I classify a backup as being either a contingency backup or an archive backup.
Contingency backups are short term backups people take to protect against computer failure. These backups represent a snapshot of computer documents at a particular instance in time and should be updated as frequently as the implemented backup technology allows. These backups protect users against deleting a file by mistake or taking a working document one notch back rather than one forward. These backups protect against external mishaps such as hardware failure, malware attacks and such like events. Online backups are best suited for this type of backup.
What I am calling an archive backup can also be termed a prosperity, point-in-time or indefinite backup. An archiving backup is the act of copying documents on a computer for the purpose of referring to them at some point in the future. Compared to contingency backups, when one takes this type of backup there is never the intent to overwrite the backup with a later copy. Certain online backup companies may offer archiving services as part of a standard package or provide the same for a fee. I have encountered situations in which forgotten contingency backups become potential archiving backups when these are discovered stashed in some remote corner of an organisation.
You can think about archive backups as equivalent to the manuscripts created hundreds and thousands of years ago. From these texts we get a glimpse of the affairs of the era, the people who ruled and the news of the day. Those who take archive backups are not thinking of such a distant future but create these snapshots for one of two reasons: legal obligations or to retain a snapshot of the organisation just in case one needs to go back at that exact point in time. For example, archive backups may be taken prior to a major system upgrade.
The frequency at which an organisation takes an archive backup can range from daily to once every blue moon depending on legal, computation, physical and financial considerations. A small SME may not afford the time, money and space necessary to make a daily archive backup and a simple “end of year” snapshot would do.
Progress and change are archived data’s main nemeses. Yet the same progress and change are technology’s main driving forces. This contradiction makes it more difficult to successfully read archived data the longer the time period between when the archive was made and when it is accessed.
Many organisations try to avoid having to delete files as much as possible. The let’s-keep-it-just-in-case syndrome is nowadays much easier to justify thanks to the dirt cheap price of hard disks. In many organisations all documents are stored on a central file server and with certain companies it is becoming standard practice to have the folder of a former employee moved as a subfolder under his boss’s directory when the person leaves the company.
There are a few tips that, if followed, increase the chances of successfully accessing archived data at some point in the future:
a. Export a copy of the files you want to archive to a format that is open. Companies come and go and with them they take their proprietary formats. Just because a company no longer exists does not mean that its file format algorithm is known. Even if the file format may have been deciphered partially or fully, it doesn’t mean that the licence holder of the file format would have granted others the right to reproduce it.
b. Export a copy of documents to a format that can be faithfully read by many programs. This compliments the point above since if a format can be faithfully reproduced by many programs it implies that the format is well documented. It does not mean that the format you are saving to today will be in common use in 25 years’ time but it does mean that the likely hood of finding a tool that deciphers it increases in relation to how popular the format was in the past.
c. No company will last forever / no software will last forever. Do not be misled by the fact that just because the manufacturer of your software is the biggest on the planet it will remain so many years into the future. The absolute majority of tech companies will eventually go bankrupt, be bought out, dwindle into obscurity or move out of certain areas of technology. And while they are on top these companies take decisions that effects the software they produce. For example, in Microsoft’s 25 years of selling Word, there have been 12 versions of the product with 5 distinct file formats. The latest version of the program does not natively read files produced by the first version and the currently available filters do not retain formatting well when importing very old formats.
d. Export to a format that is not compressed or encrypted. This allows you to extract the data if everything else fails.
e. Save documentation about the programs you use to edit the stuff you want to archive. Remember that after 25 years you will not have the vaguest recollection of what program you used to create the stuff with. Having the name of the program makes it easier to target your future searches more accurately. This should be in a plain text file format to ensure readability.
f. Save as much documentation as possible about the format itself. Just because today the internet is packed with information about the format does not mean that when your successors are researching archived data for the 100th anniversary of the company, any information about the format the documents are written in would be easy to find.
g. Upgrade your documents to the latest version of a program. A newer version of a program normally works flawlessly with the version of the program it replaces. If the file format between program versions has changed, you should upgrade all your documents to use the file. With office productivity tools this normally necessitates that you open the document and “Save As” it to the new format. After having confirmed that everything is OK, you should delete the files having the old extension if the extension would have changed. If you have a large number of documents you may want to automate the process.
h. Store the archives in a secure place. Not only does this help preserve the media on which the archives are backed up, but it also safeguards against anyone reading them—recall the point above that suggested that such archives should not be encrypted.
This doesn’t mean that you should no longer delete files. Business operations take precedence and filling the working folders with useless documents is counterproductive. Besides, an archive backup is intended to be a snapshot of an organisation and not an organisation modelled to accommodate the snapshot. In order to implement an archive backup system in your organisation does not necessarily mean that you will have to spend an enormous amount of money.
Only time will tell whether in 500 years researches will be trying to decipher optical disks and backup tapes to understand how organisations in the Cambrian period of computing used their rudimentary machines to communicate, socialise and conduct business. Scholars would be debating how society of the time allowed so much of them to be published on social networks that even allowed indiscriminate photos of third parties to be uploaded without the necessary consent.