Can I use Dropbox, OneDrive, iCloud or other online (“cloud”) storage systems to host my ATLAS.ti library?

Can I use Dropbox, OneDrive, iCloud or other online (“cloud”) storage systems to host my ATLAS.ti library?

The short answer is: Don’t do it! It could be detrimental to your project data

Using cloud storage services like Dropbox to store your project data, especially in a teamwork situation, seems like a convenient solution at first glance. However, there are certain peculiarities of such services of which many users are not aware and which, if not considered properly, can lead to damage to your projects to the point of rendering them unusable. While Dropbox and similar services offer many advantages, it really depends on exactly how you use them so you can reap benefits from their use (such as exporting your projects to save backups online). Other ways can be downright detrimental to your work (such as moving your ATLAS.ti library to an online storage system). 

To understand the problem, please consider the basic workings of ATLAS.ti: 

If you mark a segment in a document, ATLAS.ti “remembers” the exact location of that segment in your document from the last time you saved your project. If the document is later changed without “informing” ATLAS.ti about it, references will no longer be accurate and your segments will be off, your coding will misaligned, and hence your analysis would be flawed. 

Thus, if ATLAS.ti finds a document with misaligned segments, it will no longer load such a document in order to avoid further damage to your project. You can manually correct such misalignments at this point, but this is not a desirable thing to have to do and, of course, also frequently a very time-consuming business to boot. 

For ATLAS.ti to reference your documents and their content correctly (i.e., for your data to be usable), the program has to rely on the absolute integrity of these documents. This does not mean that they cannot be changed, but the changes have to made through ATLAS.ti in order for the program to “know” about them and to adjust its references accordingly. 

For this purpose—to guarantee absolute document integrity—we introduced the user and team libraries in ATLAS.ti version 7 onwards. These are specially protected internal folders that serve as repositories for your data files. 

Now, some users have thought it a good idea to move their repositories to a cloud-based system like Dropbox. The thinking is that it would be great to have access to the file from another location or to give access to this same file to other users (in a team, for example). 

Unfortunately, there is a major flaw with this concept. Users overlook the fact that Dropbox will simply replace the existing file with another copy of the same file if a user—either the original author or another user with access to the file—makes changes to this file. From the point of the document alone this makes sense, but from the point of view of an application that requires the document to remain intact this is catastrophic, because Dropbox does not “inform” the application (or the user) of the changes. The required file integrity has been undermined without the user being even aware of it. 

To put it differently, it is the very nature of Dropbox and other cloud-based services that they will simply go ahead and change data in your repository without asking if and when this data has been changed elsewhere–with or without the user’s knowledge or consent. 

This also means that some users may need to update some of the automatic settings on their computers that might be copying the ATLAS.ti library. ATLAS.ti recommends double checking which folders are included in the computer's automatic backup/syncing process, such as OneDrive settings for Windows users and iCloud settings for Mac users.

Therefore, ATLAS.ti strongly advises users against moving the library to Dropbox (or other cloud-synced services/folders). Loss of data or other unpleasantries are likely to result if multiple users work simultaneously with a repository that resides in Dropbox.

“Simultaneous” in this context means “while ATLAS.ti is running.” Taken in this sense, “simultaneous”can thus also refer to two different points in time that lie quite far apart from one another. Consider, for example, the following scenario: 

ATLAS.ti is started on computer A on Jan. 1, on computer B on Jan. 6, and is kept running on both computers. The researcher on computer B saves files to the repository on computer B. The researcher on computer A continues work without restarting ATLAS.ti. ATLAS.ti will not “know” about the files that have been “secretly” changed by the cloud service in your repository in this case and will be unaware of the changes in your documents. This will result in misaligned quotes, unconnected codes, and overall unusable documents. 

We would like to make it very clear that the above is NOT a bug or design flaw of ATLAS.ti in any way. Rather, such problems are the natural consequence of the way in which cloud services work and the user’s choice of making use of such a service. You will understand that ATLAS.ti GmbH will therefore also not be able to accept any responsibility and no liability for any data loss caused by moving your libraries to a cloud-based service.