> Well, back in September 2008 David replied to a similar question about TAN
> and PV with:
>
> 1) The header describes a TAN projection, but also contains headers
> PV2_1 and PV2_2 which are used to define additional parameters for
> some projections, but not the TAN projection. The TAN projection does
> not use or require any PV keywords.
>
> A previous version of the AST library (used by GAIA, Convert, etc)
> made use of this fact, and attached its own special meaning to any PV
> terms found with a TAN projection (they were used to describe a
> distortion to the TAN projection). This was changed earlier this year
> so that AST now attaches no special meaning to extraneous PV keywords.
> It sounds like you are using the old version of AST.
>
> Your message has, though, raised another issue, in that now, rather
> than using the PV values for its own private purposes, AST reports an
> error if it finds any such extraneous PV keywords. Whether it is
> strictly illegal to include such extraneous junk in a header is a moot
> point, but it is certainly bad practice. On the one hand you could say
> that AST should simply ignore such unnecessary keywords, but on the
> other hand reporting an error flags that something is wrong with the
> header that should really be fixed.
>
> end quote
>
> So the implication is that David thought he'd fixed things so that TAN would
> ignore PV headers.
So, a couple of other people looked at this same file, but I believe using
an older Starlink release (whereas I'm using a bleeding-edge version),
which seems to jive with what you are saying. The map contains proprietary
data, but I can replace the data array with random numbers or something
and send it around tomorrow perhaps.
Ed
|