Re: [Geopriv] I-D, Action:draft-george-geopriv-lamp-post-00.txt

From: Henning Schulzrinne ^lt;hgs@cs.columbia.edu>
Date: Mon Jul 05 2010 - 14:36:52 EDT

I admit that I fail to see the practical difference for software. You can obviously already specify an empty field if you really insist, but there are two cases:

- the recipient needs the information - being told that we know it to be empty isn't too helpful, since that's not a legal value

- the recipient doesn't need the information - whether this is <foo/> or just omitted makes no difference.

We should strive to keep things simple - special cases that are likely to be done wrong in practice don't seem to help here.

Henning

On Jul 5, 2010, at 2:27 PM, Geoff Thompson wrote:

> Brian-
>
> It seems to me that this is somewhat different than "if you don't have an element, you don't include it".
> The cases you cite are more "we do have the element, the known name is an empty value".
> (that being significantly different from "unknown" or unsupplied for an unknown reason)
>
> It seems simpler to change the syntax than to do something like what the US Feds did a number of years ago. That was to decree that (in order to qualify for federal money for local rescue squads) every house in your area had to have a street address with both a registered street name (beit a private or public street) and a monotonic "house" number. No surprise, that imposed a significant real world implementation burden with lots of hangover.
>
> I look forward to the feedback from the cognoscenti
>
> Geoff
>
> On 4/7/10 3:23 PM, Brian Rosen wrote:
>> While I do want to change the basic way the PIDF schema is defined, enforcing a restriction in the schema seems like a poor choice. I am loth to define a null road name; that's a big change to the way PIDF works; if you don't have an element, you don't include it, rather than include it with a null value.
>>
>> I'll ask around the PIDF cognoscenti to see what they think.
>>
>> Brian
>> On Jul 4, 2010, at 3:45 PM, Geoff Thompson wrote:
>>
>>> Brian-
>>> It seems lik a bad idea to let perfectly good assumptions (like a MP or a house number "needs" a road name) get screwed up by unusual exception as you have noted.
>>>
>>> Why not, instead, just charge ahead but allow a special value of road name ("NULL" ?) that notes that in this case it is known that a road name does not exist? The entered value should probably be differentiated from that which would be provided if the value of road name were "unknown".
>>>
>>> Geoff
>>>> Message: 2
>>>> Date: Sat, 3 Jul 2010 16:53:18 -0400
>>>> From: Brian Rosen<br@brianrosen.net>
>>>> Subject: Re: [Geopriv] Fwd: I-D
>>>> Action:draft-george-geopriv-lamp-post-00.txt
>>>> To: Carl Reed<creed@opengeospatial.org>
>>>> Cc: geopriv@ietf.org
>>>> Message-ID:<917C86FD-5CF8-42D8-AF6C-18339497D1FA@brianrosen.net>
>>>> Content-Type: text/plain; charset="us-ascii"; Format="flowed";
>>>> DelSp="yes"
>>>>
>>>> I am a co-author on this draft, and the milepost addition to the lamp
>>>> post original was motivated by the NENA requirement to have a milepost.
>>>>
>>>> I agree that the MP nearly always needs a road name. So does a house
>>>> number. The current documents don't have requirements like that,
>>>> primarily because we keep finding odd cases that don't meet what we
>>>> thought was a simple rule.
>>>>
>>>> Brian
>>
>>
> _______________________________________________
> Geopriv mailing list
> Geopriv@ietf.org
> https://www.ietf.org/mailman/listinfo/geopriv
>

_______________________________________________
Geopriv mailing list
Geopriv@ietf.org
https://www.ietf.org/mailman/listinfo/geopriv
Received on Mon, 5 Jul 2010 14:36:52 -0400

This archive was generated by hypermail 2.1.8 : Mon Jul 05 2010 - 14:37:26 EDT