Discussion:
[core] Binding Attributes in draft-groves-core-dynlink-02
Carey, Timothy (Nokia - US)
2017-02-25 16:34:12 UTC
Permalink
Hi,

I noticed in the latest draft the binding attributes were changed.
The LWM2M uses the binding attributes in their specification that was recently published.
In LWM2M Greater than is gt; Less than is lt and Step is stp.
In the latest dynlink draft Greater than is gth; Less than is lth and Step is st

Do we know the reason for the change?
Is it possible to use lt; gt; stp as is specificed in LWM2M?

BR,
Tim
Carsten Bormann
2017-02-25 18:08:54 UTC
Permalink
(No chair hat.)
Post by Carey, Timothy (Nokia - US)
Hi,
I noticed in the latest draft the binding attributes were changed.
The LWM2M uses the binding attributes in their specification that was recently published.
In LWM2M Greater than is gt; Less than is lt and Step is stp.
In the latest dynlink draft Greater than is gth; Less than is lth and Step is st
Do we know the reason for the change?
There was a naming issue with “lt” being used for both lifetime and less than.
This is not a problem now, but could become one.
Since lifetime was in use, attempts have been made to rename less than (and greater than analogously), but discussion took some time (maybe because lt and gt are the obvious names for anyone who has heard of FORTRAN).
Post by Carey, Timothy (Nokia - US)
Is it possible to use lt; gt; stp as is specificed in LWM2M?
Maybe we have missed a window in which this change (avoiding the name collision) could have been net positive.

Grüße, Carsten

_______________________________________________
core mailing list
Michael Koster
2017-02-25 19:33:27 UTC
Permalink
Hi Tim,

Is there a name collision with "st" that required OMA LWM2M to use "stp" instead? I used "st" in the mbed reference implementation in early 2015.

Best regards,

Michael
Post by Carsten Bormann
(No chair hat.)
Post by Carey, Timothy (Nokia - US)
Hi,
I noticed in the latest draft the binding attributes were changed.
The LWM2M uses the binding attributes in their specification that was recently published.
In LWM2M Greater than is gt; Less than is lt and Step is stp.
In the latest dynlink draft Greater than is gth; Less than is lth and Step is st
Do we know the reason for the change?
There was a naming issue with “lt” being used for both lifetime and less than.
This is not a problem now, but could become one.
Since lifetime was in use, attempts have been made to rename less than (and greater than analogously), but discussion took some time (maybe because lt and gt are the obvious names for anyone who has heard of FORTRAN).
Post by Carey, Timothy (Nokia - US)
Is it possible to use lt; gt; stp as is specificed in LWM2M?
Maybe we have missed a window in which this change (avoiding the name collision) could have been net positive.
Grüße, Carsten
_______________________________________________
core mailing list
https://www.ietf.org/mailman/listinfo/core
Carey, Timothy (Nokia - US)
2017-02-26 19:17:06 UTC
Permalink
Michael,

Well the IETF interfaces draft has been consistent - the step attribute is "st".
I went back and traced the change in LWM2M from "st" to "stp".
It occurred between:
The February 2015 (OMA-TS-LightweightM2M-V1_0-20150228-D) has "stp" and the February 2015 version (OMA-TS-LightweightM2M-V1_0-20150225-D) has "st".
The agreed to cr that brought that change was OMA-DM-LightweightM2M-2015-0008R01-CR_Attributes_

I added Thierry into this email - maybe he can answer your question better since he brought the original CR but it looks to me to be an editorial problem when he restructured that part of the document and clarified the rules for use within LWM2M... Probably just a typo that needs corrected in the OMA spec...

BR,
Tim

-----Original Message-----
From: Michael Koster [mailto:***@gmail.com]
Sent: Saturday, February 25, 2017 1:33 PM
To: Carsten Bormann <***@tzi.org>
Cc: Carey, Timothy (Nokia - US) <***@nokia.com>; ***@ietf.org WG <***@ietf.org>
Subject: Re: [core] Binding Attributes in draft-groves-core-dynlink-02

Hi Tim,

Is there a name collision with "st" that required OMA LWM2M to use "stp" instead? I used "st" in the mbed reference implementation in early 2015.

