[rkward-devel] plot history: 3rd attempt: request for feedback

[Prasenjit Kapat]

Hi,

On Sun, Sep 12, 2010 at 9:18 AM, Thomas Friedrichsmeier
<thomas.friedrichsmeier at ruhr-uni-bochum.de> wrote:
> Hi,
>
> I didn't find too much time to test, but what I've seen so far looks good, both
> in terms of speed and correctness. I've found some rather minor issues and
> corrected those in SVN, but nothing serious.

Thanks.

I was wondering, if we should remove the "Plot properties" action
completely? With the drop down list implemented, is it needed any
more? There are two extra bits of information there: (1) size of the
current plot and (2) complete plot call. Now, (1) may be useful to the
user, given that we have implemented size restriction. But (2) may not
be useful at all. What do you think?

> I did not look at the code in any detail so far, and will wait with that until
> you remove all the debugging calls (whenever you feel comfortable with that).

Assuming there are no major issues / bugs, say till Wednesday ("testing phase")?

> Admittedly, on first glance this looks a bit scary for the sheer amount of
> code, and I hope some of this can be simplified a bit. But that's for later.

Part of the reason is that I chose to handle lattice calls separately,
unlike windows ()' or quartz ()' implementation! Which I think, others
may / may not agree, is a big feature. Along with that, the modularity
should help in adding more features if / when needed.

> The important thing is that this implementation seems to work well, and we can
> improve it incrementall, later - as far as possible.

Sure. In terms of speed, the crucial step was to disengage the devices
avoiding multiple replay calls. My earlier implementation of
synchronizing the devices seems a bit naive now!

Regards,
-- 
Prasenjit