Comment On Select Just In Case

SteveInRichmond:

looks like a general handler for { check, radio, combo, select }
control types, with additional handling for 'select' and 'radio'... it's
possible the developer's intention was to eventually add the general handling code for
these four control types at some point in the future, or that the code has been "maintained"...

Come on, guys, this one is obvious.

It's simply using CASE to evaluate ControlType regardless of its case, then using IF to handle lowercase values within the CASE, and in case the case of ControlType is not lowercase (perhaps uppercase or a mix of cases) it is handled outside of the CASE or excluded by the CASE statement.

Just in case you can't figure it out ... 

Rick:

I really like that an uppercase control type will pass the 'switch' and fail on the 'if' test.

Actually no it won't. When he makes the LCase() call in the "Select Case" line, it will compare the lower-cased version of the control type, while not actually changing the case of the variable. I'm surprised that this guy would know to do that, but didn't have enough smarts to just remove the Select structure altogether.

If StrComp(controlType, "select", vbTextCompare) = 0 Then

ElseIf StrComp(controlType, "radio", vbTextCompare) = 0 Then

....

End If

Rick:

I really like that an uppercase control type will pass the 'switch' and fail on the 'if' test.

Manni:

Actually no it won't...

Dammit I'm a retard. Sorry, you are totally right. Still, my StrComp solution would work just fine.

Jeff S:

Come on, guys, this one is obvious.

It's simply using CASE to evaluate ControlType regardless of its
case, then using IF to handle lowercase values within the CASE, and in
case the case of ControlType is not lowercase (perhaps uppercase or a
mix of cases) it is handled outside of the CASE or excluded by the CASE
statement.

Just in case you can't figure it out ... 

Jeff S:

Come on, guys, this one is obvious.

It's simply using CASE to evaluate ControlType regardless of its case, then using IF to handle lowercase values within the CASE, and in case the case of ControlType is not lowercase (perhaps uppercase or a mix of cases) it is handled outside of the CASE or excluded by the CASE statement.

Just in case you can't figure it out ... 

Well, in that case, case closed![:P]

Ytram:

You just always want to have your bases covered.

Sorry, couldn't help myself.

sacha

Anonymous:

I wonder if there is a lurker on this board
working on a vblint, rubbing his hands together with glee every time a
new vb post goes up.

Anonymous:

Huh, this seems pretty reasonable to me. Whats the problem? There was
some ** general code ** for all 4 types snipped out, and the 2 if's
handle the specific sections just for those controls in addition to the
general code. Get it?



case ** general case **:



    if (**specific condition**)  do something specific

    if (**specific condition**)  do something specific



    ** general code **



    break;

Select Case LCase(controlType)

  Case "check", "radio", "combo", "select"

    If controlType = "select" Then
      'ED Snip ...
    End IF

    If controlType = "radio" Then
      'ED Snip ...
    End IF

    'ED Snip ...
End Select

I don't know much about VB, but isn't there an equivalent of

case default:

    // do stuff

    break;

Wouldn't that be where you put so called "general stuff"?

To me this just looks like what happens when someone tells an intern "Obviously you'd want to use a switch for that." And this is what they come up with.

This looks like a minor WTF and more of a logic error than anything else.

Yes, he forgets to compare the inner control types in a Case Insensitive manner, but that's more of a logic error than anything else.  Nobody here can say they have never created a logic error before (with any honesty anyway).

Now, that said, sure the code is redundant.  However, it does allow for the possibility of all four control types to share one section of code while allowing for customization of specific types. 

Notice that he only shows two types specifically mentioned in the If statement, while he covers four in the case.  This tells me that it's likely code that changed from when it was originally written (or he wrote it for something that never came into play).  I'd say this is simply a case of stumbling upon code that was maintained.

Sure, as it's written, aside from the logic error, it's fairly silly (and only really requires one extra if check, so it's not exactly killing performance).  I would only write somehting like this IF I was planning on handling all those types with the same logic and had a few exceptions crop up.  It could be handled by doing what he did or calling a function in each and every case.  This is actually better self-documenting and looks cleaner. 

Sorry, it's a minor WTF at best. 

bugsRus:

This is really just an optimization.  He won't have evaluate all of those expensive ifs if the case doesn't match.

But in all truth, this make my head ache.

Boy, I hope this is a joke.

WTFer:

It is probably a military system, or an hospital system. The redundancy is needed for it to be failsafe.

As an employee of a leading company that makes hospital systems, no, we don't do that.

We never, never, ever do that.

however anyone with half a clue would realise that to do that 'switch'
you need to do the comparisons anyway.  Think about what you
just said, "oh if none of those match then he won't need to do any of
those expensive if statements", what do you think a switch/case
statement looks like once 'compiled'?

scheky:

Sorry, it's a minor WTF at best. 

