============= Meeting 14-Mar-2012 =============

Agenda
-----------

0. Preliminaries (5 min)

1. Normative Types (40 mins, *Greg*) [3]
   NTMatrix
   NTVariantArray
   NTNameValue (Guobao doen't like present arrangement)
   more next week.

2. Normative Types self identification review of Andrew's suggestion
(20 minutes) [4]

3. Version release requirement (do we need a beta 1.1 or beta 1.0.1, or not?)
(*Guobao*, 15 mins)

4. Archiver interface demo from David would be great (It's been 2 weeks David!)


------------------------ Minutes ----------------------------

Present: AJ, TK, MK, NM, JR, GS, MA, GW, MS
Scribe: NM

************
NEW TOPIC: Normative Types
************
MK: concerns about ntfield_t (section 3). 
MK: suggests to extend ntfield_t with structure_t and structure_t[]
MK and GW: discussion about structure_t[] in the context of Maser
GW: thinks that Maser uses application-specific types but normative
types should be general.
GS, MK, GW: discussion about scope of normative types
GW: NTTable is not CORBA any
MK & GS: Maser should rely on only normative types
GW & JR: consider that normative types represent only subset 
(90%) of (many) application types, for example, archiver
GS: suggests two extensions of normative types, NTTable and NTNameValue (performance reason)
GW: good idea. If they are generic enough, then we can add
them as normative types.
JR & DH: also consider (probably, non-normative) new type(s) 
for archiver
MK: does not like "NTMatrix" and suggests like "NTNDArray"
JR: NTNDArray is reserved by AreaDetector
AJ: agrees with Marty, probably NT2DArray ...
GW: plans to think about comments ...

************
AI on GS & MK by Mar-22-12: Write proposal to intoduce a new normative type for addressing 
requirements of Maser and (probably) Archiver. Specifically for representation and transport of numbers of channels, where each channel may be a waveform valued channel.
************
GW: suggests to consider and extend NTMultichannelArray

************
NEW TOPIC: Normative Types self identification review of Andrew's suggestion
************

AJ: suggest to apply the aspect-oriented approach
AJ: The bitfield that defined the contents of the type may be extendible for 
data that are extensions of NTypes.
NM: What is teh definition of trait?
GW: let's consider NTAggregate (section 6.6) What is relationship
between types and traits?
AJ: bitfield can be associated with a structure rather than a structure member

All: hot discussion about type ids .... no convergence

GW (with the help of AJ) will try to implement some 
demo to resolve issues.


************
NEW TOPIC: Version release requirement (do we need a beta 1.1 or beta 1.0.1, or not?)
************

GS: needs a version number for his debian package

GW: Suggets libpvdata--20120315

AI on JR by (open): Release new snapshot for c++ called beta-1.0.1