On Mar 20, 2007, at 11:48 PM, M Pulis wrote:
>
> On Mar 20, 2007, at 9:38 PM, Sam DeVore wrote:
>
>> Ok on some back and forth with my client (never underestimate the
>> idiocy of a client) it turns out that they were seeing one of my
>> exception handler messages.
>
> And the client is the idiot??!?? The client without whose input you
> would not have caught a problem that could have been caught on your
> desk with regression tests?
>
> Seems to me you should thank your client for diligence, patience and
> using your product, rather than insult them behind their back.
> Consider the possibility that, from the client's perspective, you
> could be considered the "idiot" as you are the source of your product
> and its apparently indecipherable error messages.
the client was an idiot (for reasons not really related to this
issue) because when they were presented with the window that popped
up to inform them of the untrapped error (which is my fault I agree
there should be no untrapped errors) There is a button that they can
click to email me a report that sends the current call stack and the
values of some of the variables. They have done this before during
this development cycle (this is a clearly marked Alpha version that
they have offered to test as part of the contract).
Instead they told me it was a different error window that they have
never seen before. They even reported the error message as being in
all CAPITALS, which they were not. After spending a large part of
the day going back and forth with versions of the application that
had significant logging added and more error handlers added (which
are good things in the long run).
It turned out that they were looking at my error handling window and
when they sent the report from the window it was much easier to
replicate the problem, create a work around for the timestamp issue
and get them moving on to the next stage of the testing. This is an
application that is being transitioned from 5.5 and they are well
aware of the need for the reports.
My point (though poorly stated in a time of frustration) is that even
if one presents your users with what you think is a logical error
window, that makes it easy (in your mind) to send you information
that can help you isolate the issue, do not assume that the user will
read it, use it, or report it correctly.
And I whole heartedly agree that making that assumption makes me the
one with the ID-10-T problem.
They are paying me to transition this code and part of the agreement
is that they are responsible for helping to test it against the test
database in parallel with their old version
Yours in shame
Sam D
_______________________________________________
Unsubscribe or switch delivery mode:
<http://www.realsoftware.com/support/listmanager/>
Search the archives:
<http://support.realsoftware.com/listarchives/lists.html>
