[drmaa-wg] DRMAA-WG May 17, 2005 call meeting minutes

[Rajic, Hrabri]

Attendees:
	Roger Brobst
	Andreas Haas
	Hrabri Rajic
	Daniel Templeton
	Peter Troeger


A. Meeting secretary for this meeting?

	Hrabri

B. Acceptance of the May 3, 2005 meeting minutes

	Accepted w/ changes.

C. Admin
   	- status of the existing trackers


>https://forge.gridforum.org/tracker/index.php?aid=1126

Agreed to use zeros.  All agreed.


>https://forge.gridforum.org/tracker/index.php?aid=1171

Agreed to accept as proposed w/ added value.  

Agreed to make all the exceptions or errors explicit.  Except
informational interfaces all other routines should have
DRMAA_ERRNO_AUTH_FAILURE.


>https://forge.gridforum.org/tracker/index.php?aid=1172

Agreed.

>https://forge.gridforum.org/tracker/index.php?aid=1175
Agreed.

>https://forge.gridforum.org/tracker/index.php?aid=1176
Agreed.

>https://forge.gridforum.org/tracker/index.php?aid=1177-80 & 1183, 1184
Agreed to accept all of them.

>https://forge.gridforum.org/tracker/index.php?aid=1186
Agreed to accept all.

>https://forge.gridforum.org/tracker/index.php?aid=1187
Agreed to accept all.

>https://forge.gridforum.org/tracker/index.php?aid=1188
Agreed to accept all.

>https://forge.gridforum.org/tracker/index.php?aid=1189
Agreed to accept all.

>https://forge.gridforum.org/tracker/index.php?aid=1190
Agreed to accept all.

>https://forge.gridforum.org/tracker/index.php?aid=1191
Agreed to accept all.

>https://forge.gridforum.org/tracker/index.php?aid=1357
Return internal error and try to describe it well in the long error
string.


Opening multiple concurrent sessions discussion:
Single session is not a good idea for OO languages - a bit unnatural or
underpowered.   C has just one context and it looks OK there.  We will
pick up this conversation in the next call.
The initial motivation to have just one session came from lowering the
entry bar for implementations, not from C binding.


Action Items:

Peter and Dan to close the all Trackers.

We will wait for the changes from Peter in the language independent IDL
doc and go thru the C binding specs to see how well it is synchronized
after the changes.  (It is very likely that Java and C binding docs will
not be ready by GGF 14, which is OK since we are resubmitting the main
doc.)