Oh, oh, oh OmegaMom, you have nooooo idea how bad it is on the inside of the development of an OS.

High level management/marketing will keep asking: Can you meet this date?

And they keep getting "no's" until they find the one manager who says, "Sure thing!"

Presto! Promotion for that manager! And dates are announced.

By the times the dates have actually slipped, that promoted manager has made sure he/she is in another totally unrelated job.

Oh, I believe, I believe!

Even on the smallest littlest tiniest project I was on--managed by a venture capitalist and his construction project manager!--we had the exact same problems. Three software consultants in a row were fired because they couldn't "meet the deadline"--all the time the VC and PM were busy adding new features. And they couldn't understand why we couldn't keep the deadline and add those features...hell, we couldn't even meet the deadline without those features, but we had stupidly agreed.

The only difference is we didn't have multiple levels of managers, each decreasing the projected finish time in order to increase his shininess with management...

That's where I've come up with my project estimation strategy: take the furthest date you think it's going to take, then inflate by at least 33%.

I'm beginning to think I need to increase that percentage.

Of course, there are drop-dead projects that do have a firm date (like the ones I'm working on now), because, well, if you don't meet those dates, everything is going to...Just. Stop. Working. But in those cases, usually the Powers That Be aren't adding features, they're busy removing features so that the project actually goes live on time.

Johnnys link isn't working, its just "downtothis". Otherwise, none of this makes any sense to this non-geek.

Thanks, Miss C! I'll just go tidy that up...

Oh man, this is amazing, 1,000 lines of code a year? The poor bastards.

When I used to write code, we had to report our numbers constantly. It sucked.
And dates where terrible... We had the same problem as Johnny desribed. Oh man.

I learned quickly to behave like Mr. Scott. Take the time I truly think it will take, multiply by 3 and always get it done just before then. People thought I was a miracle worker...

Now what about that broken leg?

Vista sounds just like a certain in-house app project I have the good fortune NOT to document. They too will not admit an August release won't work. You see, it has to be there for the year-end sales numbers. I'm sure that's the higher-ups logic with Vista.
So, release "on-time" with a half-assed system needing megapatches and peeving the consumer, or delay the release so you have a functional system to make the customer happy and miss your year-end sales goals. What would the stockholders think?