One could imagine translating all of our files into documents accessible to modern word processors, spreadsheets, etc. There is a continuous upgrade path of translators going from these formats to the next format that comes along. The advantage of this approach is that our entire data set would be usable by all the common applications of any one time. Unfortunately, this approach also requires that we translate our documents each time a new format comes along. Many of these translators seem to lose a small fraction of the semantics of the original document; for example, the rows in an ASCII text table may no longer line up. In passing through translator after translator, more and more formatting information may become lost or garbled, until the document becomes unreadable. Therefore, if we are going to translate our documents, it is important to maintain the original document and to choose a format that will remain stable for a long time.