Re: A couple of problems

[Steve Hankin <hankin@xxxxxxxxxxxxx>]

Hi Mark,

I'll respond first to the question about Ferret on SGI, then to the @modvar
question.


  1. There is a solid working version of Ferret V5.0 built on Irix 6.5. The
     problem is that applications are not backwards compatible to older Irix
     systems. We need volunteers to port the Ferret code to other Irix OS
     levels. We can provide source code and Makefile pre-tested on Irix 6.5.
     Ferret V4.91 on SGI was built on an Irix 5.x system. I suspect the
     problems with this running on Irix 6.x are system problems that would be
     resolved when it is relinked on the later system.
  2. I'll need more description of the problem with @MODVAR before I can
     comment specifically. Here is some data I just made up and some analysis
     of it using @MOD and @MODVAR. I attached a GIF file showing the result
     with Ferret V.5 Do these results look correct on your SGI?

use climatological_axes
define axis/t=1-jan-1980:1-jan-1990:1/unit=days tday
let my_base_time_series = SIN(2*3.14*T[gt=tday]/365)
let my_seasonal_event = IF MOD(T[gt=tday],365) LT 50 THEN 2*
SIN(T[gt=tday]/10) else 0
let my_time_series = my_base_time_series + my_seasonal_event
plot/t=1-jan-1980:1-jan-1990 my_time_series
plot/t=1-jan-1980:1-jan-1990/over my_time_series[gt=MONTH_REG@mod]
plot/t=1-jan-1980:1-jan-1990/over my_time_series[gt=MONTH_REG@modvar]

    - steve

=======================================

Mark Williams wrote:

> I have a couple of problems.  Hopefully someone may have some advice.
>
> Firstly I can't get the modular statistic @modvar to give meaningful
> results.  @mod works OK but not @modvar.
>
> Here is an example.
>
> use olr.7496.nc
> use climatological_axis.nc
> let/d olrd = OLR[D=1,Y=15S:15N,T=01-JAN-1980:31-DEC-1995]
> save/file=test1.nc olrd[gt=month_reg@modvar]
> use test1.nc
> fill olrd[d=3]
>
> The results are clearly wrong, however if I do the above with @mod instead
> of @modvar the results appear OK.
>
> Another quite different problem that I have from time to time is that
> ferret partially crashes after producing a graphic.  The prompt yes? comes
> up but when commands are entered they are not echoed on the screen.  After
> hitting the enter key again yes? comes up but on the same line and it is
> not possible to see the key strokes being entered.  The screen looks like:
> yes? yes? yes?
> after 3 commands.
> Ferret is in fact still working but I'm "flying blind".  The only solution
> appears to be to shut down Ferret and restart.  Does anyone else have this
> problem?
> I'm using Version 4.91 on an SGI machine.
>
> Mark Williams
> Supmet (Victoria)
> Ph 61 3 96694902

--

                |  NOAA/PMEL               |  ph. (206) 526-6080
Steve Hankin    |  7600 Sand Point Way NE  |  FAX (206) 526-6744
                |  Seattle, WA 98115-0070  |  hankin@pmel.noaa.gov