Abstraction in C

[s0suk3]

I'm developing a project which I'm trying to keep as organized as
possible. Since C doesn't support classes, I've been trying to make do
mostly with structs and functions that operate on them. However,
sometimes some of the "higher-level" modules contain functions that
map more or less directly to functions calls in "lower-level" modules.
So I've been having trouble naming these such functions, since ideally
both of them (the one in the higher-level module and the one in the
lower-level module) should have the same names. So, are there any
naming conventions for these cases? Or are there any alternatives for
doing what I'm doing?

Thanks,
Sebastian

[Fred]

On Jul 18, 1:07*am, [e-mail] wrote:

I'm developing a project which I'm trying to keep as organized as
possible. Since C doesn't support classes, I've been trying to make do
mostly with structs and functions that operate on them. However,
sometimes some of the "higher-level" modules contain functions that
map more or less directly to functions calls in "lower-level" modules.
So I've been having trouble naming these such functions, since ideally
both of them (the one in the higher-level module and the one in the
lower-level module) should have the same names. So, are there any
naming conventions for these cases? Or are there any alternatives for
doing what I'm doing?

Add a short (one or two letter) prefix to all functions in the module.
This has the advantage of making it obvious where to find the code in
a multi-file project.

For exzmple, in Unix, the X11 library functions all start with "X",
the toolkit functions all start with "Xt", and Motif functions start
with "Xm".
--
Fred

[Malcolm McLean]

<s0****@gmail.comwrote in message news:

>
I'm developing a project which I'm trying to keep as organized as
possible. Since C doesn't support classes, I've been trying to make do
mostly with structs and functions that operate on them. However,
sometimes some of the "higher-level" modules contain functions that
map more or less directly to functions calls in "lower-level" modules.
So I've been having trouble naming these such functions, since ideally
both of them (the one in the higher-level module and the one in the
lower-level module) should have the same names. So, are there any
naming conventions for these cases? Or are there any alternatives for
doing what I'm doing?

If you have, say, EMPLOYEE and CUSTOMER structures in your program, it is
quite likely that both will have getname() functions associated with them.

So use emp_getname() and cust_getname(). It's not ideal. In fact if you use
lcc by Jacob Navia, who posts here, you can get C++ style function
overloading. The problem with the lcc route is that the code will no longer
compile under another C compiler.
I don't think it is my job to either recommend or warn against lcc, but to
point out the issues, then you can make your choice.

Another issue you might consider is that if we have

char *emp_getname(EMPLOYEE *emp)
and
char *cust_getname(CUSTOMER *cust)

then we can't have a fucntion pointer to both

char (*getname)(void *obj)

because EMPLOYEE and CUST *s aren't necessarily the same size as void *s.
However you can write emp_getname() and cust_getname() to take void *s
instead of the real structure type, and convert the parameter in the first
line. Then you can do fancy mapping and polymorphism.

--
Free games and programming goodies.
http://www.personal.leeds.ac.uk/~bgy1mm

[rahul]

On Jul 19, 11:18*am, "Malcolm McLean" <[e-mail]>
wrote:

However you can write emp_getname() and cust_getname() to take void *s
instead of the real structure type, and convert the parameter in the first
line. Then you can do fancy mapping and polymorphism.

Just be aware that C won't give you the benefits of type checking and
type safety. The compiler or C lang. won't stop you from passing you
an object of Employee to cust_getname(this may not fail as both the
structures may have a member named name). You will have to introduec
type checking yourself(that is, if you really need it).

[Friedrich]

s0****@gmail.com writes:

I'm developing a project which I'm trying to keep as organized as
possible. Since C doesn't support classes, I've been trying to make do
mostly with structs and functions that operate on them. However,
sometimes some of the "higher-level" modules contain functions that
map more or less directly to functions calls in "lower-level" modules.
So I've been having trouble naming these such functions, since ideally
both of them (the one in the higher-level module and the one in the
lower-level module) should have the same names. So, are there any
naming conventions for these cases? Or are there any alternatives for
doing what I'm doing?

It does not support classes but structs so you can use them

just assume the following structs
struct Employee {
struct EmployeeFtable *funTable;
struct EmployeeData *dataTable;
};

struct EmployeeFtable {
int (*get_id)(struct Employee *);
int (*set_id)(struct Employee *, int nId);
}

struct EmployeeData {
int id;
char *first_name
char *last_name;
};
struct OthterStruct {
struct OtherStruct *funTable;
...

struct OtherStructFtable {
int (*get_id)(struct ....
};
You can use as many struct with get_id as you can...

Howerver you can use the same approach as glib-2 and will probably
gain quit a lot of it.

Regards
Friedrich

--
Please remove just-for-news- to reply via e-mail.