Top College News Subscribe to the Newsletter

Class-finder text service blocked by UCF server

Creator summoned for hearing

Contributing Writer

Published: Sunday, June 17, 2012

Updated: Monday, June 18, 2012 15:06

Although the new prototype was only launched at the beginning of this month, U Could Finish has had more than 500 unique visitors on the site and currently 50 classes are being monitored.

Johnny Lee, a recent UCF alumnus, used U Could Finish before it was relaunched. Lee is perplexed as to why UCF would block U Could Finish from the myUCF servers.

“It seems that UCF isn’t too fond of student entrepreneurs making headway before they do,” Lee said. “UCF should reconsider the ban on U Could Finish.”

A Reddit comment suggested that U Could Finish would overload myUCF, but Arnold said otherwise.

“This version of the script is highly optimized to reduce load on myUCF as much as possible,” Arnold posted on Reddit. “It only searches for the exact classes users are trying to look for, and only on the interval they ask for. If multiple users are looking for the same class, it only queries it once for all of them.”

Arnold doesn’t believe that UCF is handling the situation fairly and started a petition on Change.org Wednesday to bring back U Could Finish.

“The purpose of the petition is to give students a voice," Arnold said. “If they agree that UCF class registration is broken and want to stand up for something to be done about it, signing the petition will help us communicate to administration that this is a serious issue that we want fixed sooner, rather than later. While I acknowledge that U Could Finish was not the best solution to this problem, it was a start. And without a start, there can’t ever be a finish.”

However, Gilmartin’s statement did recognize a myUCF issue and said that UCF is working on making changes to the website.

“UCF is in the process of creating a function that would automatically enroll students in selected classes in which space becomes available,” Gilmartin said. "Discussions are in the early stages, and no timeframe for implementation has been set.”

Recommended: Articles that may interest you

5 comments

Scott Arciszewski
Thu Jul 5 2012 15:56
A more sensible approach would have been a browser addon (or iMacros script) that queries it at an interval of once every 60-or-so seconds to simulate normal user activity (click here, then here, then here) without loading any images or stylesheets, using regular expressions to fish out the course number and the availability. Then it sends a query to your server that says "{unique token} has a seat available" and cease checking.

Next, have the server as a backbone that queries the server for all outstanding class ID numbers every hour, over port 9050 (after installing Tor on the server).

This is better for 2 reasons:
1. Decentralized. Makes it harder to stop and/or disable a browser addon that students can download.
2. Reduced load. Each account would only be queried every hour or so if the student were to shut down their computer or get kicked offline.

UCF would have been able to reduce load even further if they cooperated with Mr. Arnold and gave him API access so he could do one quick batch search on his server.

But hey, what do I know? (Other than ethics and practical computer security :P )

Tim Arnold
Thu Jun 21 2012 08:53
Hey guys, Tim Arnold here and I can help answer your questions.

1) I definitely do not know the actual load on UCF's servers - only IT knows their own architecture. However, I can estimate what a properly optimized server *should* be able to handle, as well as the load I am placing compared to normal students searching on their own. With that said, the optimization comes in two ways - first, from not requesting images, stylesheets, etc that normal users do. All I am receiving from MyUCF is the raw text results. Second, rather than requesting every class, I only search for the specific class a student is looking for. And if more than one student wants the same class, I only do one query for both, rather than each individually searching MyUCF themselves.

2) I do not have API access - U Could Finish uses a "guest" access system from myUCF, so it does not go through any credentials whatsoever. We only store the absolute essential information for users, such as Name, Email, and the Phone to text to (no billing information or passwords - handled by Amazon and Facebook respectively).

Thank you all for your comments, and hopefully UCF will implement some sort of solution internally in the near future (and not punish me for trying to help).

Anonymous
Wed Jun 20 2012 23:31
Another thing people haven't considered is security. There's no way Arnold has API access or automation credentials. Therefore chances are his system is running with his own student account information. Giving an automated system, which could be fairly easily hacked mind you (full web accessible), opens up myUCF to potentially malicious traffic.

I'm all for student innovaiton, and honestly I don't blame Arnold. It's a cool thought, a cool concept, and, given what UCF has available, a pretty sensible implementation. But that doesn't change the fact it poses dangers/performance issues to UCF and simply isn't sustainable in the long run.

Anonymous
Wed Jun 20 2012 16:37
This service sounds brilliant; I wish I would have known about it sooner. I understand why UCF may need to block an outside entity from clogging up its servers, but I hope UCF sees the viability of this idea and institutes it itself somehow through MyUCF. It would certainly help many students.
Anonymous
Tue Jun 19 2012 11:59
I'd just like to point out that Arnold is far less qualified to estimate the load his application would place on myUCF servers than the people who actually manage the myUCF servers. He might sit around and talk about how it's "highly optimized to reduce load," but since he isn't sitting on the serverside monitoring the network traffic, he is in no position to make such claims.




log out