[fitsbits] start of Public Comment Period on compressed FITS image and tables

van Nieuwenhoven, Richard Richard.vanNieuwenhoven at adesso.at
Thu Jun 25 00:02:37 EDT 2015


Hi,

That's very good, thanks for the very useful hint!

The problem is as developers we are not really interested in the maths
or manually validation of the test data. So a standard providing sample
test data that are so designed that they include most of "special" cases
would be great. If not, more time is going to be consumed collecting
representative test data and fixing bugs than in developing itself.

For example a sample tiled compressed image is good, but it is difficult
to extract partial tests from it. The standard would help a lot if it
would provide for example:

- 3 stets of raw data compressed in every compression algorithm in such
a way the edges of the spec. are hit (rounding and data lossy-ness)
- an small defined not compressed 100x100 bintable / ascii table in
arranged in different combinations of tiles.
...

Such things are extremely difficult and time consuming to come by for a
developer, but not for somebody that defines the spec. It should be a
lot simpler for him. The provided data would help all fits development
not just the one implementation the developer is working on.

    Ritchie


Am 2015-06-25 um 05:02 schrieb William Pence:
> Several compressed images that can be used for test purposes are
> available from the convention Web page at
> http://fits.gsfc.nasa.gov/registry/tilecompression.html.  The fpack
> utility program
> (http://heasarc.gsfc.nasa.gov/docs/software/fitsio/fpack/) also can be
> used to generate other test compressed images from any other FITS image
> using whatever compression parameters you desire.
> 
> -Bill
> 
> On 6/24/2015 1:06 PM, van Nieuwenhoven, Richard wrote:
>> At nom-tam-fits we wan't to implement the compression's and have done
>> so partly already. The
>> biggest problem is that the standard does not provide test data, and
>> in my opinion this is a  prerequisite for a standard like this.
>> In fact a good standard delivers a reference implementation or a
>> compliance test.
>>
>> I can see that a reference implementation is not possible in a
>> community like this, but the delivery of
>> test data would make the live of a developer 1000 fold easyer.
>> Especially if it is available in levels, like separate data files for
>> the different kinds of compression.
>>
>> Sure the compression makes fits more comple, but that's what access
>> library's like ours are for. Especially when the fits files start to
>> get as big as e few gigabytes...
>>
>>         Ritchie


-- 
BSc Richard van Nieuwenhoven
Software Architekt

adesso Austria GmbH
floridotower 26. Stock              T +43 1 2198790-0
Foridsdorfer Hauptstr. 1            F +43 1 2198790-13
A-1210 Wien                         H +43 664 88614710
                                    E richard.vannieuwenhoven at adesso.at
                                    www.adesso.at
-------------------------------------------------------------
         >>> business. people. technology. <<<
-------------------------------------------------------------
adesso Austria GmbH mit Sitz in Wien
Handelsgericht Wien FN231467v



More information about the fitsbits mailing list