I can say little more than to thoroughly second Mr. Clarke's observations.  I might also add that I completely pay homage to the library user who will take the time to point out to me the apparent oddity of an author whose heading shows an open birthdate of 1825-   .    
    First, I really don't have true users (as opposed to other library staff with too much time on their hands) calling me on the phone to discuss such issues every other day.  It does occur.   In our shared environment, such things need to be explained in the context of thousands of libraries and millions of records, as Mr. Clarke so aptly points out.
   Second, and I hope that this isn't too obvious, but if someone says, "Hey, Mr. Fields must be dead by now!"  .... well, yes, so why do we need to add a date to a heading where it's just obvious to all that the fellow is in fact dead?

    Please, I don't know about those requesting such a change, but I don't have the time to be honing these records to the ultimate perfection that someone out there might wish to see.  We already close dates for people in the "corporateness" (e.g., presidents, governors, etc.) after their term in office is complete.  Unless we're considering a change to a "make work" society, I would prefer to leave things as they are.

Sherman Clarke wrote:
[log in to unmask]">
The universe of shared cataloging and shared authority work is built on
the assumption that headings will be as stable as possible. The
longstanding policies to continue to use headings that are accurate but
incomplete are built on this assumption. Bibliographic and authority
records have long legs as they are used by libraries around the world,
at different times. Thus, we have open headings for people who are
famously dead, and we have headings without dates. I think this
conservatism in headings is generally reasonable. Unfortunately, the
users probably suffer more than catalogers do; reference librarians
have to try to justify the open date for Warhol or Lady Di, end users
don't get the date context when no dates are included in the heading.

Maintenance of records may be easier now than it was in the past but
even with automated assistance and global change, it is effort that
could be spent in cataloging new items. If we are going to have more
dynamic headings, perhaps we should think about adding more 400s with
$w for earlier forms of the heading; this could expedite vendor
authority processing. Of course, depending on the system, it would
result in unhelpful references, unless we could teach our systems to
interfile open date 1928- with date range 1928-1987.

MARC encodes our headings which are based in AACR. The new cataloging
rules are moving toward being element-based. When you think of a name
as an element, the life dates become an attribute of the name. Encoding
schemes other MARC have other ways of dealing with dates than building
them in headings. This is probably not relevant in our catalogs now but
has an effect on how LC/NAF headings are used in other systems.

I'd advocate continuing the current policy of keeping headings as they
are (with open dates or without dates at all) in most cases. I imagine
most of the irate customers and pacification efforts are done for
famous cases. Adding death dates to the Andy Warhols, Lady Dis, and
Pope John Pauls would be a reasonable cost for easing the ire and
frustration, but I continue to think that the shared cataloging
universe is better served by a general conservatism about changing
existing headings.

Sherman Clarke, NYU Libraries and Art NACO
[log in to unmask]