I've recently had to pick up a mail-merge for my colleague who has gone off sick. I've learnt an important rule: if you are ever asked to do a mail-merge in 1 day on data that you don't know, don't do it.
Doing a mail merge on data that you know well, that you know is complete, and correct is usually quite trivial. That's what I assumed this time, foolishly.
Then I got back 800 "Message Undeliverable" errors. And then I got another 20 emails asking "Why did I get duplicate emails? Which version should I use?".
If this had been 'my' data, I would have checked for duplicates, but up until now, I would have had no tools to actually validate the email address against.
But I finally got access to our LDAP server and have
been running queries (command line tool only,
but I will try to install Net::LDAP
when I get the chance) to get the *actual* email
address, rather than the ones on the list we
were provided with, which appear to have been
based on guesswork... This helped me work out
which 800 had failed...
Then it's just a case of going through the list manually trying to work out what the correct email address is. In fact, in about 75% of the cases, the name is completely wrong. The person has either left the company or changed their name...
I must check out the CPAN module Lingue::EN::MatchNames, but I don't
have time to code it for now. Hopefully next
week, because this kind of situation is just
going to crop up more and more often...
Since then, I've had daily complaints
from our internal customer that not all of their
users have received their personalised mail... so
lots of manual munging to do. Can't wait to get
the time to write a tool to automate this.
