Re: [Gpsbabel-misc] GPX file structure

[Robert L. <rob...@us...>]

P. Roosen wrote:

> Hello Babel folks,

I guess that makes is babbling heads...

> when designing import and export functions for my waypoint exchange site 
> (www.gpspunkte.de [german]) I hit on a small discrepancy w/r/to GPX tag 
> treatment:

It's entirely possible.  I was one of the first users of GPX and the
spec had (and has, IMO) a number of vagaries about it.

> As fas as I understand, the <desc> tag was intended to include 
> information *not* to be exchanged with the GPS receiver. When d/l 
> waypoints from a Magellan I received the potential comment field as 
> <desc>, though, instead of the aspired <cmt>. I think this is a design 
> fault on gpsbabel's tag interpretation.
> 
> Could any one here crosscheck this, please?

Looks like you're right.  That sure complicates things...

The problem is that we have several sources of free-form text data which
may map into zero or more sources of free-form text data.  Any given
input may set zero or more of shortname, description, notes, and now
cmt.  What's a valid <cmt> for receiver X may not be valid for receiver
Y so we have to cleanse them.  If we don't have a cmt but do have a
descr, we "obviously" want to use that to cleanse and use.

So yes, it looks like we inherit yet another field in struct waypoint
that we'll probably get to conditionally test and use in each of the
backends.  Ick.

Fixing it for the specific case you describe is only slightly
ugly.  Trying to DTRT for all combinations of input and output is a
heart-sinker.

Does anyone in the audience have a clever idea for addressing this?

RJL