Raw notes from OBIX meeting 16 April

[William Cox <wtcox@CoxSoftwareArchitects.com>]

For (mostly Toby's) correction emendation and update as draft minutes.

Many items discussed; actions are labeled ACTION.

If you avoided the meeting, you have no actions :-)

Thanks!

bill

--

William Cox 

Email: wtcox@CoxSoftwareArchitects.com 

Web: http://www.CoxSoftwareArchitects.com 

+1 862 485 3696 mobile

{\rtf1\ansi\ansicpg1252\cocoartf1038\cocoasubrtf360
{\fonttbl\f0\froman\fcharset0 TimesNewRomanPSMT;}
{\colortbl;\red255\green255\blue255;}
\margl1440\margr1440\vieww10500\viewh9600\viewkind0
\pard\tx720\tx1440\tx2160\tx2880\tx3600\tx4320\tx5040\tx5760\tx6480\tx7200\tx7920\tx8640\ql\qnatural\pardirnatural

\f0\fs28 \cf0 Selected notes from Obix TC meeting 20150416\
\
See also Jira item comments this date.\
\
Scribe: Bill\
\
[11:16] William Cox: 227 - historian write this down in this format. Other pattern - that history I asked about I'd like it now in this format.\
\
[11:17] William Cox: T: pick one. Which pattern? Were saying care about large data sets? Do you want to require taht the server is building an enormous doc, and  you'll say what  you want when you order it, not when you pick it up?\
\
[11:17] William Cox: C: Based on initial request on OBIX side? T: if small, hardly care.\
\
[11:18] William Cox: Do you want to reformat as you serialize it?\
\
[11:19] William Cox: Do you want to "push an XSLT and have it processed"? Know what you asked for, and what you want, AND that you can transform.\
\
[11:20] William Cox: T often record for a week, I'll get it in a week.\
\
[11:20] William Cox: C: one thing definite - the data in the historian is almost not stored in obix XML, stored in a format efficient for the historian. In the 1.0, already a requst to take the data you have and transform and send.\
\
[11:21] William Cox: T: format when you ask for it?\
\
[11:21] William Cox: If specify formats, have to decorate the URI to pick it up, or to have it saved?\
\
[11:24] William Cox: B: irrelevant the storage form, what's important is what you've asked contratct?  Could an obix server optimize that is being asked for a lot of stuff? legal, service.\
\
[11:24] William Cox: T: BUT\
\
[11:24] William Cox: T: what if one request came an hour later? Think of aspects of the service that assumes db optimization?\
\
[11:27] William Cox: T; if I can have different formats - decoration on the initial request. on the Request to write down.\
\
[11:27] William Cox: Reqest to pick up does not change format. (Does it include it or reference it?O\
\
[11:27] William Cox: C: kind of OK with what we've agreed on, figure out the implications for the spec language.\
\
[11:28] William Cox: History filter request in obix is the request - decoration (specify a format string).\
\
[11:28] William Cox: B: I've asked for this format, when and if you give it to me it will be in this format.\
\
[11:28] William Cox: C: Then get a URI and the obix client can do an HTTP GET to get what's there...\
\
[11:29] William Cox: B: give me the data in the format I asked you for.\
\
[11:29] William Cox: T: Assume since doing URIs - past the mechanism - here's a URI initiating a request ?zinc\
\
[11:30] William Cox: If I don't have the "?" then I do it the old school format and we're done.  If i want Zinc do ?Zinc and servercan say I don't know Zinc, refuse, or says "sure". Gives backward compatibility.\
\
[11:30] William Cox: C: Server or the client? Old server won't know the format string.\
\
[11:31] William Cox: T: It refuses it - ill formed request. But if request of a client gets the old way.\
\
[11:31] William Cox: T: Skating past the "?" trick. The old school server says HUH? And old school client won't add it.\
\
[11:32] William Cox: T: so all that remains is "is it ?zinc or something else?\
\
[11:33] William Cox: C: foo.bar - put something in the example, if implementing a server would give back a uri that's somewhat ref\
\
[11:33] William Cox: B: not standardized URIs.\
\
[11:33] William Cox: C: because not mandating anything about request - in WD39/40 the answer is a history query. If can't decorate, possible that you might be doing the two pass request always.\
\
[11:34] William Cox: T: Why are we straining over that?\
\
[11:34] William Cox: C: Bryan's comment.\
\
[11:37] William Cox: C: new behavior - request for these dates, really inefficiently encoded.\
\
[11:38] William Cox: T: Whether or not a URI in the middle "can you give me ?zinc" and the servercould say "error" or "yes and here it is."\
\
[11:39] William Cox: If set and forget in the future, might have something static - give me a URI and I'll pick it up later. Allows long running requests - part of the thing. Bryan lives part of the time in "give me last 5 minutes"; but the Zinc issue is around "give me six months."\
\
[11:39] William Cox: C: sounds like what he's saying is give me zinc for five minutes.\
\
[11:40] William Cox: B: large effort, large complexity for no significant benefit.\
\
[11:41] William Cox: T: 10 years ago can't do because I have a 386 in the closet. Now you have a 10 core... earlier was COAP, simple. More powerful than it used to be. Much of what we and CEA are talking about aren't tied to these fancy linguistic patterns.\
\
[11:44] William Cox: <xs:simpleType name="contract">\
<xs:list itemType="xs:anyURI"/>\
</xs:simpleType>\
\
[12:16] William Cox: T: if a contract list is empty it must be expressed as Nil.\
\
[12:16] William Cox: C: Why do we have an empty contract list? Nil is used as an empty contract list?\
\
[12:17] William Cox: T: If a contract list must have a member then I express Nil as a value, and may not be a member of any multivalued list?\
\
[12:18] William Cox: C: Something Tridium may have assumed but hasn't been expressed formally - capture that obscure knowledge.\
\
[12:18] William Cox: C: Something Tridium may have assumed but hasn't been expressed formally - capture that obscure knowledge.\
\
[12:18] William Cox: T: must be explicit. Nil Nil Nil\
\
[12:20] William Cox: see 11.1 - core contract - WD39 1533 -\
\
[12:20] William Cox: T: with corrections for voicing are the two sentences (added to the section here)?B: no, needs to be in 4 and referenced here.\
\
[12:21] William Cox: ACTION B propose wording.\
\
[12:21] William Cox: C: something that us'ns understand is the case, but if didn't understand that might implement differently.\
\
[12:24] William Cox: C: most in list have put comments in Jira.\
\
[12:24] William Cox: ACTION T comment on https://issues.oasis-open.org/browse/OBIX-243\
\
[12:24] William Cox: T: controls world has always gotten alarms wrong. Events. alarms are events you have to do something about.\
\
[12:25] William Cox: T: three of us should have arguments on the thread to wake up the people who've been asleep. NOT Jira.\
\
[12:26] William Cox: I think before CSPRD03 approved all the changes. Chair was going to move to closed.\
\
[12:28] William Cox: ACTION B do WD41 with proposed for OBIX-238\
\
[12:29] William Cox: ACTION T clean up old issues from CSPRD03 and earlier that are closed.\
\
[12:30] William Cox: ACTION B make PDF posted. With changes and questions shown and comments included.\
\
[12:30] William Cox: (Last PDF posted)\
\
[12:33] William Cox: B proposed normative wording about this stuff.\
\
[12:34] William Cox: ACTION T change 23 April meeting to Workshop. Bill can't attend.  Next regular meeting is 30 April.\
\
[12:34] William Cox: B will provide WD41 early next week.\
\
[12:34] William Cox: Adjourned}