So far it seems like we are making progress on the first phase of our project. The bookstore scanner is a fully formed idea but some security and privacy concerns stand in our way. Our main goal is automation.
Eric Lorenz is the IUB Bookstore manager and he agreed with us that keeping the process easily transparent and quick is a must. He has in his possession two tables. The first table has a list of student's University ID's as the primary key. The other attribute of interest is a the list of classes each UID is registered in. The second table has class sections as the primary key and the corresponding books required for that section. So to visualize things easily:
Table 1: UID | CLASSES
Table 2: CLASS SECTION | REQUIRED TEXTS
The main problem is that the student ID cards do not contain the UID anywhere including the barcode and the magnetic strip. To combat this we're probably going to implement an opt-in system, either immediately at the kiosk itself or through oncourse/onestart. Once students register they are set for the life of their student ID card or until their fingers fall off...
Tomorrow Carlos and I are meeting with Steven Gooley, an associate at Campus Card Services to determine our options without violating any terms and agreements the card imposes.
13 comments:
Chris and Carlos' idea of using biometric security is a convenient and secure way to automate the print out process. Not only does it allow a quick scan of your finger to register, but it also can't be forgotten like passwords and ID cards.
My image of how a first time user would register proceeds like this: 1) They approach the computer and click register. They login to the CAS like anyone does into oncourse/webmail/onestart. They then register their fingerprint and voila.
The problem was getting the UID to match up with a student. Our card swipe didn't incorporate using the CAS login. The current situation uses the CAS login to retrieve the UID, which would be available to us but only used once to link a fingerprint to a UID.
CAS -> UID
then
CAS -> Fingerprint
thus
Fingerprint -> UID
I checked out the website Chris sent us in email.
There were two models I particularly liked.
One had a stand that protruded up about 5 inches. It allowed the user to wrap their hand around it, almost the handle of a hammer, and the user puts their thumb on the top of the base.
The other model was a mouse with a built in scanner where most newer mice have the back buttons. This way users could use the mouse to select registration and register without having to learn a new device.
I personally like the mouse scanner better but it was more expensive by around $30. The website also had some enterprise software that could be incorporated into our idea.
Today we met with Matt for the first time. We had to kind of start from the beginning because he was uninformed with the complete details of our progress. We had felt we changed direction with the Student ID card but apparently not. Matt expressed our discontent in cooperation from the university and offered to talk to a fellow named Brad, who sounds higher up than Steve Gooley and Brian Bartley. If we end up getting access/authority to pull the UID off the card that might save us some potential registration work. Users of our kiosk would be able to either register their fingerprint with our database after a CAS login/swipe of the card. Or they could skip the fingerprint registration and simply use their card for their first time use and every subsequent use of the kiosk.
I think our first goal while this card problem is discussed should be to establish a welcome page that would communicate with the CAS. I posted a couple links at the top of the page that I will familiarize myself with hopefully by the end of the weekend.
-Michael out...
Today is the 7th of February. On the 5th we figured out which modules will will probably need installed on our spock/sulu accounts. Once those are installed we can work on getting the CAS integrated into our website. The idea for having our website integrate CAS will allow users to register with our database their fingerprint if they feel like not using their Student ID card.
Also as some good news, Eric finally gave us access to the database yesterday. Last night I was running some queries on myself and everything seems to be in order. One thing we need to decide is how to interface with the database. It is MS Server 2003 and I am unsure about how to connect with CGIs. One possibility would be to migrate the data to our spock accounts, which would make our project a little easier. Ultimately the data needs to remain on the bookstore's end as they would update the database every semester with new information.
One other hurdle will be printing the information we get from the query directly to the printer without pulling up an extra dialog boxes.
The more I think about it the more I realize that this bookstore print out may be the total project. We could add more functionality to it, but I don't think the rewards card idea will be considered, nor the apple database...but we shall see.
We ordered a mini usb card reader the other day off eBay. It can read all three tracks and does come with software that allows windows machines to use it. Currently I am waiting to hear back from the seller. We also have given our survey to 40+ students to confirm some of our suspicions about the method of interaction with our proposed kiosk. What really remains is getting a working prototype that allows for authentication with CAS and connectivity to the database we will be querying. I am very unfamiliar with .Net but have installed IronPython on my home machine which allows you to use python with access to .Net's libraries. I was fooling around with Windows Presentation Forms(avalon) at a very elementary level.
After looking at the results from the survey it doesn't seem like the right direction to go towards biometrics. This being a college campus most people are informed or just naturally worried about their privacy and information(especially a fingerprint) being in possession by IU.
We also received the magnetics stripe reader and it reads all 3 tracks perfectly. I tested it out on my own ID and all the information we need was on it. We met with Dennis and he suggested using an index to make sure our database behaves intelligently and responsively.
We recently had issue again with going between python and Visual studios. I have tried for 3 days straight to have python recognize the magtek magnetic stripe reader without success. I have seen successful implementations on the internet but in order for them to work the following needed to be installed and configured correctly: cygwin, python, pywin32(python windows 32-api compatibility functions), libusb-win32, and pyUSB. I had them all installed successfully but couldn't get python to recognize the usb modules....It could be a permissions problem but I think going with visual studios will save us compatibility headaches but give us new language headaches instead.
I recently have been trying to get any remote connection to our MS SQL database. We originally thought python would work well for this but discovered we were wrong. Python has no problem connecting to other ODBC databases but for some reason it doesn't jive with MS SQL. Once again after extensive googleing there are successful implementations out there but they require the use of modules that are either outdated, now broken, or modules in tandem with other modules to work. We thought we could use PHP to connect instead. I will go in that direction.
After trying for about 2 hours, I cannot get PHP connections to our MSSQL databse to work. There is a PHP function called mssql_connect() which seems to work for other people. I even found sample code on PHP's development website but again, I am unable to connect to our database. I submitted a tech request so I will wait back to hear if it might be a PHP version problem or the administrators have disabled the function. I will update when I find out.
Ok so we officially made the switch to Visual Studios using .net and C#. We have a big hurdle to get the usb device to output information the way we would like. We have a design welcome page that has basic functionality. We have some hard work ahead of us, but it is not above us.
Ok so visual studios is not that big of hurdle anymore. Dennis helped point us in the right direction. We simply modified the demo program that came with the Magtek Magnetic Scanner to output the card data to a text file named "cardData.txt". Our visual studios program can now access the text file and parse it for the University ID which is sent to a php page in the URL, and then queries the MS SQL database for the relevant information. Our task now is to get a working prototype so that we can begin usability testing.
One problem that I was unaware of for a while was how notepad functions. Apparently when notepad accesses a text file it 'touches' the file twice. It first opens the file and then saves the file, even if no changes were made. One problem I was having was that when the cards wipe occurred the actions were executed twice. I avoided triggering the events twice by using a counter. Also what was important for me to note is that on the first swipe notepad only touched the file once, so that in order to get a printout the first time we needed to swipe the ID twice.
Our program now is essentially broken. Eric from the bookstore gave us access to the server throughout the year, but after the capstone fair he removed the permissions, breaking our connections to the MSSQL database that outputs the important information.
Post a Comment