I've seen few examples on the usefulness of a module.  Why use a module
instead of a class?

If I put a method into Module1, how is it called from other classes?

Thanks,
Brett

Methods are easy ways of making shared assemblies. All public procedures are
intrinsically Shared. The module can not be instantiated. So you would call
it simply as module.method


Show quoteHide quote

I believe you mean "Modules are easy ways..." :)

Mythran

Brett wrote:
Because you have global entitites that don't meaningfully belong to any
class. Class instance members are for <things> that have something to
do with a particular object; Class shared members are for <things> that
have something to do with a class in general; Module members are for
global <things> (where a <thing> is a variable, a function, whatever).

OOP purists will decry the use of Modules, and indeed global <things>,
but they have their place.

Module Module1

Public Sub MySub()
'...
End Sub

End Module

then from anywhere in the app:

MySub

In addition to the other replies, you will find previous discussions on this
topic here:

<URL:http://groups.google.com/groups?q=dotnet.languages+vb+modules+class>

Hi Brett,

If I was doing a code review on your code, and you had as much as a single
module in your code, I'd ask you to justify it.  To date, very few have
succeeded in justifying one of these, and most have had to rewrite.

(In C#: a class containing only static methods would qualify the same way).

To date, I've seen valid use for "modules" only when similar, in concept, to
the System.Math class, where an external source of well-understood
algorithms is the foundation for the mechanism.

(I do not use the original Data Access Application Block either, for the
same reason).

In most cases, an object composed of static methods (Module in VB parlance)
is an very strong indicator of very weak design.

Show quoteHide quote

Nick,

I think this strongly depends on what type of application you are
developing:  Reusable libraries, large applications/systems, or small tools
which help you to get your job done.  For the latter design has a very low
prioritiy because code won't be reused and rarely extended.  /Designing/
such a small tool won't be worth the effort and modules are a great
accelerator when developing such applications.

Hi Herfried,

A fair statement.  I realize that my characterization was over-reaching.
When I re-read it, I want to rewrite it.

Also a fair statement.  On the other hand, I wouldn't normally perform a
code review on a "one-off tool" :-)


I find that I cannot disagree with a single statement you made.

My response was over-reaching, as I said.  It was a visceral response,
perhaps.  I would defend the "rule of thumb" that a well designed system of
objects is usually better off if classes composed entirely of static methods
can be avoided.  I'd consider it good practice to avoid this type of object,
regardless of system size.  However, you are also correct: it is frequently
better to set aside such rules when creating small apps that will not be
extended.

Thank you for your response.
--- Nick

Nick,

We don't agree about this last message from you. When a theoretical question
is placed. Than the answer has to be theoretical.

That you with a relative unimportant program (there are more reasons) will
go earlier to a simple sollution, when you don't succeed in a good
theoretical approach, is obvious.

However no reason to take that as the first approach.

In addition, when you review an application than it is a program that has to
be checked on good theoretical implementation, when that is not important
you would not review it.

Just my thought,

Cor

Hello Cor,

May I ask who else you are speaking for?

I respond with all due respect, Cor.  However, I find that I disagree.

As you probably know, I've been posting responses on Usenet groups for about
the last decade.  I do this because I want to help people.  I believe we
share this motivation.  I've had my share of online arguments, but I don't
find them valuable anymore.  These days, my approach to these boards is
simple: "Humbly share, then inform, and finally encourage."  My philosophy
doesn't include the rule of "theory has to be answered with theory."

I'm not sure I understood you.  Here's is what I think you are stating.
Please correct me.  You feel that I would try a good theoretical approach
and fail.  Upon this failure, I would recommend a simple solution.  You are
saying that I would benefit from skipping that first step.  Did I understand
you correctly?

I did not intend to imply that I would take a theoretical approach and fail!
My friend, I can make a good OO program out of *any*simple but useful
application.  (A completely trivial application would look the same as
either an OO or a Structured Programming app, so it would be hard to tell
the difference.  It would still be a good OO program.) Other folks may not
be as comfortable with the OO approach as I am.  For those individuals, my
rules would not be helpful.

It is my personal practice to approach *all* applications this way.  I don't
have any desire to dump my practices on others.  I share my experiences.  I
encourage others to try them.  But I am not offended if another person does
not agree.  Even when that person is a respected MVP, as Herfried is.  I
hope that he is not offended either.

When I am in a position of managing a code base (as I am in my work), I
enforce the coding standards of my team.  On that basis, I will emphatically
reject code that uses classes composed entirely of static methods or VB.Net
modules unless they meet the criteria I explained in an earlier post because
this violates the standards of my team.   Everyone knows the standards
before they write the code. This is not theoretical.  This is entirely
practical.

I beleve that we are saying the same thing, and that I made this point in my
earlier post.  I do not review every snippet of code on a project,
especially code written as a small tool, or something that is not
fundamentally important to the delivery of the application.   I have better
things to do, as so do the smart and talented individuals I have the good
fortune to work with.  I am surrounded by intelligent people.  I am very
very lucky.

While my original statement was more visceral than it should have been, I
stand by my practice of _avoiding_ "global functions" in every application
that I write and challenging it in every application that I am fortunate
enough to inspect.

Nick,

Maybe wrong. However, when I write that I disagree with the one I write too,
than I use "we", because "we" have not the same opinion.  (Although I can
use for the same "I don't agree with you").

I hope not and that comes back in the end of your last message. I wrote my
previous message because I was fully agreeing with your first message in
this thread. Therefore, I was disapproved that you corrected that in a way
that it "seemed" if that one was not right.

Maybe it was not right to place the message in the way I did, however I had
fun when I wrote it in that way, because I expected such a reaction as this
from you. In fact do we full agree (Sorry, do I full agree with you).

I thought I could do that because we share almost forever the same method of
helping. Do not make it to theoretical, just help on the level that is good
for the one who asks a question. As in your nice sentence:  "Humbly share,
then inform, and finally encourage."

Here was the question of the OP "Why use a module instead of a class"

This needs in my opinion a little bit other approach. In addition, you first
answer was very correct in my opinion and needed not any further
explanation.

However, I was disapproved because you made it at least for me sound in
another way when you answered Herfried.

I hope we agree (That you do agree with me) again and thinking it over that
you did obviously express not quot the part "about this message",  did you
probably answer me in the same style as I was doing.

Cor

doh

disapproved has to be disappointed

Please don't read it in the way it is now, this is a very often by me made
error by the way when I am writing.

Cor

Nick,

I don't disagree with what you are saying.  The intention of my post was to
point out that there are people who don't have the knowledge we have and
don't need this level of knowledge to get their work done. It's a legitimate
interest of them to be able to write small, not-reusable tools without
needing to worry about OO too much.  To make a concluding remark, I believe
that even if people like you and me don't make use of modules, others will
do.

You are working in a development company, I think :-).  However, there are
many users who are not developers but need to write smaller applications
simply in order to get their work done.  I agree with you that in larger
teams coding standards play a crucial role, especially if a high quality of
the code is a requirement.

Herfried,

That we (you and me) agree about 100%. Nice text.

Cor

Herfried,

LOL

Nick is working at Microsoft for about almost an half year now. That is a
development company, you are right.   :-)).

He has it in his signature.

Cor

Cor,

That's why I put a ":-)" next to the "I think" ;-))).

VB3 must have been difficult for you then!

Show quoteHide quote

sql Server not found
What does this mean? (nested abstracts...)
What happens to debugger here?
Writing a little app to modify a file
Existsing of File
2d array question...
forecolor of button ???
Installing VS.NET 2005 Beta 1
shell command in windows service doesn't execute
Checking the window state of a running process.