Spaghetti Testing | Peter Smith

It just kept going and going and going and ...

At the end of last week, I made some tweaks to the #gc20 twitterbot I built — and today, the @gc_20 account started re-tweeting the same stuff over and over. Obviously stuck in a loop of some kind.

Of course, my little bot blew up while I was attending a training session — so I was nowhere near a PC for the whole day. At break this afternoon I made a valiant attempt to kill off the account via mobile, but that didn’t work — I kept getting errors partway through the confirmation process, so I couldn’t finish the job properly.

So it just kept on keeping on – as robots will when there’s nobody around to take care of them.

I’ve now managed (I think) to turn it off. About 15 hours after it started spewing garbage. That’s a long time in a 24-7, always on, realtime world.

Takeaway: when you are experimenting in Twitter, there’s nowhere to hide when things go wrong.

So if you are doing this kind of thing while representing your org, then it’s really important to keep close tabs on your creations — especially when you change how they work. And to make a contingency plan. Neither of which I did. ‘Course I created this thing without much forethought either. It was just a little experiment in getting the hang of building a hashtag bot. Which is another no-no if you are doing this more seriously — did I mention that it’s not a good idea to be going about this without a plan?

Anyhow, credit to my network — a couple of folks tipped me off (thanks !), otherwise @gc_20 might still be re-tweeting madly. Also credit to my peeps for not chewing me out for screwing it up. Then again, maybe they’re just practicing what mama taught them — “if you don’t have anything nice to say, then don’t say anything at all.”

Still, if I was an org, this could have been very damaging. Annoying the audience is never a good idea.

Image credit: Creativity+ on Flickr