[Foundation-l] Long-term archiving of Wikimedia content

Samuel Klein meta.sj at gmail.com
Mon May 4 23:51:45 UTC 2009

I'm splitting off a separate thread about long-term archiving.  The
original thread is important enough not to derail it.

This is a big topic, and also one that has been addressed in many
different bodies of planning and literature.  The Long Now foundation
has considered a 10,000-year library project, and their Rosetta
Project tests a technique for 5,000-year preservation of texts.
Sadly, an earlier forum devoted to these ideas has been taken offline,
robots.txt'ed out of the internet archive, and I can't find a copy...
[ a long now apparently doesn't require archival public discussion? :)

Kevin Kelly on long-term backups:
The original y2k event:

Related research into long-term archival engineering has turned up
good ideas: laser micro-etching into nickel provides an excellent
price/size/weight point per archived page, and requires only the
[re]creation of decent, bootstrappable optics to recover lost

You could create and distribute etched-plate copies of the 10B words
of all Wikimedia text [and thumbnails?] on perhaps 100 thin nickel
sheets, for roughly $100k / 50kg / 0.01 m^3 (incl padding).  If this
laser etching process were scaled up, it would drop significantly in


On Mon, May 4, 2009 at 6:41 PM, Thomas Dalton <thomas.dalton at gmail.com> wrote:
> 2009/5/4 Nikola Smolenski <smolensk at eunet.yu>:
>> It seems to me that you are joking, but I was seriously thinking about
>> cooperating with the Long Now project on long term preservation of Wikipedia.
> No joke, I thing the long term preservation of knowledge is a very worthy cause.
>> Printing Wikipedia on acid-free paper every year or at least decade in several
>> copies dispersed on several continents should ensure that the contents last
>> for several centuries at least. It wouldn't be prohibitively expensive either
>> and it could gather some media attention (= sponsors).
> Acid-free paper won't last for several centuries without decent
> storage, and we're talking about a small library worth of paper. (See
> http://en.wikipedia.org/wiki/Wikipedia:Size_in_volumes - and that's
> just the English Wikipedia. Include other languages and other projects
> and you have a very sizeable amount of content.) That kind of storage
> isn't particularly cheap. Air tight containers in a cave might work
> pretty well though - caves have very stable temperature, and the air
> tight containers would control humidity - and the caves already exist
> so no need to spend money constructing somewhere.
>> For a really long term, a cooperation with some brickworks, where a brick
>> printer would be introductd in the brick producing process, so that Wikipedia
>> (and other important works) would be printed on every brick produced. We know
>> that Sumerian tablets have lasted for thousands of years, so these bricks
>> would surely last that long too.
>> And for even longer, do the same with bottle manufacturers.
> Yeah, bits and pieces would survive a long time, but you wouldn't get
> any significant portion of the projects saved that way. If you got it
> written on bricks that were being used to build a building you have
> good reason to believe will be around a long time, then it might work,
> but you would need a lot of bricks.
> According to the page I linked to above, the English Wikipedia has
> 7,484,527,350 characters. Let's assume an 8pt font (any smaller and it
> becomes difficult to write or read easily) on a standard brick (which
> Wikipedia tells me is, in the UK, 215mm by 65mm), that's about 18
> lines of text and maybe 17 words per line. That's about 300 words per
> brick (I'm assuming only one face will be written on). That works out
> at 25 million bricks. That's well over 1000 typical houses just for
> one copy of one project. Since the vast majority of these bricks
> aren't going to survive you are going to want massive redundancy. I
> don't think it is practical.
> Engraving on bottles isn't going to work - the bottles will
> (hopefully!) get recycled.

More information about the foundation-l mailing list