Best regards,

Michael
Post by Carsten Bormann
(No chair hat.)
Post by Carey, Timothy (Nokia - US)
Hi,
I noticed in the latest draft the binding attributes were changed.
The LWM2M uses the binding attributes in their specification that was recently published.
In LWM2M Greater than is gt; Less than is lt and Step is stp.
In the latest dynlink draft Greater than is gth; Less than is lth and Step is st
Do we know the reason for the change?
There was a naming issue with “lt” being used for both lifetime and less than.
This is not a problem now, but could become one.
Since lifetime was in use, attempts have been made to rename less than (and greater than analogously), but discussion took some time (maybe because lt and gt are the obvious names for anyone who has heard of FORTRAN).
Post by Carey, Timothy (Nokia - US)
Is it possible to use lt; gt; stp as is specificed in LWM2M?
Maybe we have missed a window in which this change (avoiding the name collision) could have been net positive.
Grüße, Carsten
_______________________________________________
core mailing list
https://www.ietf.org/mailman/listinfo/core
Carey, Timothy (Nokia - US)
2017-02-26 19:18:59 UTC
Permalink
Carsten,
Probably for gt and lt; they have been used in LWM2M since it looks like around 2013.
Changing the values to gth and lth will probably affect implementations.

BR,
Tim

-----Original Message-----
From: Carsten Bormann [mailto:***@tzi.org]
Sent: Saturday, February 25, 2017 12:09 PM
To: Carey, Timothy (Nokia - US) <***@nokia.com>
Cc: ***@ietf.org WG <***@ietf.org>
Subject: Re: [core] Binding Attributes in draft-groves-core-dynlink-02

(No chair hat.)
Post by Carey, Timothy (Nokia - US)
Hi,
I noticed in the latest draft the binding attributes were changed.
The LWM2M uses the binding attributes in their specification that was recently published.
In LWM2M Greater than is gt; Less than is lt and Step is stp.
In the latest dynlink draft Greater than is gth; Less than is lth and Step is st
Do we know the reason for the change?
There was a naming issue with “lt” being used for both lifetime and less than.
This is not a problem now, but could become one.
Since lifetime was in use, attempts have been made to rename less than (and greater than analogously), but discussion took some time (maybe because lt and gt are the obvious names for anyone who has heard of FORTRAN).
Post by Carey, Timothy (Nokia - US)
Is it possible to use lt; gt; stp as is specificed in LWM2M?
Maybe we have missed a window in which this change (avoiding the name collision) could have been net positive.

Grüße, Carsten
Christian Groves
2017-02-27 03:36:21 UTC
Permalink
Hello Tim,

Given the overlap in "lt" it was either people's resource directory
implementations or binding attribute implementations that would be
impacted. It seemed that there were more resource directory
implementations and draft-ietf-core-resource-directory was further along
so it seemed the preference of the list was to change
draft-ietf-core-dynlink.

As editor i'll leave it up to the Chairs to determine which should
change. We could leave gt as it is and only change lt if it helps.

Regards, Christian
Post by Carey, Timothy (Nokia - US)
Carsten,
Probably for gt and lt; they have been used in LWM2M since it looks like around 2013.
Changing the values to gth and lth will probably affect implementations.
BR,
Tim
-----Original Message-----
Sent: Saturday, February 25, 2017 12:09 PM
Subject: Re: [core] Binding Attributes in draft-groves-core-dynlink-02
(No chair hat.)
Post by Carey, Timothy (Nokia - US)
Hi,
I noticed in the latest draft the binding attributes were changed.
The LWM2M uses the binding attributes in their specification that was recently published.
In LWM2M Greater than is gt; Less than is lt and Step is stp.
In the latest dynlink draft Greater than is gth; Less than is lth and Step is st
Do we know the reason for the change?
There was a naming issue with “lt” being used for both lifetime and less than.
This is not a problem now, but could become one.
Since lifetime was in use, attempts have been made to rename less than (and greater than analogously), but discussion took some time (maybe because lt and gt are the obvious names for anyone who has heard of FORTRAN).
Post by Carey, Timothy (Nokia - US)
Is it possible to use lt; gt; stp as is specificed in LWM2M?
Maybe we have missed a window in which this change (avoiding the name collision) could have been net positive.
Grüße, Carsten
_______________________________________________
core mailing list
https://www.ietf.org/mailman/listinfo/core
Carsten Bormann
2017-02-27 06:03:58 UTC
Permalink
As editor i'll leave it up to the Chairs to determine which should change.
The chairs can only channel the consensus of the WG, so please keep the comments on this (bikesheddy, but icky) issue coming.
We could leave gt as it is and only change lt if it helps.
(Chair hat off:) It seems to me personally that a change here hurts more than it helps, both with respect to requiring a change from the released LWM2M spec and with respect to permanently installing a POLS*) violation in the spec.

Grüße, Carsten

*) POLS: Principle of least surprise, which shouts loudly for less than to be abbreviated lt.
Loading...