Ronald P. Larkin wrote:
> In article <33D6EE25.766C@ibm.net>, Anthony Ayiomamitis
> <firstname.lastname@example.org> wrote:
> > Since when was SAS responsible for explaining to a user the
> > "statistical issues" behind a PROC or an option? SAS is, in one sense, a
> > library of tools for you to use as needed....
> >...I don't think I
> > should be faulting SAS for not explaining it to me in their manuals.
> > There are ... books for this purpose.
> This point of view seems kind of quaint in 1997 because the concept of a
> printed statistics text separate from statistics software is outdated. I
> wonder why, when we do statistics only with computers, we learn statistics
> from books. I look forward to a softcopy, highly interactive introduction
> to statistics (or to a specialty thereof) that is configurable to provide
> syntax, examples, and examinations using any one of several statistical
> Ron Larkin
I am quite of the opposite opinion. When you go and buy a manual for
Microsoft Windows 95, do you expect Bill Gates to explain the
fundamentals behind GUI design, structured programming techniques,
object oriented programming methodology and design, principles of
documentation etc that went behind the development of Windows 95?
I DON'T think so since the purpose of the documentation is to
describe the FEATURES of Windows 95. Why should SAS manuals be any
different? Their purpose should be to describe the features, options and
syntax of the product.
We have enough problems with SAS documentation when it comes to
volume. Now, add pages and pages of examples and discussions, and our
200 pounds of manuals will become 350 pounds.
How many times have you looked for something very particular but had
to go through pages and pages to find it? Can you imagine how that task
would be magnified if manuals were to get that much larger just because
now we are looking for dual-purpose manuals?
My experience in SAS goes back 17 years (solid - day and night). I
still have my first manual ... the one for version 79.8. In this one
inch document you basically have ALL of SAS (at the time) descrcibed. At
best, you had one brief example of syntax for each PROC step.
I am not advocating such a BRIEF document. But at the same time, I
would prefer a document that has to do with SAS (the Statistical
Analysis *** SYSTEM ***) and not a statistics manual with some SASese
Just my two cents worth ....