Agenda

-----------

1. Proposed ICALEPCS papers. Please give some thought to abstracts,

 due at end of February.

2. Recent work in streamlining basic APIs (Matej)

- asyncRPC

- Unifying java and cpp simple APIs (add “PVObject get()” ?)

- Adding monitor to easyPVAJava

- easyPVACPP - what role does Sinisa’s existing work have for this.

- Collecting up experiences and simply stated user level requirements

  would be a good start.

3. Short status updates

Minutes

----------

Present: GC, DH, GS, MK, MS, GW, RL

Chair: GW

Scribe: GW

******

NEW TOPIC: ICALEPCS Meeting

******

GS: Possible paper on V4 for beam commissioning

GW: Overall EPICS V4 summary

Re EPICS meeting Tutorial, at MSU, MAY 2015

DH: Re Oak Ridge meeting, some but little discussion on prep for EPICS meeting in East Lansing.

Probably monday for the training.

******

NEW TOPIC: Tutorial and Examples help

******

Group: Some discussion of complexity of pvAccess API especially w.r.t. server side.

DH proposing that pvDatabaseCPP makes it very simple.

GS: saying it is pretty difficult and giving BNL experience.

MK: Agrees that for asyn of course asyndriver is recommended, but for

non-asyn non-driver use cases pvDatabase is good.

AI on DH: Write chapter in Getting Started on writing a pvAccess server with pvDatabase.

*********

NEW TOPIC: Recent work in streamlining basic APIs (Matej)

*********

GW: Do we need a “pvObject”?

MS: Depends on what users want?

Discussion: MK: In Sinisa’s pvObject, each get gets you a new pvObject - so for instance different gets with different pvRequests, are a new pvObject.

MS/MK: underlying code does cache connections.

MK: Looking at easyPVAJava for monitors

… one thing Sinisa had was like “getDouble” on the channel.

MK: Propose adding monitors to easyPVAJava first, then turn to simplifying easyPVAJava API before pvDatabaseJava.

GW:MK: MK proposes adding a optional “poll” function for updating - which would be perfect for Matlab.

*******

NEW TOPIC: Short status updates

*******

GW: MAD modelling of LCLS, monitor and using multi in easyPVAJava. Monitoring archive data.

DH: Working on performance with matej - performance of monitor used to be faster. DH and MS diagnosing why performance appeared to have degraded since some time in past. Think they have found answer.

DH/MK/MS currently in discussion on how to do copy properly to support sharing in an optimal way.

MS: working on pvAccess performance. Will go back to multicast, esp make APIs simpler. Server side now good, but client is complicated.

MK: was doing pvDatabaseJava, now turning to easyPVAJava monitor and API.

GS: (nothing to report on V4)

MS asks GC status of CS Studio build. Basically broken.