Re: [gtkmm] General C++ Design Question: relation of GUI toapplication

["Stephen G. Tashiro" <tashiro trac wsmr army mil>]

> >>What's my score? How many questions did I answer, versus NEW questions? --
>
    One thing I don't understand is why  static gint _callback( void *arg)  { ....
is preferable to static gint _callback( Foo* arg) { ...
Experiment shows the GCC compiler will accept things like  static
gint _callback( Bar* arg) { ... where Bar is a class.
However, I don't know whether the code does what is expected.

The separation of Gui and Application is a noble sentiment, about which I don't
need persuading.  The question is how to accomplish it.    As the architect
Christopher Alexander points out,  a useful approach to conscious design is to
identify pairs of requirements that are mutually contradictory.   In computer
programming, I would say these are  1) individuality or eccentricity  versus  2)
standardization and uniformity.   On the one hand, nobody wants a big book with
guidelines on how he should name his variables.  We want to program as if were were
speaking to ourselves in a private language which we are constructing as we go.  On
the other hand, when we build on the work of others, we don't want them to keep
changing things and forcing us to re-learn how their system works.   My interest is
finding a system of doing things that strikes me as a happy compromise between
these.

Stephen Tashiro

tashiro trac wsmr army mil
tashiro zianet com
"Never let the task you are trying to accomplish
  distract you from the study of computers."