A. Validate

   1. Run NSGMLS on the received file. 
   
   2. May need to add character entities to 
      the entities file (propagate additions
      to all copies).
      
      (see any Readme.doc that may have accompanied
      the file)
      
AA. Sample

   Note: this procedure will require some modification if there
   is a large number of unnumbered pages, or if there are 
   multiple sets of pagination that overlap.


 (1) Get list of page numbers:
 
   1. Extract list of page-break tags from the .sgm file
   
      (Using TextPad, search-in-files, specifying file name,
      search for <PB[^>]*> file-type=binary, regular expression,
      all matching lines; copy the list to another file; remove
      the line-number info [replace ^[^:]: with nil])
      
   2. Supply artificial page numbers for <PB> tags without "N" values,
      using decimal values (770.1, etc.)
      
   3. Reduce to bare numbers (replace <PB N="  and "> to nil).
   
   4. Save list of page numbers to a file ("pages.txt")
   
 (2) Use random number generator to generate sample
 
   5. Open Excel, import the text file.
   
   6. Note the number of cells occupied by the list, i.e. the
      total number of pages. Calculate the number of samples
      needed (total pages x 0.05). Use the Excel "sampling"
      option (under Tools/Data analysis) to generate a 
      random sample containing the necessary number of pages.
      Sort it. Look for duplicates. Replace duplicates with
      arbitrarily selected pages from the original set.
 
 (3) Copy sample pages from .sgm file
 
   7. Open the .sgm file in a text editor.
   
   8. Replace <PB with \n\n<PB to make page breaks stand out
   
   9. Proceed through the file copying and pasting the
      pages corresponding to the sample page numbers into
      a separate file.
      
   10. Close the .sgm file without saving.
   
   11. Save the file of samples as xxx.test.sgm
       (where xxx is the NOTIS ID no.)
       
 (4) Create .html file for proofreading
 
   12. At the command line in the working directory
       for the title in question:
       
       copy \mecode\perl\cmeA\vnd2html.pl
       
       perl vnd2html.pl xxx.test.sgm > xxx.test.html
       
   13. copy \mecode\perl\cmeA\striptag.pl
   
       perl striptag.pl xxx.test.sgm > xxx.stripped.txt
       
       (this creates a "stripped" version in order to establish
       a character count against which to evaluate transcriptional
       accuracy.)
       
   14. Load the html file in a browswer; if there are obvious
       uglies or inadequacies, revise the perl script and/or
       make specific changes in the .html file, e.g. adding
       missing closing tags.
       (Look especially for characters that do not display
       satisfactorily.)
       
       Otherwise, print the file.
       
 (5) Proofread the print against the original, noting errors
     or omissions of transcription.
     
B.