> Arty Ecock wrote:
> >    Why not adopt cpint and be done with it?  Why re-invent
> the wheel?

> I think that the new module is the better soloution. It
> splits the work into a userspace component and a kernel
> module proper instead of trying to deal with everything in
> kernel space. Everyone is free to chose his/her prefered wheel now.

I disagree, for technical and user interface reasons.

Technically, having the command fail or truncate output because the user
didn't anticipate the size of the response correctly seems a major step
backwards. Principle of Least Astonishment dictates that the command should
do the Right Thing w/o user intervention. Hcp does, vmcp does not, forcing
the user to deal with a unnecessary (and incompatible) implementation
restriction.

Second, from a user interface perspective, previously there was a single
clearly indicated method to do this task that was well documented and is
already in the distributions. We now have two methods, both accomplishing
the same thing, and one demonstratably technically less capable. IBM will
undoubtedly use one method, and the rest of the world will likely continue
with CPINT. This does no one any favors -- we now have two methods to
maintain, and have doubled the confusion possibilities for J Random User.

Not good by any measure.

If there's something wrong with cpint, say so, and we'll fix it. Otherwise,
I think you're doing the community a disservice by introducing this
"feature".

-- db

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390

Reply via email to