The reliability of dates on websites

[ADSGarson O'Toole]

Ben Zimmer's recent message about the verb "drone" contained the following:

[Begin excerpt]
http://burnpit.us/comment/9732#comment-9732

Michael Shouse, Burn Pit (comment), Oct. 21, 2011
Who is to say who is a terrorist? That is a very dangerous label, and
will be used against us, even as we drone people into martyrdom.
[End excerpt]

The date of "Oct. 21, 2011" seems to be based on the text located at
the website. Since it is possible to retroactively alter the dates
listed on a website I think that it is natural to wonder about the
reliability of such dates.

Have the editors who are compiling the OED, Merriam Webster
Dictionary, American Heritage Dictionary, and others developed
policies on which dates listed on websites provide acceptable
evidence? Perhaps the archive capture dates recorded by The Wayback
Machine of the Internet Archive should be acceptable. (In principle
even these dates can be modified retroactively.)

The sentence with the verb "drone": "That is a very dangerous label,
and will be used against us, even as we drone people into martyrdom"
is present in a snapshot within The Wayback Machine database that is
dated December 7, 2011:

http://web.archive.org/web/20111207213830/http://burnpit.us/comment/9732

Even with a date of "December 7, 2011" instead of "Oct. 21, 2011"
Ben's discovery is great. Maybe both dates should be given.

Ideally, an electronic historical archive should periodically copy
data to a storage format that is fixed and difficult to alter, e.g.,
data should be burned onto write-once archival quality CDs, DVDs or
Blu-Ray discs (if they exist). If archival storage of this type does
not exist then it should be developed.

Backup snapshots in write-once storage could be used to provide
assurance that data is not being retroactively altered.

Garson

------------------------------------------------------------
The American Dialect Society - http://www.americandialect.org