On Wed, Feb 18, 2009 at 4:07 PM, tedd <tedd.sperling@xxxxxxxxx> wrote: > At 3:54 PM -0500 2/18/09, Andrew Ballard wrote: >> >> You're missing my point. Yes, e-mail addresses are unique delivery >> points. They can not, however, uniquely identify one and only one >> person -- which is what one would need in the OP's situation. >> >> Andrew > > Andrew: > > No -- I did not miss you point, your point is obvious. > > I simply said that if it were me, this is what I would do. I also added that > my method ensures one vote per email address. I did not say that an email > address ensures one person. > > I am sure we both agree. > > Cheers, > > tedd > It all depends on the domain of the problem in which one is working. I agree that you could restrict it to one vote per e-mail address. Obviously, I can't speak for the OP. I've worked with applications where e-mail addresses were limited to a single domain and every user had one, and in those cases the e-mail address made an excellent key. I have also worked in situations where the correlation between people and e-mail addresses was n:m rather than 1:1 or even 1:m. In those cases, the e-mail address was totally unusable as any kind of key. Then there is a broader scope where one decides that, given the lack of a better solution, the overall population is broad enough to tolerate the imperfections since there is no better solution. To go back to what I said in my first reply on this thread, I consider that more about polling and statistics than voting. I'll be happy to let it go at that, though, since we all appear to be in agreement that there is no "magic" solution; only those that are "close enough for government work." :-) Andrew -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
