IETF
ASDF
asdf@jabber.ietf.org
Thursday, March 11, 2021< ^ >
cabo has set the subject to: https://codimd.ietf.org/notes-ietf-110-asdf
Room Configuration
Room Occupants

GMT+0
[11:39:45] cabo has set the subject to: https://codimd.ietf.org/notes-ietf-110-asdf
[11:50:03] Niklas Widell_web_925 joins the room
[11:50:03] Michael Richardson_web_674 joins the room
[11:50:03] Oh Seokbeom_web_437 joins the room
[11:50:03] Alessandro Amirante_web_990 joins the room
[11:50:03] Yoshifumi Atarashi_web_105 joins the room
[11:50:17] alex-meetecho joins the room
[11:51:00] Francesca Palombini_web_639 joins the room
[11:53:25] francesca joins the room
[11:54:55] <Niklas Widell_web_925> test
[11:55:03] mcr leaves the room
[11:55:07] mcr joins the room
[11:55:39] <Michael Richardson_web_674> hello.
[11:57:06] Ari Keränen_web_368 joins the room
[11:57:09] Nigel Davis_web_166 joins the room
[11:57:19] Ari Keränen_web_368 leaves the room
[11:57:24] Ari Keränen_web_262 joins the room
[11:57:48] Andreas Ruest_web_612 joins the room
[11:58:07] Alex Huang Feng_web_169 joins the room
[11:58:53] <mcr> hello. This is the co-chair. If you want something taken to the mic, please prefix with mic:
[11:58:59] Alan Soloway_web_156 joins the room
[11:59:02] <mcr> notes are at: https://codimd.ietf.org/notes-ietf-110-asdf?edit <https://codimd.ietf.org/notes-ietf-110-asdf?edit>
[11:59:27] Michael Breuer_web_985 joins the room
[11:59:41] Carsten Bormann_web_726 joins the room
[11:59:57] Wei Pan_web_448 joins the room
[11:59:58] Trent Adams_web_773 joins the room
[12:00:13] Michael Koster_web_476 joins the room
[12:01:07] Dave Robin_web_913 joins the room
[12:01:24] Stan Srednyak_web_101 joins the room
[12:01:52] Kenneth Murchison_web_641 joins the room
[12:02:48] Wei Pan_web_448 leaves the room
[12:03:01] Andre Bondi_web_557 joins the room
[12:03:56] Christian Amsüss_web_732 joins the room
[12:04:22] Steve Donovan_web_780 joins the room
[12:04:39] Toerless Eckert_web_786 joins the room
[12:05:46] Michael McCool_web_673 joins the room
[12:08:00] Dave Robin_web_913 leaves the room
[12:08:03] Dave Robin_web_900 joins the room
[12:09:09] Dave Robin_web_900 leaves the room
[12:09:12] Dave Robin_web_721 joins the room
[12:10:59] Juan-Carlos Zúñiga_web_885 joins the room
[12:13:27] Toerless Eckert_web_786 leaves the room
[12:22:48] <mcr> CVS $Tags$
[12:23:14] Lixia Zhang_web_233 joins the room
[12:24:00] Lixia Zhang_web_233 leaves the room
[12:27:16] <Carsten Bormann_web_726> https://github.com/ietf-wg-asdf/SDF/issues/26
[12:27:39] <Carsten Bormann_web_726> CVS tags are interesting, because we learned that we don't want to have them
[12:29:04] Joris Baum_web_880 joins the room
[12:29:35] Joris Baum_web_880 leaves the room
[12:29:48] <mcr> Yeah, CVS $Tags$ were a disaster and git had to get rid of them.  For others, they were a way to embed history and other info into a file that was managed in the 1990s Concurrent Version System (CVS). Actually, they go back to "RCS" and SCCS.  But, some tags could be used to establish this kind of contributor history into the file. (but, three-way commit).
[12:33:14] <cabo> Poll?
[12:35:12] Umberto Fattore_web_285 joins the room
[12:37:29] Nigel Davis_web_166 leaves the room
[12:37:32] Nigel Davis_web_289 joins the room
[12:39:58] Oh Seokbeom_web_437 leaves the room
[12:44:10] Andre Bondi_web_557 leaves the room
[12:44:15] Andre Bondi_web_307 joins the room
[12:45:38] <mcr> Additional Units for Sensor Measurement Lists (SenML)
RFC8798
[12:51:10] Julien Maisonneuve_web_601 joins the room
[12:51:13] Julien Maisonneuve_web_601 leaves the room
[12:51:16] Julien Maisonneuve_web_191 joins the room
[12:52:33] Julien Maisonneuve_web_191 leaves the room
[12:55:32] Julien Maisonneuve_web_904 joins the room
[12:56:20] Oh Seokbeom_web_484 joins the room
[12:58:26] Alan Soloway_web_156 leaves the room
[12:58:30] Alan Soloway_web_929 joins the room
[12:59:10] <Michael McCool_web_673> btw in WoT we also have "safe" and "idempotent" properties for actions, following with the concept that these are functions.
