Making WolfAlert Better

[Mike Vysocka]

While I was sitting in my office this afternoon, meeting with some colleagues, someone in the room received a WolfAlert message. Immediately, the others in the room looked for verification on their mobile device, while I spun around to my computer to check GMail and www.ncsu.edu for more information. Surprisingly, nothing. Panicked? Confused? Maybe. Was it a real alert? Was it an errant test? Why hadn't anyone else received one? Finally, a few minutes later I had both a text message and an email, success!

So then it occurred to me.

If one of the perceived inadequacies (and I'm sure there are others) of the WolfAlert system is the time it takes for the actual 'alert' to reach you (despite most users having multiple arteries of communication), how do we make it better? How do we make it faster?

Enter "Maps".

What if a campus administrator could click on a campus building (where an alert has originated) and immediately get a list of all of the students who are scheduled to be in class in that building (+/- 30minutes), all of the faculty scheduled to teach, all of the staff listed with offices, and do the same for the 5 or 6 closest buildings nearby automatically,... with one-click.

• Would this reduce the stress on the system?
• Would individuals closest to the WolfAlert be alerted sooner as a result?
• Would you reduce disruption and panic from spreading to otherwise unaffected areas? Granted, you could always send out a tier-2 alert to the remaining community members.

Other ideas for WolfAlert delivery...

• Automatically expire all WRAP cookies upon receiving an alert, put an issue on the WRAP page
• Do the same for Nomad?

What are your thoughts?

For the record. I received the WolfAlert at mail.google.com (in Firefox) at 12:51pm, the text message at 12:59 from Verizon Wireless.

[Jonathan Champ]

Thoughts:

 - Alerting people in descending risk order is an interesting idea, but is an attempt to optimize an inherently broken system. Also, alerting only "affected" individuals would mean that the people who are just there unscheduled would be left out in the cold.

 - Why can't WolfAlert do a broadcast message rather than sending the messages in serial (which is what causes the large delays)? This is a more difficult problem to solve, but e-mails can do this already, so it's mostly a matter of having the appropriate kind of interface with the external systems.

 - WRAP cookies can't auto expire unless the WRAP module on the servers do a check a WRAP command and control server (which I don't believe exists).

This is definitely an issue that would be good to speed up, but it looks like they were able to contact everyone I've spoken with within 10 minutes of the time in the e-mail (12:50 - 13:00), so they've improved quite a bit since the last time I timed it. My e-mail arrived at 12:57pm and my text message was at 12:56pm.

[Mike Vysocka]

Jonathan - we did think about those folks who may be in an affected building (maybe for an unscheduled advising session) that wouldn't appear in any of our data feeds, however ... looking at my own experience yesterday, we felt it was pretty likely that those individuals would receive the alert from the folks around them, or pick up on the fact that there was a lot of chatter/movement in the hallway outside their room without having to actually receive the alert themselves. Admittedly, this is a pretty bold assumption.

Glad to know you and others received the alert within a 10 minute window. Anyone else experience any unusual latency?

[Jason Austin]

It was slow.  I think I saw it on twitter first then got the text message a little bit later.  A few folks I was with never got the alert at all.  I did get the email before the text though, which in my opinion, seems counter intuitive.