Your name doesn't start with "A" or "B", by chance? ;-)

You say it in jest, but I know a group whose head's surname started with a B and she was a staunch defender of alphabetical order... until someone joined who started with an A.

It doesn't, no.

I'm currently supervising 2 PhD students whose last name is alphabetically later than mine. I explained my reasons for my preference to them; that's how far I'll go. With one, we've had multiple collaborations outside the institution. In those cases, alphabetical order was typically arrived upon with all authors. In other cases it depends - we've had quite equal contribution papers which ended up alphabetically. We're now working on one that he conceived and executed. I'm basically the co-writing sparring partner, where I strongly prefer him to be first author.

The other student feels it's important for his career to be first author, so he is. I certainly hope he's wrong (for my career so far it didn't matter), but I'm not taking that chance.

I'm 2nd author on a 2-author paper where my only contribution was answering a couple of questions on how to use some software, and then proofreading the couple of paragraphs that mentioned it. I didn't even read most of the paper, let alone know what it was really about. Surely there's a need to communicate the fact that my contribution is less than that of the real author?

I agree, but order isn't the way to do that, as it could mean you did 49% or 4%. A better system would be to just explain what each author did.

No doubt. Or not even have such minor contributors as authors in the first place. Perhaps you should have to have actually written part of the paper, like an actual author, or at least done some novel research of your own. Technicians and others who assisted with it, even did all of the grunt work could just be acknowledged if they want some fame. But I guess that authorship is a kind of payment-by-exposure instead of actual money and enough people are desperate enough to take it.

In the last few years, an increasing number of journals are adopting that kind of system.

Could you elaborate on your gripe with order, or how that would be abused?

In my field, 1st author did most of the work, typically the responsible grad student, then authors 2 to n-1 supporting characters, maybe contributed some data, some script, wrote a paragraph. Last author is PI.

Sometimes you’ll have a 2nd author who ended up contributing like the 1st as the paper progresses, that’s really the only ambiguous situation, but you can switch that around for a conference or follow-up.