Date: Fri, 17 Oct 1997 12:48:59 -0600 From: Warner Losh <imp@village.org> To: Nate Williams <nate@mt.sri.com> Cc: Bill Fenner <fenner@parc.xerox.com>, Garrett Wollman <wollman@khavrinen.lcs.mit.edu>, Donald Burr <dburr@poboxes.com>, ports@freebsd.org Subject: Re: 8 days until 2.2.5... Administrative notices. Message-ID: <199710171848.MAA03209@harmony.village.org> In-Reply-To: Your message of "Thu, 16 Oct 1997 22:02:06 MDT." <199710170402.WAA18900@rocky.mt.sri.com> References: <199710170402.WAA18900@rocky.mt.sri.com> <199710151603.KAA11877@rocky.mt.sri.com> <97Oct16.150242pdt.177487@crevenia.parc.xerox.com>
next in thread | previous in thread | raw e-mail | index | archive | help
In message <199710170402.WAA18900@rocky.mt.sri.com> Nate Williams writes: : > Ok, I just modified fetch to keep a moving average of throughput over : > the last 25 seconds : : Can't you have it print the average throughput since it began, and : therefore the time-remaining since it began based on it's throughput so : far, rather than averaging it 25 second chunks? That's what most : browsers do, and it works 'fairly' accurately (much more accurately than : the results you got it appears.) Actually, I found that an exponentially decaying average over the last 3-5 minutes gave the best "range" of ETA times when I was playing with oidl, a tool that watched files and told you, based on how big you told it it should be how long it would take to get there. I said range because I kept the standard deviation as well and used that to give a 2 sigma interval of ETA times, when the 2 sigma distance was > 1 minute, otherwise I just used the exponentially decaying average. This ad-hoc approach could be improved on only in one way, which is what netscape does with "stalled" connections. Warner
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199710171848.MAA03209>