Re: [Ecrit] draft-ietf-ecrit-similar-location-03 feedback - additional

[Randall Gellens <[e-mail]>]

Instead of a boolean, it might be worth considering an integer, to 
inform the client how many similar locations could match.  Depending 
on the situation, the client might report an error and attempt to get 
a less ambiguous location.

At 9:45 PM +0000 10/17/16, Dan Banks wrote:

>  I'm thinking of an example where you might a large apartment 
> building, and a query that did not include unit or floor.  If the 
> response only lists the first 5 units as similar locations, a 
> person looking at this might (incorrectly) assume that the server 
> doesn't have data for the other 13 units in that building.  Of 
> course, there is already a good disclaimer in the draft not to make 
> such assumptions, but the more often we can give the users a 
> complete list of similar locations, the more effective this 
> extension will be.  So, I think an explicit indication will be 
> helpful.  At the very least, it should save a support call or two 
> ("If the LVF knew that unit 501 was valid, then why did it only 
> suggest 101, 103, 201, 202, and 203?")
>
>  Dan
>
>  From: Rosen, Brian [mailto:[e-mail]]
>  Sent: Monday, October 17, 2016 4:24 PM
>  To: Dan Banks; [e-mail]
>  Subject: Re: draft-ietf-ecrit-similar-location-03 feedback - additional
>
>  I don't really have a problem with having this option, but one 
> question I have is what would the client do with that information? 
>  I can't think of any behavior change that return of that 
> information would trigger.
>
>  Brian
>
>  From: Dan Banks <<mailto:[e-mail]>[e-mail]>
>  Date: Monday, October 17, 2016 at 4:18 PM
>  To: "<mailto:[e-mail]>[e-mail]" 
> <<mailto:[e-mail]>[e-mail]>
>  Cc: Brian Rosen <<mailto:[e-mail]>[e-mail]>
>  Subject: draft-ietf-ecrit-similar-location-03 feedback - additional
>
>  There is one thing I would like to add to the feedback which I sent 
> recently on the similar location draft:
>
>  Section 4 discusses briefly the possibility of the server limiting 
> the number of returned similar locations.  Although the current 
> text expresses the general idea that there may be a few that are 
> the "most likely" to be the correct location, there are also 
> scenarios where many similar locations could all be equally likely, 
> and the server might need to simply cut off the list at a 
> reasonable count.  In those situations, I believe it would be 
> helpful if the server indicated when such  a limit is actually 
> applied.
>
>  I suggest that an attribute be added to the locationValidation 
> element of the findService response, perhaps 
> rli:similarLocationsLimited, having a data type of boolean, with 
> instructions that a server SHOULD include this attribute if the 
> number of returned similar locations is limited due to 
> configuration or policy.
>
>  Dan Banks
>
>  _______________________________________________
>  Ecrit mailing list
>  [e-mail]
>  https://www.ietf.org/mailman/listinfo/ecrit


-- 
Randall Gellens
Opinions are personal;    facts are suspect;    I speak for myself only
-------------- Randomly selected tag: ---------------
Don't worry about what people think; they don't do it very often.