Public comments are a very important part of the OGF document approval process. Through public comments, documents are given scrutiny by people with a wide range of expertise and interests. Ideally, a OGF document will be self-contained, relying only on the other documents and standards it cites to be clear and useful. Public comments of any type are welcomed, from small editorial comments to broader comments about the scope or merit of the proposed document. The simple act of reading a document and providing a public comment that you read it and found it suitable for publication is very useful, and provides valuable feedback to the document authors.
Thank you for making public comments on this document!
Comments for Document: Distributed Resource Management Application API Version 2 (DRMAA)
|Author(s):||P. Tröger, R. Brobst, D. Gruber, M. Mamon Ì?ski, D. Templeton|
|Public Comment End:||19 Nov, 2011|Posted by: troeger 2011-09-21 05:01:42"slots" attribute in SlotInfo
The "slots" attribute in SlotInfo currently has a string type. This is a typo, it should be defined as numerical type.
Posted by: troeger 2011-09-21 05:25:34Attribute naming
Some of the attribute names are chosen in a way that they repeat the scoping of the surrounding data structure ("Job::jobId" , "Reservation::reservationId", "JobSession::sessionName"). This was noted as a minor issue for some of the language bindings.
So far, the spec favors consistent attribute names, since attributes such as "jobId" are also used in other structures.
Posted by: merzky 2011-09-21 08:19:14exception extensibility
The idl spec is silent about implementation specific exceptions - this should either be explicitly allowed, or explicitly forbidden.
Posted by: merzky 2011-09-28 04:33:10CPU Speed?
It seems not to be possible to specify CPU speed on resource requests
- is that intentional left out or is that an oversight?
Posted by: troeger 2011-09-29 01:28:48Behavior on session closing
Our discussion about the C language binding at OGF33 identified the following issue:
If a job / reservation session is explicitly closed by the application, the already existing job / reservation instances should be in a defined state. The IDL spec makes no clear statement here.
The 'InvalidSessionException' implies that such instances should be invalidated automatically. However, the spec also allows session-independent instances to exist, for example from monitoring calls.
Posted by: troeger 2011-09-29 01:30:11Scalar attributes
"Access to scalar attributes (string, boolean, long) MUST operate in a pass-by-value mode."
This is a very strict requirement, which might not be fulfillable by languages such as C.
Posted by: loewis 2011-11-04 06:24:12Section 3 issue: ZERO_TIME etc. is a value, not a type
Section 3 declares ZERO_TIME, INFINITE_TIME, and NOW to be native types (along with all the other types defined there). However, the narrative text explains that they are actually values, or either TimeAmount or AbsoluteTime. Using IDL const declarations would be appropriate, except that one has to specify a value for these also.
I propose to use the following definitions
const TimeAmount ZERO_TIME = 0;
const TimeAmount INFINITE_TIME = 0x7fffffffffffffff;
const AbsoluteTime NOW = "now";
Posted by: loewis 2011-11-07 04:02:125.4: slots should be a long value
5.4.2 describes slots as "the number of slots reserved on the given machine", so it is clearly meant to be an integer - yet it is declared as a string.
Posted by: loewis 2011-11-07 04:37:126: SessionManagementException is undocumented
In the elaboration below the IDL definitions, SessionManagementException is not explained.
Posted by: loewis 2011-11-07 04:43:178.4: Raised exceptions not declared
The text explains that certain operations may raise predefined exceptions, yet the IDL signatures fail to declare these exception. My proposal is to add appropriate raises()-clauses, such as
void suspend() raises(InvalidStateException);
Posted by: loewis 2011-11-07 04:47:3413: Line break missing
Daniel Gruber's address is missing a line break before the Postleitzahl.
Posted by: henyen 2012-12-07 03:57:37Rolex Watches
The Rolex replica watches
are revered since the prosperity of this constitution of parade. The for show uses no activity unless known is a adjust network label (joyous numbers). The Phosphor E-Ink field of fake watches
are suave further modern. I regard the pitch-dark polyurethane strap is the surpassingly "stealth-like" of the three strap choices (unmarred or leather eclipse orange accents are the peculiar two).