[Difx-users] difx2mark4 bug {External}

[R. Craig Walker]

Note that SCHED preserves the case of the station codes in the catalog.  
But if you specify station names by code in a schedule file, the 
comparison with the catalog is done case independent. Upper case 
versions of both are compared.  So, if you specify two stations in the 
catalog with the same letters, just different cases, you can end up with 
problems.

SCHED also allows the use of 3 characters which I think will become more 
important as the bigger arrays come on line.  They might need to be even 
longer.  At some point, one wonders if the codes are even that useful.  
I think their main advantage is when making human readable tables, such 
as those in Sched of baseline lengths, they can be reasonably compact.  
Maybe there should be just one name/code with a maximum character 
count.  Even as it is, there are multiple antennas in the catalog with 
the same code - usually different antennas or pads at single stations 
such as the VLA or the DSN stations.

Cheers,

Craig



On 2/14/24 8:43 AM, Geoffrey B. Crew via Difx-users wrote:
>
> I knew I was opening a can of worms on this one.  As a relative 
> new-comer I was bit by this sort of thing a few times...and just 
> assumed I was the only one confused.  Now I understand:
>
> Fortran did not have good string handling and the (original) character 
> set was limited; hence capitalization in AIPS.  VEX came later, 
> particularly *after* K&R decided character (string) handling was 
> something that needed to be fixed in C and there is no fixed 
> convention other than Vex can handle mixed case and AIPS cannot.
>
> With arrays such as SKA, ngVLA, &c under discussion, it seems that 
> perhaps 2 characters are not enough and some guidance within the DiFX 
> code base on "what is the plan" might not be a bad idea so that the 
> developers can move into some coherence in the future.
>
> And if AIPS is to survive into that era, some sort of mapping of 
> station codes into all-upper-case two letter codes might eventually be 
> necessary.  difx2mark4 already has an option to provide a map of 
> two-letter codes to one letter codes.  (And we already decided HOPS4 
> isn't going care and will use what is in the VEX.)
>
> Somewhat earlier Leonid wrote:
>>> Dear colleagues,
>>>
>>>     Can you clarify, is there a convention on a case of two-letter station
>>> codes:
>>> gs, GS, or Gs?
> -- 
> Geoff Crew
> MIT Haystack Observatory
> gbc at mit.edu
> (NOTE NEW EMAIL: gbc at haystack is no longer reliable)
>
> _______________________________________________
> Difx-users mailing list
> Difx-users at listmgr.nrao.edu
> https://listmgr.nrao.edu/mailman/listinfo/difx-users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listmgr.nrao.edu/pipermail/difx-users/attachments/20240214/3030ce43/attachment.html>