Re: Re[2]: paper vs online

[Bill White <bwhite -at- INMET -dot- CAMB -dot- INMET -dot- COM>]

>> My question is, how do we structure information so it does fit and stay
> in memory for the amount of time it takes to return to the application
> and apply it?

> Eric
> ejray -at- okway -dot- okstate -dot- edu


This depends on the information.  Some information is hard to remember
because it is arbitrary, but has to be captured precisely.  Other
information is not so arbitrary, or its use does not have to be so precise.
    o For example, if you are programming, and you need a type definition,
      or the name of a type definition or a funcition, the document
      reading tool should allow you to cut the declaration out of the
      tool and paste it into your program.
    o If, however, you are proving a theorem, and you want to apply a
      lemma, you either have to have the other lemma in your head, or
      else you have to have the other lemma's proof in your head.  I
      suppose the second is more likely.  In any case, you don't need to
      recall the precise statement of the lemma.  Any logically
      equivalent sentence will do nicely.
    o If you are following a procedure, say an installation procedure
      for some piece of software, you probably have to have the procedure
      written down in front of you as you run the procedure.  Most people
      who install something like TeX or emacs care only that things go
      in the right place, but they don't care where the right place is.
      If they have to know what each file is and make a determination
      about where each file goes they will not be happy with the procedure.

There is another question which I believe is related, but I am not sure
how.  It's very important to figure out how much information to provide.
    o If you are following a procedure, say an installation procedure
      for some piece of software, you probably have to have the procedure
      written down in front of you as you run the procedure.  Most people
      who install something like TeX or emacs care only that things go
      in the right place, but they don't care where the right place is.
      If they have to know what each file is and make a determination
      about where each file goes they will not be happy with the procedure.
    o There is a large software company in the PC business whose C/C++
      manuals are the worst I have ever read.  There is practically no
      information in them.  The reason, I have decided, is that the
      technical writers have decided, or else have been told, that
      programmers aren't smart enough to understand what is really
      happening when they program, and so the technical writers have to
      present a highly simplified model.  In particular, I am talking
      about segmented memory, but there are other examples.  This sort of
      works if the model works, and programmers don't stress it
      very much.  In the particular case I am thinking of, the model
      does not work, and any call to a library routine, or to any I/O
      procedure overstresses it.

                                        Peace,
                                        Bill White