>>>>To me, getPem() returning an object is equally non-obvious as a cryptic macro expansion. And if we assign this object to a properly prefixed local variable, it's even more readable than a macro or an eval() call.
>>>>
>>>>But then, de gustibus non est disputandum.
>>>
>>>Taste should never be as important as engineering. :)
>>
>>"Less obvious" tastes like having a taste... towards "I like this better because I'm used to it". For me, getpem() is pretty much a straight system call, while eval() or a macro are workarounds.
>
>Dragan
>
>You've convinced me. I take back my statement. It should not be avoided, just properly documented.
>
>MS should document that GETPEM will return any contained object and not just PEMs. As long as variables are named appropriately GetPEM should be fine.

But they did document:) Any contained object is just a member of PEM:)
Cetin