[Rets-dev] Feedback requested - RETS 1.7 COMPACT format <COLUMNS> interpretation

[Paul Stusiak]

This is a request for feedback on the interpretation of the COMPACT and 
COMPACT-DECODED response formats <COLUMNS> tag to all developers of 1.5 
and 1.7 systems. In particular, this is targeted at server vendors over 
client vendors since server vendors originate the response.

A recent thread identified an ambiguity in the 1.5 and 1.7 specifications.

The specification is silent on the field names used in the <COLUMNS> tag 
of the COMPACT family of response formats. The specification should 
probably give some indication of what may be returned, even if it says 
that it doesn't specify it.

Two interpretations have been presented:

1. The field names used are the SystemNames.
2. The field names used are dependent on the state of the StandardNames 
argument

A third and fourth interpretation is possible.
3. The field names used may be intermixed SystemNames and StandardNames.
4. The value of the <COLUMNS> field names is an implementation detail.

I would like to know which of the above interpretation is used by your 
implementation. There may be other interpretations possible, so if you 
do something different, please let me know.

It seems like we should consider either an engineering note or a point 
revision to the 1.7 specification if we can come to a consensus on the 
interpretation.

There is anecdotal evidence that 1 was favored by Bruce Toback, however 
there is no definitive statement that I can find. I think that option 4 
is probably not very desirable for client developers.

Impact
-------
If necessary, either an engineering note can be created to clarify this 
point or the specification can be changed. If the specification is 
changed, an estimate of the impact is given below.

Option 1. Changes to section 7.4.2 and section 13.
Option 2. Changes of option 1 and a change to section 7.4.7
Option 3. Change to section 13.
Option 4. No change.

Paul