposal and creating this website as the DMP people requested. GIMP is awesome for making buttons, but I seriously should get some other html editor besides Notepad.

On Thursday we had a meeting and discussed how our projects are going. Over lunch we listened to a talk by some guy named Cristoph Koch who was visiting from Cornell University. He was talking about probabalistic databases and I understood a little of it although most of it went right over my head.

I designed my GUI for extending Scenic and started modifying a copy of Scenic to give it some added GUI features.