I got a financial statement today, and on the envelope it asked me to go paperless, as its simpler and safer. However I like to keep my paperwork in a filing cabinet because i never can tell if the servers 15 years from now are going to have my documents that I may need.
That got me thinking about how historians are going to deal with this issue? A lot of the primary records are financial documents, letters and other papers that groups and civilizations keep. But 200 years from now, how are historians going to reference our records if they are all stored on a hard drive somewhere?
I can speak to this a little bit as it pertains slightly to my work, though I'm on the a/v side of things and not usually text documents. Keep in mind this is a relatively new and constantly developing area of archives management. I think archivists and not necessarily historians are better equipped to answer this question (though there is, of course, crossover).
What you're thinking of are usually referred to as "born digital" materials/records. Many companies like the one that likely sent you your statement have in-house digital archives and records management, though whether or not these digital records will survive the company let alone be accessible for research is a very good and very large question. Company (as well as personal) papers have often made their way to archives through strange and meandering paths if not directly donated, which will likely change when they're all digital. While paper seems lasting, remember that paper records have often been tossed, destroyed purposefully, or destroyed/damaged in disasters. We don't have most of the 1890 US Census because it was destroyed in a fire, for example. Digital material is just as fragile, and neither is perfect, so like paper conservators might with their materials, we take certain steps to preserve access.
Let's now talk about university, museum, and independent archives. Archives that have made the transition to working with born digital materials they receive have the task of perpetually keeping up with hardware and software before failures and obsolescence render their work meaningless. It's much like mold and mildew is to works on paper, now that I think about it.
In a/v (and I'm sure in text documents as well) this would mean keeping the file archived in the original format as well as recoding it to a newer, more universal format if that's what needs to be done to play it, as well as keeping a log of this information (which is the original, which is the recode and when it was made, etc). All file formats should be reassessed at regular intervals to make sure they're keeping up with the times. All storage drives obviously need a backup system that scans for bitloss and stores to multiple sites. Any special hardware that may be needed to play specific files should be inspected at intervals to ensure that it's still in working condition. I sometimes deal with artwork that has multiple components like odd or older hardware pieces (i.e. a 1983 video game console, or electronics created by the artist) - where replacement parts may be more difficult to procure in the future. In these cases, keeping a spare/spare parts around if you can get one is a good idea.
In all cases, policies and procedures should be revisited at regular intervals to ensure they account for both changing technologies and the archive's current and future capacity, allowing future historians (and others!) to continue their research unimpeded.
Sources for those wanting to learn more about born digital archives management:
Disclaimer to mods: This may mention somethings that have developed in the past decade and may not come with in the 20 year mark. It's a field that is picking up in the last decade and requires a little bit of contemporary mentions.
Digitization of records has been something historians and digital humanitarians have been working on for the better part of 10 or more years. Entire collections and archives are slowly becoming digitized in both small and large institutions. I myself have digitized some records and used records that are part of an online collection - this does include primary sources that are in paper. It is also part of a preservation effort as paper, vellum, cloth, etc. will deteriorate to a point where it cannot be read. Digitization also allows researchers and scholars to access these sources without actually having to pull the physical source out and possible damage it. The effort is more than just a nominal scanner but something of a higher quality so that the resolution is crisp even when zoomed in.
The Massachusetts Historical Society is one example of online archives. They have a huge collection of early American artifacts that are online that include letters from John Adams and Thomas Jefferson, paintings, historic weapons, etc. The fact that they are the oldest historical society and have a huge online collection shows that historians are not stuck in dusty cobb-webbed archives but understand the change in technology and its benefits. Another is the archives.org where thousands of books are scanned and allowed to be accessed for free. These are of course historical books or those without copyright restrictions. Many are primary sources as well and often books not easily found in a local or even university library. There certainly is promise within the digital field for historians to grow and develop. More than digitization but publications too. Many of the old journals are selling online subscriptions or copies of their volumes. It is cheaper and for those that may live in an area that is too expensive to get a mailed copy of East European Studies it allows access.
Tools like Omeka, Scalar, and even websites help with this. I've heard of historians worrying that this will cause destruction to the field back in the 90s and 00s but so far it's doing pretty well. The idea of going paperless is quite an old one too. Vannevar Bush talked about the 'memex' or memory extender back in 1945. Think of it as something like a microfilm projector and computer. The ideas was for information to be portable and easily distributed to other scholars. By means of a code, a memex user could access book, pictures, or other information and project it on a screen. Collections of materials could be linked in some way that resembles a prototype for google or even wikipedia. It maybe a longtime before other archives are digitized but there have been some valiant efforts in the past decade or so. We've come along way from microfilms with introductions of disks, drives, and clouds.