Re: [Evolution] Full name parsing

Well, FWIW: in America at least, if we use a comma, it always implies
that the last name comes first.  Just like the default "File As" field
(I wish that were configurable) shows Last Name, First Name.


On Wed, 2001-09-26 at 05:54, Pascal DeMilly wrote:
I agree with you. But what about a rule that will help the parser in
deciding what is what. We could use the comma as a separator when in add
mode. So for example:

John & Jane, Doe will parse correctly. The comma could be removed
automatically by Evolution once the parsing is done. Obviously this rule
should only apply to full name not company name.

Therefore the rule will be as follow. Start by the end and find the 1st
comma. If there is none, apply existing rule. If there is one, we have a
last name. Then continue to the next comma, If we have none apply
existing rule. If we have one, we have a middle name. The rest if the
first name.

Therefore I should have inputed "John & Jane,, Doe" to be parsed
correctly. Evolution will remove the commas, and print it as "John &
Jane Doe".

Could that be implemented ?


On Mon, 2001-09-24 at 22:08, Jon Trowbridge wrote:
On Mon, 2001-09-24 at 20:38, Pascal DeMilly wrote:
When parsing a full name, I think Evolution should have a rule that know
how to parse John & Jane Doe correctly. I propose that John & Jane
should be firstname and Doe should be a lastname. Right now, John is a
firstname, "&" is a middle name and Jane Doe is the lastname. 

What do you think ?

I agree in principle, but trying to construct parsing rules that work
really well in a lot of unusual cases like this is really hard.  For
example, I've seen e-mail addresses like "Joe Smith & Family" and "John
Doe & Jane Roe".  Doing the right thing with &-separated names in all
(or even most) cases is a hard and murky problem.  And this is just one
of an essentially infinite number of obscure cases.  And to make things
even more complicated, the '&' character might have totally different
semantics in other languages, causing problems for all of our
non-English-using friends.

Not that trying to transparently doing the right isn't an important and
worthy goal.  It is just much trickier than it usually looks.

In the short term, I'd recommend using the contact editor to 'manually'
parse unsual constructions like 'John & Jane Doe'.  As you probably
already know, clicking on the "Full Name" button when editing a contact
lets you specify how the name should be broken into first/middle/last


evolution maillist  -  evolution ximian com

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]