I sometimes suspect that I am actually an AI. I’ve always struggled with captchas and I comment my code exactly as you’ve just described.

You don’t for the one time codes because there is a standard that is supported by many authenticator apps.

App based 2FA is better. Either the app generates a time based code that you enter into the site or the site sends a push notification to the app asking you to verify the login attempt.

Passkeys are good too as they replace the password completely and leave the 2FA part to the device.

It’s better than nothing and some people would really struggle to do other types of 2FA.

I find this very odd because Azure DevOps is hardly ever down in our experience.

Microsofts documentation is also increasingly just outright _wrong_:

There used to be a spot on joke about Microsoft documentation taking the piss out of the fact that it was always 100% accurate but at the same time pretty useless. That joke hasn’t been relevant for a while.

It’s so frustrating trying to find out how to do something in one of the admin centres for M365 and you find a Microsoft document with exactly what you need in it only to find out that the UI has changed and the steps don’t work now. Did they move it? Did they remove it? Who knows?

The AI Fix podcast had a piece about how someone let an AI agent do the coding for them but had a disaster because he gave it access to the production database.

Very funny.

https://theaifix.show/61-replit-panics-deletes-1m-project-ai-gets-gold-at-math-olympiad/

I’m pretty sure that 5 is a feature because the button that moved is usually replaced with a clickable ad.

Sounds like SAP.

I think that there is always an implied design requirement of the program shouldn’t crash.

There’s a setting on the phone app to block notifications when you’re active on the desktop. It works well for me.

I don’t think that I get meeting notifications on my teams on desktop either so they just come up in outlook.

There are people in my org that always start with a “Hi” and then send each sentence as a new message. Very irritating.

The idea is that you can have more data online than you can fit on your computer.

It makes sense for SharePoint when there can easily be enough data to cause space problems on employee computers.

It doesn’t really make sense for it to be the default for personal OneDrives though.

I think that the main difference is that developers tend to test for success (i.e. does it work as defined) and that testers should also test that it doesn’t fail when a user gets hold of it.

I have worked with some excellent testers but I have also worked with a team that literally required us to write down the tests for them.

To be fair, that wasn’t their fault because they weren’t testers. They were finance people that had been seconded to testing because we didn’t have a real test team.

The current team is somewhere in between.

I wish our test team was like that. Ours would respond with something like “How would I test this?”

This is very frustrating! I get so many requests from customers asking why we returned response code 400 when we gave a description of the problem in the response body.

I remember when our company split up and we had to give them the source code of some older versions that they still used. We couldn’t do that because the repo was corrupt meaning that we couldn’t access some older revisions. We had no problems using it day to day so nobody noticed which meant that all backups were also corrupted.

It’s not the mechanism of branching that I prefer.

It’s the fact that Mercurial tags the commit with the name of the branch that it was committed to which makes it much easier to determine whether a commit is included in your current branch or not.

Also, Mercurial has a powerful revision search feature built in which I love (https://www.mercurial-scm.org/doc/hg.1.html#revisions).

It’s definitely up with Git in my opinion. I much prefer the branching in Mercurial.

It’s certainly very offensive to lump it in the same band as SVN and TFVC.