Agreed.  As I said, these were off-the-cuff comments.

I wasn't thinking about porting between a real 407 and pipes.  I was
thinking about keeping consistent with the model architecture being
emulated for the sake of the humans who have to understand 407E, a
non-trivial task.

"Simply resisting change" was also not my point.  If it were me, I would
want to know that any impact of a change would add a significant capability
that did not already exist, and which could not otherwise be achieved.  In
this case, adding SELECT SECOND would seem to be a tolerable work-around,
versus a number of broken pipes where the knowledgeable people are long
gone..  Also with 6.4 having updated pipes means that there may not be an
easy fallback when a bunch of things go wrong all at once.

I, too, am a purist that prefers clean, consistent, and easy-to-use
interfaces.  If costs vs. benefits justifies a change, then I'm behind you
110%.

Have a good weekend.

On Fri, Feb 19, 2016 at 8:40 AM, Rob van der Heij <[email protected]>
wrote:

> > Two off-the-cuff comments:
> >
> > I'm really rusty on the interaction between the various 407E features.  I
> > think it's possible that this is consistent with the actual 407.
>
> We don't have that many people exchange programs between their real 407 and
> the CMS Pipelines code ;-)
>
> > Second, there may be breakage if you "fix" this.  There may be pipes out
> > there that are using this behaviour, either intentionally or
> inadvertently,
> > which might produce different results if this is changed.
>
> Yes, that is a concern with anything that gets fixed. There are quite a few
> cases where we can't go back and change things the way we want because too
> much water went through the pipeline. We do care about compatibility.
>
> This particular issue shows even in the book - the example with that option
> just did not work. If your plumbing relies on options not doing what they
> are
> supposed to do, then maybe your pipe will break. I tend to challenge people
> to explain me what about the erroneous behaviour was so desireable.
>
> Simply resisting change does not work. Plumbers tell me that not fixing
> bugs
> and not adding new function for two decades is not good either ;-)
>
> Sir Rob the Plumber
>



--
OREXXMan

Reply via email to