Next: Summary
Up: Schwab: Preparations
Previous: Testing of documents
Storing a document on cdrom leads to three significant consequences,
which an author needs to know about when writing a document:
- Space limitation:
Currently SEP cdroms can contain 350 MBytes of document. The authors
of a document have to ensure that their document does not exceed
350 MBytes (clean your document with Tour clean and then
run du -s . in your top directory). Usually,
data files and figure files consume most of the space.
You may want to consider compressing some of your files using gzip.
Unfortunately, removing data and figure files affect directly the
reproducibility and beauty of a document.
The space limitation also prevents us from offering on a cdrom the entire
computational environment we enjoy on our home system. For example, we
only supply a subset of SEPlib executables on each cdrom (See SEPHELP
cdrom).
- Read-only file system:
In a read only file system, existing files are not simply overwritten.
Attempting to write in such a file leads to errors which do not occur
on SEP's standard file system. Please make sure that your clean rule removes
all intermediate files. It would be great if someone could come up
with a test for this (e.g. changing a the file protection on a documents
files after cleaning it up).
- Unknown mount point:
The cdrom can be mounted at very different points on very different
systems. It is safe to assume that, when the cdrom is mounted
at Elf in the French back country, there will be no directory called
/scratch/me. A rule using such a directory is bound to fail.
Instead the author should limit himself to refer only to files and
directories on the cdrom and by using relative path names. Especially,
since SEPlib header files point to their corresponding data files
by absolute pathname, the author needs to edit the header file or
to combine the header and data file into a single file (cite SEPlib).
I wish I could think of a way to test this.
Next: Summary
Up: Schwab: Preparations
Previous: Testing of documents
Stanford Exploration Project
3/8/1999