> Oh well... it seems like there's no way to ask a question without it's
> meaning being high-jacked :-)
>
Yep. :-)
> You are one of the original designers of .NET and you have the option of
> adding a feature that you know will have performance repercussions. You
> also know for a fact that some people, either through ignorance or their
> malignant nature, will misuse this particular feature. My question was
> what are therefore the overwhelming reasons that still persuade you to go
> ahead and implement the said feature? Are there any more reasons than the
> obvious ones that meet the eye?
>
To answer your question generically, yes, I still implement the feature--if
people misuse it, on their heads be the consequences. One can only protect
people from themselves so far.
To answer your question specifically, I've not had the need or desire to
play around with the exception filters in the CLI, so I can't speak to that.
I can hypothesize, but nothing beyond that.
Ted Neward
{.NET || Java} Course Author & Instructor, DevelopMentor
(http://www.develop.com)
http://www.javageeks.com/tneward
http://www.clrgeeks.com/tneward
----- Original Message -----
From: "Cristian Diaconu" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Friday, May 03, 2002 2:50 PM
Subject: Re: [DOTNET-ROTOR] [OT] double-pass exception semantic - Why?
> Oh well... it seems like there's no way to ask a question without it's
> meaning being high-jacked :-)
>
> Ted, I don't believe either Don or myself ever argued that using
> exceptions for flow of control is an acceptable practice. I can't read
> Don's mind, but as for myself, I promise you I don't do it even when
> nobody's watching :-).
>
> Second, nobody is "blaming the tool for their problem" like you said
> either. This is *by no means* the intent of my question. I tried to keep
> it as cool and factual as possible.
>
> But picture this for a second:
> You are one of the original designers of .NET and you have the option of
> adding a feature that you know will have performance repercussions. You
> also know for a fact that some people, either through ignorance or their
> malignant nature, will misuse this particular feature. My question was
> what are therefore the overwhelming reasons that still persuade you to go
> ahead and implement the said feature? Are there any more reasons than the
> obvious ones that meet the eye?
>
> So the focus of my interest is not .NET bashing, but simply trying to test
> my understanding of a design decision from the point of view of the
> original designer, *not the user*.
>
> I could have asked the same as it relates to Win32's SEH as well, but
> that's not the point of this forum and in the case of Win32 no language
> generic appeal was being sought. .NET appeals to language interop so
> obviously I assume that filters must be useful in other languages that I
> haven't heard of. If so, what are those languages/circumstances that
> require it?
>
> Thanks
>
> Cristian
>
> On Fri, 3 May 2002 13:11:03 -0700, Ted Neward <[EMAIL PROTECTED]>
> wrote:
>
> >That's sort of like saying that "you can't guarantee that violence isn't
a
> >standard practice for resolving conflict; however dubious you personally
> may
> >feel it is, this is a standard technique for many people". Just
> because "it
> >happens" doesn't mean we shouldn't try to STOP it. :-)
> >
> >The fact is, if they want to engage in dubious practices, then they pay
> the
> >penalty for doing so. In this case, the penalty is decreased speed and
> >performance. They want to improve the speed, they need to come into line
> >with how the tool's supposed to be used. Let's not blame the tool and/or
> >platform here for what's essentially their problem!
> >
> >Ted Neward
> >{.NET || Java} Course Author & Instructor, DevelopMentor
> >(http://www.develop.com)
> >http://www.javageeks.com/tneward
> >http://www.clrgeeks.com/tneward
>
