I track my time, too, and my biggest complaint is how long it takes. I think it's a great idea for companies to show why time tracking matters so it doesn't feel so futile.
My current company it makes sense to try my best to the hour/half-hour, because we're consultants and that's how we bill our clients.
A past employer for me was the worst of both worlds because they had a consultant's mindset towards billable hours having been born of a law firm, but not the consultant's approach to billing software development work. We were classified purely as "overhead" and not allowed to "bill" our (internal) clients. Meanwhile, despite starting life as a law firm it was by that point mostly a call center, and call staff to lofty law firm partners look like babies that need a strict panopticon.
Where things fall apart for me is logging overhead time, like time spent waking up with the first cup of coffee while reading company emails or doing IT-ish tasks on my dev system. I have no idea how to account for those times that doesn't make a mockery of the "Other" time entry.
In fact, I had to log X keystrokes and Y LOC every hour in order to be paid for that hour...talk about being a factory worker.
One tracked time to attempt to get r&d tax credits on certain work. The other was an agency that tracked time in case of disputes with unruly clients (such as the time we had gone through 3 project managers and 2 CFOs while we worked with them, and they were desperate to recover money from their own bad mistakes). The rest didn't track time at all.
Not to mention that LOC as a metric has been an industry joke at least since the early 80's: http://www.folklore.org/StoryView.py?story=Negative_2000_Lin... I've heard of funnier versions of that story, but haven't tracked them down yet.
Some of the most productive hours are LOC removed.
Counting keystrokes?
That sounds nightmarish, man.
Now that I manage my own team I let them abuse the Other time entry. I only expect 70-80% billable anyways, which lines up with studies on productivity. Anybody expecting 100% productivity is an idiot and anybody reporting 100% productivity is a liar.
Now I think about it - should have been a cron job, maybe that was why it was rejected.