Dalmady, Otto [mailto:[EMAIL PROTECTED] wrote:

>Unfortunately that's what I was expecting but I thought I would ask.
>For all it's strengths, the biggest weakness I've encountered in Labview
>seems to be with it's interface to other applications (on Windows). 
>I'm sure this is partly due to it's platform independent nature since
>DLLs and ActiveX interfaces don't exist on non Windows machines.
>It's too bad because we could do some really cool things if they were
>better supported.

Well, I do think we do quite some cool things without using ActiveX. And
I do not really see to many things we couldn't do without ActiveX
 
>I've also, in the past, had the need to register a callback function in a
>DLL (non activeX) but the concept of passing a pointer to a LV function
>doesn't exist. Wrapper DLLs, created using other compilers, are often
>needed to solve these issues. This tends to undermine the use of LV in
>the first place.

Not for me. If I encounter such an issue I simply write a wrapper DLL and
am done with it in a few hours. Using Active X costs me just as much time
but I have never had a problem that such a DLL suddenly didn't work after
having upgraded Windows or LabVIEW or even just installing an MS Office
application.

>I think we should include these enhancements in our LabVIEW wish list
>(hope you are listening Mr McKaskle). Perhaps NI developers can somehow
>tie activex into the event structure so a programmer can define events
>that correspond to activeX methods. Both the event and method would be
>created by the programmer. Also, allow for pointer to function capabilities
>for DLL callbacks. 

I have some other peewees I would rather have NI spend their time with ;-)
I think having Active X methods representing event states would completely
obscure the event architecture in LabVIEW. I already find it less than clean
the way it is now.

Function pointer capabilities is another problem. It could be done in principle
with a VI getting translated in a function pointer but the setup of this would
be quite complex, more in fact than the current Call Library function dialog.
A lot of possible problems exist with multithreading or the context in which
the Callback is called, e.g. interrupt time, being some of them.

Active X itself specifically distinguishes between Methods, Properties, and
Events. Why go and confuse these things altogether in LabVIEW?

Your mileage obviously varies but I try to avoid ActiveX as much as possible
and have been quite successful in it.

My reasons:
- Windows only
- still unstable and that is not only the problem of the LabVIEW ActiveX interface
- installation issues with registration of the ActiveX server and version issues
- A very complex technology to create a sort of universal interface which typically
  has low level changes in every new Windows version
- A lot of ActiveX object developers don't understand ActiveX and consequently create
  unstable or unusable objects
- Most "Standard" Windows objects change so often, either in their registered name
  or even worse in their entire Active X Object interface. 
- Windows only and proprietary

Rolf Kalbermatter
CIT Engineering Nederland BV    tel: +31 (070) 415 9190
Treubstraat 7H                  fax: +31 (070) 415 9191
2288 EG Rijswijk        http://www.citengineering.com
Netherlands             mailto:[EMAIL PROTECTED]
  



Reply via email to