Re: "At long last, Sir, have you no shame?"
- To: mathgroup at smc.vnet.net
- Subject: [mg18541] Re: "At long last, Sir, have you no shame?"
- From: sidles at u.washington.edu (John A. Sidles)
- Date: Sat, 10 Jul 1999 02:18:42 -0400
- Organization: University of Washington, Seattle
- References: <7luk6s$l95@smc.vnet.net> <7m3l4p$si1@smc.vnet.net>
- Sender: owner-wri-mathgroup at wolfram.com
>David Withoff <withoff at wolfram.com> wrote: >>independent experts often disagree about what is or is not a bug. Savvy Mathematica users learn not to use the word "bug" when talking to Wolfram Technical Support! I suggest the phrase "unexpected behavior" instead. This avoids pointless discussions of what the documentation does or does not guarantee. As a concrete suggestion, how about a Wolfram web page listing, for *every* built-in Mathematica command, the "unexpected behaviors" that users have reported, together with workarounds. Special priority should be given to Gotchas that occur silently or sporadically --- these are by far the biggest threats to software integrity. In fact, it would make sense to restrict the list to *only* silently-failing "Gotchas". How many Mathematica users would consult such a page? I sure would. A "Gotcha" Page solves many problems for Wolfram management --- for example, it would not obligate Wolfram to release software patches, or even to acknowledge gotchas as bugs, because almost all gotchas are easy to work around, *if* you know about them! And it would be easy to organize --- just list the commands in alphabetical order, and then for each command, sort the reported gotchas in descending order of frequency reported. And it wouldn't be excessively long -- most Mathematica commands have no more than one or two silently-failing "Gotchas". And finally, need I say that users and low-level tech support personnel should be the ones to determine what goes on the Gotcha page, *not* Wolfram management? The job of managment should be to read and learn from the Gotcha Page, not censor it!