Cls's vs. Mod's

[Networks]

Lately I've become really confused between classes and modules. Some people tell me I use them too much others say it's ok in any event I would like to know when it is best to use a class over a module and vice versa.
For example let's say I want to create a lastseen index using an array, would a class be best for that? For a sub or function that parses user commands, which do I use? How a database that stores users and allows editting of them,  which do I use? I'd like also to go back to basics..What do classes allow you to do differently from modules and vice versa. I've heard different things from Stealth and Sidoh, and other people as well. What If I don't ever unload my database class until the application ends should it have been a module? Anyway please post.

[hismajesty]

We discussed this in detail here

[Stealth]

Classes are for code that needs to be duplicated, and accessed on a custom scope. Because you can instantiate a class as many times as you want, you can create copies of the same code dynamically as your program runs. Also, you can instantiate classes on any scope - module-level, global, procedure-level, as you please. They're a lot more flexible in that way.

Modules are for code that does not change and needs to be globally accessible.


IMO:

Your lastseen index that uses an array, is probably best implemented as a class. That way if you need more than one index you can instantiate another one. Also, you can write a simple wrapper for the Collection class that you can reuse later.

Your database is probably best implemented as a class as well.

Your function that parses user commands can go either way.

[ChR0NiC]

For a sub or function that parses user commands, which do I use?

Well Functions return data, Subs do not. Ask yourself, is this procedure going to return any data perhaps a true/false statement, using the write type can also help save memory

[Grok]

For a sub or function that parses user commands, which do I use?

Well Functions return data, Subs do not. Ask yourself, is this procedure going to return any data perhaps a true/false statement, using the write type can also help save memory

Hmmm...

Code Select Expand

Public Sub ReturnSomeData( ByRef DataSetToReturn As DATASET)
   'modify DataSetToReturn
   With DataSetToReturn
   .Val1 = 1000
   .Val2 = 2000
   .Tm1 = Now
   End With
End Sub

[ChR0NiC]

I meant for example

Code Select Expand


Private Function Blah(Data As String) As Boolean
If Data = 1 Then Blah = True
If Data = 2 Then Blah = False
End Function


You can't do that with a sub.

[Eli_1]

I meant for example

Code Select Expand


Private Function Blah(Data As String) As Boolean
If Data = 1 Then Blah = True
If Data = 2 Then Blah = False
End Function


You can't do that with a sub.

Ignoring how you forgot quotes.  

Code Select Expand


Private Sub Blah(ByRef sData As String, ByRef bResult As Boolean)
bResult = IIf(sData = "1", True, False)
End Sub

[ChR0NiC]

To hell with you all then, fine !!

[Networks]

For a sub or function that parses user commands, which do I use?

Well Functions return data, Subs do not. Ask yourself, is this procedure going to return any data perhaps a true/false statement, using the write type can also help save memory

You took the question the wrong but I understand how you did my bad =p

I meant should I use a class or a module for my function/sub that parses commands.

[Newby]

Can you do this exact thing with a sub?! o.O

Code Select Expand

Dim I As Integer
I = SomeFunctionName(ArgumentOne)
NOPE. You'd have to do.

Code Select Expand

Dim I As Integer
Call SomeSubName(I, ArgumentOne)
Which looks unsexier.

[Eli_1]

Can you do this exact thing with a sub?! o.O

Code Select Expand

Dim I As Integer
I = SomeFunctionName(ArgumentOne)
NOPE. You'd have to do.

Code Select Expand

Dim I As Integer
Call SomeSubName(I, ArgumentOne)
Which looks unsexier.

There you go lieing to everyone again, Newby.  The "Call" and two parenthasis are completely unnecessary.