RE: Email type

Notification (warning) and assistance is better than rejection – particularly if the correct valid may not yet be known or even possible to know.

 

For example, XSD rules seem to require that xsd:decimal fields be populated, which makes no sense for stratml:PerformanceIndicator(s) of the actual results type to be measured and reported at future dates. To overcome that problem in StratML Part 2, we had to create a NumberOfUnitsTypeType as a union of an EmptyStringType.  http://stratml.us/references/oxygen/PerformancePlanOrReport20160216.htm 

 

Owen

 

From: ebruchez@gmail.com [mailto:ebruchez@gmail.com] On Behalf Of Erik Bruchez
Sent: Friday, June 23, 2017 6:58 PM
To: Liam R. E. Quin <liam@w3.org>
Cc: Steven Pemberton <steven.pemberton@cwi.nl>; public-xformsusers@w3.org
Subject: Re: Email type

 

Validation schemes coming from specifications such as RFCs might be correct, but they are not very useful to end users.

 

It is true that some web sites/web apps reject email address which are definitely valid in practice (I remember the rejection of the "+" character in emails such as `erik+test@example.org <mailto:erik%2Btest@example.org> ` by some sites).

 

But, witness Steven's example: it is technically true that `steven@cwi` is correct, yet there are virtually zero users in practice for whom that kind of email address will not be an actual error.

 

So there is some tension here which is hard to solve: you don't want to block users, but you also want validation to catch likely errors.

 

A possible way around this could be to have two kinds of email validations:

 

1. per RFC, which would catch a wide net and accept `steven@cwi` and other rare-in-practice addresses

2. practical, which would reject `steven@cwi` and other funny cases

 

This would work better if XForms had a concept of "warning" validation (which our implementation supports): you could have a hard validation catching certainly incorrect addresses and soft/warning validation for the narrower validation. This would allow a determined user to enter a likely incorrect email address if she is sure that it is correct, after accepting the warning. A form author could choose how to deal with the various options of email validation.

 

-Erik

 

On Fri, Jun 23, 2017 at 12:51 PM, Liam R. E. Quin <liam@w3.org <mailto:liam@w3.org> > wrote:

On Fri, 2017-06-23 at 13:28 +0200, Steven Pemberton wrote:
> Our definition of email accepts the following as a valid email
> address:
>
>       steven@cwi
>
> Are we OK with that? I'd expect at least one "." after the @.

https://tools.ietf.org/html/rfc5321#section-2.3.11
has:
[[
A domain name (or often just a "domain") consists of one or more
   components, separated by dots if more than one appears.
]]

so steven@cwi should be allowed.

Liam


--
Liam R. E. Quin <liam@w3.org <mailto:liam@w3.org> >
The World Wide Web Consortium (W3C)

Web slave for www.fromoldbooks.org <http://www.fromoldbooks.org>