Live data from Hacker News

Ask HN: How Would You Build a Foolproof Attendance System?

news.ycombinator.com

1–5 of 5 posts

Ask HN: How Would You Build a Foolproof Attendance System?

#1
I'm building a privacy-focused, secure attendance system that works solely with students' smartphones—no additional hardware. Here's my approach so far:

    Biometric login: Ensures only the phone owner can log in.
    Geo-fencing: Verifies the user is within the classroom perimeter.
    OTP/QR Code: Adds a layer of real-time verification.
I'd like advice on how to improve this, because there are still some flaws in such a system. I want to avoid buying any additional hardware, and just depend on smartphones, and make it safe and anti-cheat in nature. Please chip in your expertise and experience. Thanks!

Re: Ask HN: How Would You Build a Foolproof Attendance System?

#3
post #2

How do they access it? Wifi or internet? How do you verify that the data you receive is true? Any data you receive must be expected to be forged.

There is going to be a an application on the phone, which a student/personnel can't log out of. To log out you'd need authenticators existing in applications with the management, along with a log-out request that must be accepted first. As for taking attendance itself, at the end of each lecture, a professor would share an OTP, or a qr that changes every 5 seconds and is also valid for that period. Even before, one can scan the QR code, there would be a biometric and a geolocation check. Once all of this is complete, the attendance metadata would be sent to a server hosted in-house, and presence will be acknowledged. The problem is in someways, this system is still not foolproof. And this is the place where I need help.

Re: Ask HN: How Would You Build a Foolproof Attendance System?

#4
If you have a building access control system just add the students to that, they clock in and out of rooms (most access control systems don't just control locks, they are widely used to clock staff in/out of work). There are solutions that use updated and rotating keys to combat cloning when phones are used instead of RFID/NFC cards. QR codes can be sent as photos or even live streams if you have short validity so you can't guarantee proximity - but RFID/NFC will except someone is clocking other students in. That would be easily noticeable if the readers are within the view of teachers, staff and other students. On many phones RFID/NFC also works just fine without having to open up an app first.

Re: Ask HN: How Would You Build a Foolproof Attendance System?

#5
post #2

How do they access it? Wifi or internet? How do you verify that the data you receive is true? Any data you receive must be expected to be forged.

There is going to be a an application on the phone, which a student/personnel can't log out of. To log out you'd need authenticators existing in applications with the management, along with a log-out request that must be accepted first. As for taking attendance itself, at the end of each lecture, a professor would share an OTP, or a qr that changes every 5 seconds and is also valid for that period. Even before, one c…

Geolocation can be faked.

Not every phone has biometrics but even if it does, it provides little practical security value in this case. It's unlikely that someone will share their phone but in case they are inclined to do so, what will prevent them from enrolling a classmate as a guest user?

Posting/scanning QR/OTP codes and authenticating users is logistically inconvenient and time consuming for both teacher and students.

QR/OTP can be photographed and sent to classmates that are not present.

An easy and nearly foolproof circumvent is to simply setup an empty/spare phone that can be passed to a classmate to use for logging you in.

Last but not least, it may not be legal to *require* a battery draining app installed on student's phones.

Bottom line --- It's difficult if not almost impossible to implement strong security using only devices that others control.