Why should we care about file formats?
08/02/2017

Many funders require you to share your research data after your project, to increase its value and reach, and potentially increase your impact – and secure preservation makes your data easier to retrieve and understand if you want to reuse it yourself in future, too. But if data is preserved for a decade or more, how likely is it that it will actually be usable in the future?
We used to store data on DV tapes and floppy disks, but how many people could read these today? My remaining floppy disks certainly make more effective drinks coasters than storage devices! In a similar way to how physical storage formats change, file formats change, too, especially those created by companies that then go on to develop new formats or even go out of business. It’s therefore important to choose file formats wisely.
Wherever possible, you should save in an open format (such as csv), rather than a proprietary format (such as xls) that requires specific software to read it properly and is unlikely to be as interoperable. Some proprietary software only runs on particular operating systems, which excludes some users from being able to open the files. Such proprietary formats might not even be fully readable by the same software a few versions down the line – Matlab and Microsoft products in particular are known to have this issue. (Microsoft Works was discontinued in 2009 but files need to be converted to be readable today and may lose their formatting or some content in the process; have you also encountered Compatibility Mode issues in Word due to version changes?)
So to help keep your data usable for the long-term, try to choose from an open format in the table below; the more extensive UK Data Service format advice is also a valuable reference.
Type of data | Recommended format(s) | Format(s) to avoid |
Qualitative textual data | .xml, .rtf, .txt, .pdf (PDF/A) | .doc |
Quantitative tabular data | .por, .csv, .tab | .xls and other proprietary database formats |
Images | .tif, .png, .svg, .jpg | .psd |
Audio | .flac, .wav | .wma, .ra, .ram |
Video | .mp4 | .wmv, .mov, .avi |
If you’re depositing the file in CORD, our new data repository, we will accept all formats at the moment but are reviewing how to handle these in order to ensure the files’ long-term usability, as we are required to retain data accessibly for 10+ years. And of course sometimes you cannot avoid using a proprietary format – this is fine, but it will be important to make this clear when depositing the file on CORD or another repository, so anyone viewing the record will know what requirements there are to be able to access the data in the file. See the RDM intranet site for more advice and contact details for further help.
Image: “I find your lack of space disturbing” by N. Hussein, CC-BY-NC-SA 2.0, available at: https://www.flickr.com/photos/nhussein/3583560407/
Categories & Tags:
Leave a comment on this post:
You might also like…
Register for a British Library Reader’s Pass
The British Library (Euston, London) is open to students and researchers who wish to use its world-class collections and resources. To do this, you need to register for a reader's pass. Begin by reading the ...
Unlock your full searching potential with a personal account in EBSCO or ProQuest
Some of our journal-searching resources include features that are only available when you have registered for an individual account. These features can help you to: Save time Document your work Keep up to date. Sound ...
How to use an appendix in your assignments
You may wish to include appendices in your assignment where appropriate, for example to include a survey, raw data or other relevant material which you refer to in your work. The following are some of ...
Bank holiday hours for Library Services: Monday 5 May
Library Services staff will be taking a break on Monday 5 May for the early May bank holiday. You will still be able to access all the resources and help you need via our library ...
How do I reference… a table of data from multiple sources?
If you have read our previous APA7 post on Referencing ... tables, you will know how to cite a table of data taken from another source, but when you are creating a new table which ...
Finding full-text Economist articles…
If you’re looking for The Economist, the place to go is ProQuest One Business. Follow these step-by-step instructions to get full-text access. Login here and click on the Publications option at the top, above the ...