I can't believe how difficult it is to convince people in a *tech* company to automate things.

"oh yeah it's only a couple of times a year. How much work is it gonna take to automate it, it'll take us only a few minutes to do manually"

That's how you get stuck in the past.

@Lensflare so with a few times a year and only a few minutes to spare per time you need to finish that automation in less than 30 minutes before its a waste of time :)

My clients fall into this so hard.

"Automate this for me!"

"OK"

"But I need it to handle these 217 corner cases I have"

"That... complicates stuff and will take longer"

"Just do it"

Then they only use it once a year...

And they paid a week worth of time when it would take me 2 hours to do it manually...

And of course, next year is a new corner case the script can't handle anyway.

Automated decisioning is hellishly complex to where usually no one knows how its supposed to work on a good day. When it breaks, it breaks for good. Same for automated workarounds of a poor process.

@iceb certbot?

@SuspiciousBug No. to check some special conditions with signup. Funny thing is I'm pretty sure we have selenium for normal use cases. Just not the special rules we wanted to check in this case

There are some factors that make even no time saved argument invalid.

Consistent and correct execution for example if the impact is significant.

@CoreFusionX if there are 217 corner cases than yeah the process might be fucked or it actually really is with automating because now either the logic already lives as documentation and the time spent on handover is significant.

Of course other things can be much more valuable to automate. Giving loads of credit to the comic.