Set up study files
The following tasks are done by the Spectre administrator to set up files in the programs directory before any programming commences
- Run the script "crprottmpl" to create "protocol.template"
- Copy "protocol.template" to "protocol.txt"
- Edit "protocol.txt" to put in study specific details
- Run the script "crtitlestmpl" to create the titles template file "titles.template"
Develop report programs
The following tasks to create the reports are done iteratively by each programmer
- Define program name (without extension) to the exported Unix variable "prog"
- Run the script "titles" to edit the titles member to add titles and footnotes (one title needed but the others might be defined in the reporting code)
- Run the script "crtitlesds" to activate the changed titles member
- Develop the reporting code which will use the %titles, %openrep and %closerep macros
- Run the code in batch using the script "sasb"
- Check for errors and warnings alerted by "sasb" completion
- Check the report output which will have a ".lis" file extension
- Make code amendments if required
Run all the reports
Running all the reports is done periodically and also finally by the Spectre administrator
- Run the script "makerun" which will create the script "runreports"
- Run the script "runreports" using the command "./runreports 2>runreports.err"
- Alert the programmers that the run has completed and ask them to check the files "runreports.err" and "runreports.chk" (the latter using the script "myrepchk") to ensure their programs did not create unexpected error and warning messages
- Make a backup by creating a status directory for the run just done and copy (cp -p) the files donelist.txt, runreports.err, runreports.chk, ALL.TITLES and copy all outputs using "cp -p $(donefiles) "
- Run "ddiff" to compare the new outputs with the previous set of outputs (if not the first time outputs produced) and save results
Create the PDFs
Creating PDFs of the collected reports is done periodically and also finally by the Spectre administrator. The description below assumes the output needs to be split into multiple PDFs where a single PDF would be too large for the documentation system. This might be done in the status directory containing the backup of the run just completed.
- Copy and paste the file "donelist.txt" into donelist section files "donesect1.txt", "donesect2.txt" etc. using a text editor
- Run the script "bigps" on each donelist section file to create a PostScript file
- Convert each PostScript file to a PDF using the script "a4ps2pdf" (or "usps2pdf" for US Letter output)
- Run the script "printpdfbookmarks" on each PDF to create a table of contents for each PDF
- Run the script "titlesvsbkmarks" on the collection of PDFs to see if any table/reference numbers used in the .titles members are not present in the PDF bookmarks
- "ftp" the PDFs over to the Windows environment where the documentation system resides
QC the output
The PDFs "ftp"ed to the Windows environment are recommended for use as the basis for QC'ing output since they are very late in the reporting chain and immediately precede submission to the documentation system. Not only programming faults but also potential defects in the conversion of report output to PDFs will be detected here.The following tasks are done by the QC programmer and project statistician.
- QC the output
- Sign off output if correct or report findings to the programmer
Check here for
run and
related to
script
Page
powered by
FREE Go FTP