Anonymous:

WTFer:

It is probably a military system, or an hospital system. The redundancy is needed for it to be failsafe.

As an employee of a leading company that makes hospital systems, no, we don't do that.

We never, never, ever do that.

As an employee of another leading company that makes hospital systems with the development done in India (BTW, I love losing control of the application I've developed for the last 7 years), you'll never know what wackiness they'll come up with next. [:P]

smitty_one_each:

vblint?

Anonymous:

bugsRus:

This is really just an
optimization.  He won't have evaluate all of those expensive ifs
if the case doesn't match.

But in all truth, this makes my head ache.

Boy, I hope this is a joke.

Anonymous:

however anyone with half a clue would realise that to do that 'switch'
you need to do the comparisons anyway.  Think about what you
just said, "oh if none of those match then he won't need to do any of
those expensive if statements", what do you think a switch/case
statement looks like once 'compiled'?

Anonymous:

I don't know much about VB, but isn't there an equivalent of

case default:

    // do stuff

    break;

Wouldn't that be where you put so called "general stuff"?

Yes, there is an equivalent for the code you posted. But it is for non-covered possibilities. Select 'breaks' after a Case statement. Hence, it would never get to the 'case default'. General code goes outside of the Select statement. Is it so difficult to write a function that does the common stuff?

kengray:

To All Those "Base-Coverers":

No No No!

if you need an if, use an if; if you
need a case, use a case.  To say you  should use an if inside
the case statement "just to cover your bases" makes me wonder whether
you should even be working in the "vodoo-wicky" field of computers and
programming. 



PS: sorry if i offended anyone, but if we start to question the
validity of functions like "case" then we need to start questioning the
validity of mathematical functions like addition as well.

Anonymous:

bugsRus:

This is really just an
optimization.  He won't have evaluate all of those expensive ifs
if the case doesn't match.

But in all truth, this make my head ache.

Boy, I hope this is a joke.

Anonymous:

however anyone with half a clue would realise that to do that 'switch'
you need to do the comparisons anyway.  Think about what you
just said, "oh if none of those match then he won't need to do any of
those expensive if statements", what do you think a switch/case
statement looks like once 'compiled'?

Anonymous:

Anonymous:

however anyone with half a clue would realise that to do that 'switch'
you need to do the comparisons anyway.  Think about what you
just said, "oh if none of those match then he won't need to do any of
those expensive if statements", what do you think a switch/case
statement looks like once 'compiled'?

It appears you are assuming that select gets compiled into a long
if/else sequence.   This might be true, but there is no
reason to assume so.   I can think of other ways to do it,
and they might be faster for some situations.


Select COULD build a state machine, checking
each letter for all currently valid combinations - each letter can be
an index into an array of function pointers.  If there are many
cases where this case won't match, only 4/26 will need to compare two
letters, while the standard if/else will need to compare 4 letters just
to see that it matches none.



Another obvious optimization to select it to build a hash table for all
possible values.   This might be more CPU time than a 4 place
if/else sequence, but if there are many other cases it could be a win.



A smart compiler could recognize a long if/else sequence and do either
of my suggested optimiztions (and perhaps more that I have not thought
of).  In the real world such optimizations are hard to
recognize.  I don't think any compiler does that, but one could.



That still doesn't explain why you don't have a different case for each
option.  (Unless there is a lot of common code two all four that
Alex sniped without telling us, but if that is the case there should be
a function call or something to handle the common code)



Of course I'd be surprised if visual basic was smart enough to actually
do a select as anything other than a long if/else.   For that
matter I'd be surprised if many real world select statements have
enough different cases, and are used enough that the cost in extra code
is worth the code bloat. 



Come to think of it, the most common way to do select in python is a hash table of fucntion pointers.

WTFer:

It is probably a military system, or an hospital system. The redundancy is needed for it to be failsafe.

Anonymous:

Actually, Rick is right. The 'if' test doesn't lcase() the control name, so 'SELECT' = 'select' will fail.

Unless Option Compare Text is on.

Anonymous:

Ytram:

You just always want to have your bases covered.

All your base are belong to us.

This code is fast. If it fails the case check it will jump over the if statements.

If controlType = "Check" it will jump into the case statement but all the If's will fail.
Or he forgot all the lcase's in the if lines.

I wonder how he implemented the case else, else if construction.

redtetrahedron:

Anonymous:

Actually, Rick is right. The 'if' test doesn't lcase() the control name, so 'SELECT' = 'select' will fail.

Unless Option Compare Text is on.

Then why would he be using LCASE() in the CASE?

Jeff S:

redtetrahedron:

Anonymous:

Actually, Rick is right. The 'if' test doesn't lcase() the control name, so 'SELECT' = 'select' will fail.

Unless Option Compare Text is on.

Then why would he be using LCASE() in the CASE?

That's assuming the author knows what s/he is doing. [;)] This is a WTF, so you can't assume anything...