Re: [obix-xml] oBIX Scheduling

You are right about needing a definition for an active event, I will
find iCalendar's definition or make one up.

First and foremost, I don't think we want to stray from the iCalendar
model.  You have to ask yourself, can you map the iCal model to your
own?

If you define your schedules as a series of successive start times,
then it is very easy to map dtend as the next start.  In other words, I
don't see an impedance mismatch, the mapping between iCal an your model
seems straight-forward.

If your schedule doesn't have a default value (explicit or implicit),
then are you using it as an event triggering mechanism?  If so, the
section call "Edge Events" is admittedly weak, but covers the trigger
concept.

I hope you can make the meeting, this is good stuff.

Best,

Aaron

Thorsten Roggendorf wrote:

  
  

  
  

  
Hi

  
 

  
“active
events” is not clearly defined in the document.
I assume it refers to events that have

  
(dtstart
< now) AND ((dtend > now)  OR ((dtstart +
duration) > now)).

  
What
about events that have neither dtend nor duration? I think
this is the standard in building automation. In our scheduler the
active event
is (disregarding priorities) that event with dtstart < now that has
the
highest dtstart. Our scheduler has something like dtstop but no
duration. dtstop
is a separate event in our scheduler. We have no default value.

  
 

  
   
Regards

  
        
Thorsten

  
 

  
 

  

  

  
From:
Aaron Hansen
[mailto:] 

  Sent: Wednesday, October 22, 2008 5:50 PM

  Cc: 

  Subject: Re: [obix-xml] oBIX Scheduling

  

  

  
 

  
1. Sorry, priority is 1-9, and 0 means no
priority.  1
is the highest priority.  This is an ical construct.

2. No problem on the op name change.

3. Dashes vs camel case - I was trying to preserve the iCal names (obix
name =
ical property name = xcal element name).

4. I was thinking Vevent and Period type match the parent calendar
type. 
So if the calendar is a bool, then the events and periods would be
bools.

5. Period end vs. duration - just passing along the iCal structure.

  

Brian Frank wrote: 

  
Yes
let’s definitely setup a meeting.

  
 

  
This
is great that we finally have some stuff written down.
 Couple comments:

  
 

  
-
I do not think vcalendar should implement Point –
intuitively it doesn’t seem like calendars are hard or logical points

  
 

  
-
I think it ok that we can’t capture the contract of
consistent values;  no type system can perfectly model every
contract;  so I don’t think we should complicate things by trying to
create specialized bool, real, etc calendars;

  
 

  
-
I don’t get the priority of 1 or 0 – is that the
range?  If so why not use a boolean?

  
 

  
-
getSchedule doesn’t sound like an operation – it
sounds like a JavaBeans getter; I’d vote for something like flatten,
periods, toSchedule, scheduleFor

  
 

  
-
We should translate dashes to camel case for example
last-modified should be lastModified

  
 

  
-
Shouldn’t Vevent and Period have a value slot?

  
 

  
-
Does it make sense for Period to include end and duration when
that can be derived?

  
 

  
 

  

  

  

  
From:
Aaron Hansen [mailto:]
  

  Sent: Monday, October 20, 2008 6:59 PM

  To: 

  Subject: [obix-xml] oBIX Scheduling

  

  
 

  
Lets schedule a meeting to discuss.

  
Issues/Todo:

  
    
Should Vcalendar implement
obix:Point.

    
Should we define calendars and
events of specific types (bool,real...)?

    
defaultValue object under Vcalendar?

    
How can the getSchedule operation be
modified to more closely support free/busy requests?