|
The Forgotten Art of Email Address Validation
By Steven Smith
(This article is courtesy of Joshua Finer, who asked that it be published here)
As webmasters and database administrators struggle to keep email address databases clean,
web site form validation becomes key. At this point, the state of email address validation
is poor. Feel free to prove it to yourself: go to your favorite site that requires
registration, and enter “dlkjfdklj@djflkjdlkj.com” as your email address. More often
than not, it will let you through no problem. Keep in mind that not only is it letting you
through, but it is also storing that email address in an internal database. These email
databases get littered with bad addresses, which companies usually handle in one of two ways:
- Ignore the problem and send membership emails anyway.
This results in many bounced emails that waste bandwidth both to (when sent) and
from (when returned) the Internet.
- Hire temps to painstakingly pick out individual bad email addresses. This is a
costly and time-consuming endeavor. Additionally, this is a reactive, not proactive,
solution.
What you should be thinking at this point is “Why not stop the bad email address before it
gets into your database?” As you might guess, there are different approaches to validating
email addresses on web site forms. Generally, there is an inverse relationship between ease
of implementation and quality of validation. However, with the recent popularity of developer
components, these types of solutions can typically be “snapped in” with little learning curve.
The following is an overview of the various levels of email address validation.
- Base Level Validation - This is the type of email validation that most current web sites
use. Code will generally be in-line and simply look for a “@” and a “.” in the email
address. This is extremely inadequate and needs to be upgraded. The reason that most
sites use this method is that it is easy to code. An example in VBScript would be:
StrAddress = Request.Form(“emailaddress”)
If InStr(1,StrAddress,”@”,vbTextCompare) = 0 OR InStr(1,StrAddress,”.”,vbTextcompare) = 0 Then
Response.Redirect("badaddress.html")
Else
<continue form processing>
End If
- Base Domain Level Validation – This email validation is a step up from Base Level
Validation, but still extremely inadequate. This level of validation only requires that
the domain name of the email address be a valid registered domain. Many domain names
are registered but cannot receive email. Additionally, the majority of registered domain
names are not currently active mail servers. This is usually implemented as code or a
component that does domain name registration lookups or possibly even standard DNS lookups.
For purposes of example, let's say you choose to use a DNS lookup component. After
installing on the server, you need to first make sure the DLL is referenced in
View-->References or Project-->References. Keep in mind that most DNS objects are
expecting a domain name and not an email address. The following code is a generic
example of how you might implement a DNS component for email validation:
set DNSobject = Server.CreateObject("CompanyABC.DNS")
emailaddress = Request.Form("emailaddress")
thedomainname =Mid(emailaddress,InStr(1,emailaddress,"@",vbTextCompare)+1,Len(emailaddress))
DNSobject.domain = thedomainname
DNSobject.server = "XXX.XXX.XXX.XXX"
goodorbad = DNSojbect.dolookup
If goodorbad = "" Then 'Keep in mind some DNS object may return something other than ""
Response.Redirect("badaddress.html")
Else
<continue form processing>
End If
- MX Domain Level Validation – This email validation is currently the highest practical
level. It is not 100% foolproof, but is a marked improvement over any lower form of
validation. This level guarantees that the domain name of the email address is not only
registered, but also that it is a live Internet host that can actually receive email.
This is generally considered as the base practical level of email validation. This level
of validation can be easily implemented via a developer component. Similar to #2, you must
first install the component package on the server. After referencing the DLL, you might
use code such as:
set MXValidate = Server.CreateObject("CompanyABC.Emailvalidate")
emailaddress = Request.Form("emailaddress")
MXValidate.serveraddresses = "XXX.XXX.XXX.XXX","XXX.XXX.XXX.XXX" '(backup DNS)
MXValidate.ValidateIt(emailaddress,resultcode)
If resultcode = False Then
Response.Redirect("badaddress.html")
Else
<continue form processing>
End If
- Absolute Validation – This is a theoretical level of email address validation.
In this level, email addresses could be validated with 100% accuracy. Currently, this
is nearly impossible to achieve. Internet servers run various operating systems, which
handle email in various ways. Additionally, due to security considerations, many operating
systems are configured in such a way that specific email account validation is impossible.
Although this is theoretically the best, it is functionally impossible to implement.
The closest way to implement this method is to send the user an email requiring that he
reply in order to become registered for the web site.
The bottom line is that any web site that has little or no email address validation is just
a victim of lazy developers. These days data is heaping around us at an alarming rate. We
must strive to keep it clean and scrubbed if we expect it to be a productive and useful asset.
Please visit Component Source, http://www.componentsource.com to see a selection of email
address validation components.
Written by Joshua F. Finer September 24, 2000
Joshua is currently a 2nd year MBA student at The Pennsylvania State University, concentrating
in Entrepreneurship. Additionally, Joshua is the President of Finer Technologies, a software
publisher. Finer Technologies recently released EmailScreen, an email validation developer
component that can be found at
http://www.email-screen.com. Joshua welcomes anyone
to write
him: Joshua@finertechnologies.com
|
|