On Thu, 28 Aug 2003, Mike A. Harris wrote:
This brings up an important point I'd like to make to people. ... available, those people will become upset and rant and rave.
Considering now that we have announced some dates in the beta ... missed, we are seeing people rant and rave about it.
If people want to see estimated dates for things, then they need to also realize that such estimates are JUST estimates, and that
Maybe this needs to be listed in 26 point font with bright red capital letters next to any dates, with a little * afterwards pointing to the fine print below.
Basically, wether or not any date is announced for a release from *ANY* software project, any software project will be truely only released "when it is ready" to be released.
Sadly, this isn't true. Have you ever used Netscape 6? The proxy dialog didn't even work right ;( I am very glad that redhat releases when it's ready instead of when they said it would be ready. It may be frustrating at times to wait a long time for software to be released, but I think it is far more disappointing to wait and then have something unusable released.
If no date has been ... also potentially may not be met.
Thus proves the rationality of the previous policy of not pre-announcing release dates. ... So dates truely do not provide people with any truely valueable information, since the information is not something that can be 100% relied upon.
I agree with you, but I also write software, and people do always want to know. I think they don't realize that programming work is typically not done very uniformly. Maybe they immagine I can write 100 lines of code per hour, and the project will be 4000 lines of code, and so take 40 hours, which is one work week. They might not realize that one day I might write 1500 lines of code, and that the next, I might write only 5 lines. They might not know how tracking down a stupid bug took 4 days, or that sometimes I actually plan before typing.
At any rate, you're not going to keep people from asking, and "when it's ready" sounds smart-ass, no matter how accurate it is.
I think the best solution is this: Put up a conspicuous estimated schedule on the web site where it will be very easy for everyone to find. Also include in big bold letters that it isn't accurate, and might not be met. Every few days, update it. It doesn't have to be remotely accurate, but if the project is going to miss the date, just move the date up a month or something. Then when people ask, you can just say "look at http://rhl.redhat.com/sched.html" and get back to work.
Of course, this is all just my own personal opinion, and does not reflect any views of Red Hat.
of course ;)
OS Systems Engineer - XFree86 maintainer - Red Hat
-- noah silva