On 8/19/07, Fraser Jackson <[EMAIL PROTECTED]> wrote:
> The default behaviour for viewmat and plot is clearly different but both are
> defined as classes and can be used to open as many windows as wanted  (see
> plot_class.htm).  It is easy to name the windows to distinguish them if
> wanted.

It's simpler than that, as anyone willing to put together a dozen lines
of pd commands can accomplish a lot with plot.

But that's not what I'm talking about -- I'm talking about people using
the single-line interface to plot (or viewmat, but I did not report any
viewmat problems).

> The viewmat default is fine for the situation where beginning users want to
> use a colour display of some function.  However I cannot see any advantage
> in changing the default plot class behaviour. In experimenting with a plot
> using a script and ctrl-r you could easily  end up with a whole series of
> windows to close. For me at least, a graphic goes through multiple
> refinement steps and I do not want to have the whole sequence as separate
> windows.

I do not see that plot and viewmat are all that different.  Anyways,
none of this addresses why I get an error with the new command
(identical error with the use command and assigning PForm directly).

> > If someone runs open'jzplot' then accidentally types something
> > and uses control-Z to undo that typing, they lose the contents of
> > the window.  This is one keystroke away from losing a significant
> > chunk of J -- this seems too easy and of no particular benefit.
> > Perhaps J should have a way of discarding undo history.
>
> This seems to be slightly overdoing the risk.  Someone who is opening the
> class scripts to modify them should have enough knowledge to know that if
> you click on close, then answer Yes to the question if you want the script
> modified something will be changed in the system.

But people might be using open'jzplot' to study the code with no
intention of modifying it.  Here, the risk is considerably higher:

[1] As the documentation for plot does not cover all issues, examining
the code is sometimes the only good way of figuring things out.

[2] As people who intend to read the code are not likely to make
any intentional changes, the risk of undoing all changes after
they accidentally make a change is considerably higher than for
users who are intentionally modifying the code.

[3] People reading the code to study it would typically be less advanced
than those who are prepared to modify it.

-- 
Raul
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to