Regarding: Framemaker to text/SAS conversion

Has anyone out there in SAS-land had experience converting Framemaker files to SAS data sets ? (or just to text for that matter)

The Framemaker (version 5.5) files are in a Windows NT environment. Ultimately the files will be uploaded to a UNIX System V Release 4.0 where SAS resides (version 6.12). I also have SAS version 6.12 available in the Windows NT environment if needed.

Here is the task. I have inherited about 150 Framemaker files in a Windows NT environment. (This is my first exposure to the Framemaker product.) Each Framemaker file is a questionnaire of about 10-25 questions. (In clinical trials, these are the inclusion/exclusion criteria). Obviously, there are graphics on the page such as checkboxes and frames for elements. What I want to do is save the file as close as possible to the original format (i.e. like a snapshot) while eliminating the graphics. (i.e. preserving all the text) This would allow me to then write SAS code to parse out the questions and assign them to variables. However, doing a straight "save as text" file does not seem to work, as elements within frames get moved around and don't preserve the original alignment at all. I assume in part this is related to anchoring and frames. Is there a simple way to preserve this alignment? If I can achieve this alone, I would be quite happy.

Now, assuming the first part is doable, here is the frosting on the cake: When saving as "text", all the symbols and special characters imbedded within the questions become "?" or "??". It would have been nice if the program would have had the smarts whenever possible to convert graphics characters to a text equivalent. Such as:

the "greater than or equal symbol" to something like: >= the trademark symbol to (TM), superscripts (as seen in lab values) to 4 x 10*6

A typical questionnaire will have a dozen of these "problems" per page.

Thus, the big question is how to handle these 2 problem situations "programmatically" if at all possible since I have 150 files to convert. What is the easiest way to create the analogous text value substitutions without doing a zillion manual "find these problem characters and do a global change and replace" in either Framemaker or other product? Creating files in other intermediate file formats (i.e. RTF,) is acceptable if it lends itself to a solution. We also have most of the popular software PC products.

For your reference: Framemaker can output files in the following formats (I'm excluding ones I have no access to such as for the MAC and WordPerfect):

Text Export Filters - HTML - MIF - Microsoft Word 6.0, 7.0 for Windows - PDF (Portable Document Format) - RTF 1.2, 1.3

Additional for UNIX - Interleaf 5 (IAF 6.4)

Graphics Export Filters (when embedded in exported text files) - EPS - EPSF - EPSI (TIFF and WMF Preview) - FrameVector - PEG - PICT - TIFF - WMF (Windows Metafile)

Graphics Export Filters (to HTML) - Compuserve GIF - JPEG - PNG (Portable Network Graphics)

Thanks in advance for any creative solutions!

Paul Gill Author of "The Next Step: Integrating the Software Lifecycle with SAS Programming" Director of Bioclinical Data Services Inc. (408) 629-6399