[ Previous chapter ][ This chapter ][ Next chapter ]
In order to analyse data with a computer, you must get the input data into electronic form. This can be achieved by typing in a query string to a retrieval program, or by entering a sequence in a specific program (called editor) via the keyboard. Alternatively, you can use data which have been entered before by yourself or others. These data are stored on the local computer's hard disk or can be accessed via electronic networks. Data transfer , therefore, is important to you if you keep data for analysis on different computers, such as your desktop computer and the mainframe.
NOTE:
If you keep data on different platforms with identical names,
make sure that
you are aware of the synchronization problem:
Changes on one computer are not automatically
propagated to the others.
SECURITY NOTICE:
Depending on your environment, data on personal or mobile computers
are potentially
unsafe as password protection on these devices is
usually not guaranteed. This applies particularly
to industrial environments.
Unfortunately, computers show a rather low tolerance with respect to data processing if the format of the data is unknown. Therefore, many programs will require that you honor certain conventions with respect to the type of data to be processed. Many program packages will do this automatically for you, but different program packages require different formats. Therefore, reformatting will be of importance if you use different programs in different environments.
Many programs do not only take a single input but also allow changes in the way a specific algorithm is run, or how the output shall be presented. Many of these options will not necessarily require that you know or change the suggested values (called 'default' values) , but some of the parameters will affect your result severely. Make sure you know about the power of the parameters before you change or modify parameters. The GCG programs as described below in the BioCompanion, if run via command line, will present default values in braces, such as a value of 40 in this form: (* 40 *) . Other options will not be shown at all and require that the option is entered by commandline and typed explicitly. Many public domain software packages follow a similar approach.
All programs will produce output in one way or another. Some will display only a value, some will write output files, and others will present a figure as a result. Sophisticated programs will take the output of another program as input. It is important that you review the output critically.
SECURITY NOTICE: Bad data don't smell. The plain fact that a computer program completed and produced output correctly is not an evidence that the output is correct.