Author mark.dickinson
Recipients Aaron.Meurer, Nofar Schnider, benjamin.peterson, ebarry, georg.brandl, gvanrossum, haypo, lisroach, mark.dickinson, merwok, ncoghlan, pitrou, python-dev, rhettinger, sbaird, serhiy.storchaka, skrah, stutzbach, tim.peters, veky, waldir, zach.ware
Date 2016-08-12.09:06:07
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <1470992768.21.0.223329478975.issue12345@psf.upfronthosting.co.za>
In-reply-to
Content
BTW, if we're talking about useless constants in the math module, it's hard to get more useless than `math.e`: it's rare for `e` to turn up in formulas except in the form `e**<something>`, and the latter case is better catered for by the more accurate (and usually faster) expression `exp(<something>`. I'd even go so far as to call `math.e` *worse* than useless, since its presence leads people astray by encouraging them to write `math.e**x` instead of `exp(x)`.

To give just one example, on my machine, the result of `exp(500)` has an error of 0.42 ulps, while `math.e**500` gives an error of over 150 ulps.
History
Date User Action Args
2016-08-12 09:06:08mark.dickinsonsetrecipients: + mark.dickinson, gvanrossum, tim.peters, georg.brandl, rhettinger, ncoghlan, pitrou, haypo, benjamin.peterson, stutzbach, merwok, skrah, Aaron.Meurer, python-dev, sbaird, zach.ware, serhiy.storchaka, waldir, veky, ebarry, Nofar Schnider, lisroach
2016-08-12 09:06:08mark.dickinsonsetmessageid: <1470992768.21.0.223329478975.issue12345@psf.upfronthosting.co.za>
2016-08-12 09:06:08mark.dickinsonlinkissue12345 messages
2016-08-12 09:06:07mark.dickinsoncreate