The alternative viewpoint is that rolling the in-the-money call consists of 1) closing the call at a loss, and 2) selling a new call. On considering this further, I think those two actions should be independent.
The loss on the expiring call is irreversible. So why should the strike and expiration of the new call depend on how far underwater it is?

You *can* think of it as two separate trades. But you may or may not find that the most useful approach.
Yes, the first leg lost money. The second one might or might not have a gain, which may or may not be more than the first loss, so you might or might want to do the second one. But if your investment thesis hasn't changed and the first one lost money, then the second one is an even better deal that the first one, so it's a continuation of the same single initial investment decision. You're just waiting for it to work out, which hasn't happened yet.

If you buy a stock thinking it will soon go up in price by 30%, the hold period to hit that return goal might be slow or quick. Since options expire, and and investment thesis can take time to work out, I personally find it more useful to think of it as a single position that has certain events during its lifetime: an investment thesis, a position opening, [possibly maintenance bookkeeping stuff in the middle], and a position closing. For me, rolling an option is just maintenance work, like renewing a magazine subscription each year.

A good parallel example might be deciding to hedge your portfolio for a year with short index futures. You would normally do this three months at a time, as liquidity is stupendous for the near quarter contract and very much lacking after that. I don't see any particular utility to thinking of that, or doing your own bookkeeping, as four separate transactions with their own profit and loss---you hedged for the whole year, with one decision. If you were a CEO who decided to do this, you'd only have instructed the banker once. I think of options much the same way...it's the length of the position that matters, not the length of the (possibly multiple) instruments you used to implement it.

To me, it's just a long term position. I have some Berkshire long calls that I've been rolling for a decade or more. I don't even calculate how much I've gained or lost on the individual legs. It's just one method of having the exposure to the stock.

But, as you say, it's certainly true they are technically separate contract positions. If nothing else, your tax man or stock broker might care about that distinction.

Jim