WF Rules 2011-04-20
===================
Rule 1:   Errors are not the problem; not finding them is.
Rule 2:   Put active Stata files in the working directory.
Rule 2.1: Friends don't let friends click and compute.
Rule 2.2: Do not hard code directory information in do-files.
Rule 3:   It is very hard to have too many comments.
Rule 3.1: Used "// #" to number parts of the do-file.
Rule 4:   If you output wraps, fix it immediately.
Rule 5:   Do not use abbreviations shorter than three letters.
Rule 5.1: Use abbreviations reluctantly.
Rule 5.2: A do-files should create a log file with the same name.
Rule 5.3: Start do-files with "capture log close" and end them with "exit".
Rule 6:   If you lose a file. Take your time. Do not make it worse.
Rule 7:   It is OK to ask for help, but do it carefully and reluctantly.
Rule 8:   Reflect before hitting the send button.
Rule 9:   Name do-files so that they run in alphabetical order.
Rule 10:  Never type anything that you can obtained from a saved result.
Rule 11:  It is faster to document it today than tomorrow.
Rule 12:  Document things necessary for replication.
Rule 13:  The research log only needs essential information, not everything.
Rule 14a: Planning saves time.
Rule 14b: Organization makes it easier to find things.
Rule 14c: Documentation is essential for replication.
Rule 15:  Preserving files is hard but necessary.
Rule 15x: Make at least two copies of important files.
Rule 15y: Store the copies at different locations.
Rule 15a: Verify that copies are exact duplicates.
Rule 16:  Only share results after the files are posted.
Rule 17:  Once a file is posted, never change it.
Rule 18:  Never name it final.
Rule 19:  All datasets need "label data".
Rule 20:  Notes should document the provenance of datasets.
Rule 21:  All datasets should have datasignatures.
Rule 22:  Never change a variable unless you give it a new name.
Rule 23:  Use names that are at most 12 characters long.
Rule 24:  Original names should be 10 or shorter to allow Vx.
Rule 25:  Let the name of a 0/1 variable indicates what 1 means.
Rule 26:  Do not use names that differ only in capitalization.
Rule 27:  Try names and labels before you decide.
Rule 28:  Every variable should have a variable label.
Rule 29:  Put important things in the front of a value label; rarely use more
                than 30 columns.
Rule 30:  Test labels before you post the data.
Rule 31:  Use notes to document the provenance of new variables.
Rule 32:  Categorical variables should have value labels unless the variable
                has an inherent metric.
Rule 33:  Value labels should be 8 or fewer characters in length.
Rule 34:  If a value label is assigned to only one variable, the label
                definition should have the same name as the variable.
Rule 35:  If a value label is assigned to multiple variables, the name of
                the label definition should begin with L.
Rule 36:  Carefully evaluate why missing values are missing.
Rule 37:  Use consistent missing value codes for all variables.
Rule 38:  Make redundant, verified backups in multiple locations of anything
                you want to preserve.
Rule 39:  You must preserve materials necessary to replicate your work.
Rule 40:  Assume your data is wrong. Prove it is right.
Rule 41:  Assume data conversion was done incorrectly.
Rule 42:  Stay as close to the original data as possible rather than
                converting a converted dataset.
Rule 43:  Always create new variables. Never change existing variables.
Rule 44:  When you create a variable, verify that it is correct.
Rule 45:  When you create a variable, add labels and a note.
Rule 46:  Almost never delete a source variable.
Rule 47:  When creating a variable, verify where missing values came from.
Rule 48:  Change the labels of variables Stata automaticlly creates.
Rule 49:  Make data changes only once and save them in a dataset.
Rule 50:  Add two digits to the name of your dataset. Increment the number
                with each revisions.
Rule 51:  If a dataset is temporary, start the name with x-.
Rule 52:  Remember why you are doing what you are doing.
Rule 53:  Do not create variables in analysis do-files.
Rule 54:  The provenance of every result must be documented.

Rule 99:  Enjoy! This too hard if you do not enjoy what you are doing.

===================