Re: Persnickety comment about visad tutorial

Hi Kevin,

> I just have one little comment about the visad tutorial, which overall
> is very clear and useful.

I agree, Ugo and Tom are doing a great job, and the new sections
they're working on are going to very helpful.

> In Section 3.1 "Using a Integer2DSet" it is stated that:
> <quote>
> We assume we have some NROWS x NCOLS pixel values. The important thing
> to observe is that the pixel samples are in raster order, with component
> values for the first
> dimension changing fastest than those for the second dimension.
> 
> That is, suppose we have an array of 6 values per row times 5 per
> column, like
> 
>        pixel_vals = {{0, 6, 12, 18, 24},
>                      {1, 7, 12, 19, 25},
>                      {2, 8, 14, 20, 26},
>                      {3, 9, 15, 21, 27},
>                      {4, 10, 16, 22, 28},
>                      {5, 11, 17, 23, 29}  };
> 
> then these values should be ordered as the values indicate.
> </quote>
> 
> This is obviously a contradiction. The example shown is in column
> major(FORTRAN) order, while _raster_ order is row major(C/C++, Java,
> PERL, ...) order by definition.

You have a point, but the tutorial is very careful not to say
either "row major" or "column major".

I think the error is where the text says "6 values per row times
5 per column" which should be "5 values per row times 6 per column".
That is,

  NCOLS = number of values per row, and
  NROWS = number of values per column.

> if im not being annoying enough :) shouldn't the section title be "Using
> an Integer2Set" ?

Blame the proof readers (including me) for this.  I could
never get my "der", "die" und "das" right in school.

Thanks for your help,
Bill
----------------------------------------------------------
Bill Hibbard, SSEC, 1225 W. Dayton St., Madison, WI  53706
hibbard@xxxxxxxxxxxxxxxxx  608-263-4427  fax: 608-263-6738
http://www.ssec.wisc.edu/~billh/vis.html