[12:59:29] <Michael McCool_web_673> s/properties/attributes/
[13:00:28] <Michael McCool_web_673> for events, we also have multiple data models and interactions for subscribe, unsubscribe, and of course the notification itself
[13:00:48] <Michael McCool_web_673> although one could also use an action to subscribe/unsubscribe
[13:03:18] <Andre Bondi_web_307> How does one distinguish between "readable" and "observable" ?
[13:03:38] <Andre Bondi_web_307> Question answered.
[13:04:12] Emmanuel Baccelli_web_258 joins the room
[13:04:57] Alessandro Amirante_web_990 leaves the room
[13:05:03] Alessandro Amirante_web_532 joins the room
[13:05:12] <mcr> I lost cabo audio.
[13:05:19] Juan-Carlos Zúñiga_web_885 leaves the room
[13:05:25] Juan-Carlos Zúñiga_web_142 joins the room
[13:06:01] Trent Adams_web_773 leaves the room
[13:06:07] Trent Adams_web_218 joins the room
[13:06:17] <Michael McCool_web_673> for events, pls clarify how subscribe/unsubscribe are handled.  Actions?
[13:06:19] <mcr> is it me? shall I reload?
[13:06:22] Kenneth Murchison_web_641 leaves the room
[13:06:33] <mcr> reloading. sorry.
[13:06:36] Michael Richardson_web_674 leaves the room
[13:06:39] <Niklas Widell_web_925> mcr: not sure, I think it was Carsten
[13:06:43] Michael Richardson_web_435 joins the room
[13:07:04] Wouter van der Beek_web_766 joins the room
[13:10:56] Nigel Davis_web_289 leaves the room
[13:11:02] <Alan Soloway_web_929> printer queue?
[13:11:16] <Michael McCool_web_673> i think a printer queue is a good example
[13:11:20] <Alan Soloway_web_929> serialized, but all "active"
[13:11:27] <Michael McCool_web_673> righgt
[13:11:52] Qin Wu_web_480 joins the room
[13:11:57] <Michael McCool_web_673> anyway, REST apis often have "creatable" resources
[13:11:58] <mcr> airport jetway ?
[13:11:59] Nigel Davis_web_722 joins the room
[13:12:52] <Michael McCool_web_673> in the wot directory service we also ran into this, since you create new resources when you register entries
[13:14:05] <Nigel Davis_web_722> Action appears to be what is called a task or a job in other groups. The task can be long lived or fleeting. In the case of a long lived task then it could be paused, resumed etc. (you can abort a missile btw)
[13:14:42] <Michael McCool_web_673> in this case you are essentially creating new properties, but they follow a template, and can be handled in other ways... not all resources need to be directly in the data model, either
[13:16:56] Qin Wu_web_480 leaves the room
[13:17:16] <Michael McCool_web_673> btw the reason we have "safe" attributes in WoT is that you might have "actions" that are purely computational and do not have real-world side effects.  An example would be a AI inferencing job (RPC) on an edge computer service.
[13:18:04] <Michael McCool_web_673> we can argue about whether that is IoT but you do want to be able to easily integrate such things into an iot system
[13:20:44] Georg Murzik_web_661 joins the room
[13:21:06] Georg Murzik_web_661 leaves the room
[13:23:20] <cabo> Now I need the big "thumbs up" button (things with mappings)
[13:23:56] <Niklas Widell_web_925> you can do emojis here
[13:28:27] <cabo> MMC: sdfFunction...
[13:29:17] Eve Schooler_web_739 joins the room
[13:29:23] <cabo> Mapping files *should* be a separate spec
[13:29:38] <Niklas Widell_web_925> yes, I think that makes sense
[13:32:25] <cabo> But core SDF needs to be ready for it (see the scaleM..imum discussion)
[13:34:21] <Michael McCool_web_673> cabo: thanks, will look at sdfFunction
[13:34:40] <cabo> I just invented that :-)
[13:35:33] <Wouter van der Beek_web_766> sdfunction == sdfaction
[13:35:44] <Wouter van der Beek_web_766> so why do we need another construct?
[13:36:02] <Michael McCool_web_673> I thought I had overlooked it ;).  Anyway, actions are already doing things that functions need (input and output data schemas, etc) so I think some additional attribute is the way to go here.  These let you know whether it's ok e.g. to retry an action
[13:36:08] Stan Srednyak_web_101 leaves the room
[13:36:13] Stan Srednyak_web_581 joins the room
[13:36:36] <cabo> Wouter: We don't.  This could just be a quality of an sdfAction.
[13:37:11] <Wouter van der Beek_web_766> about OCF using mapping files, I am not sure if OCF will use that, since OCF is already using OAS files that very explicit in what is on the wire.
[13:37:13] <cabo> W15
[13:37:29] Qin Wu_web_787 joins the room
[13:37:41] <Michael McCool_web_673> also if an action is safe you know it won't e.g. affect other properties and invalidate your cache of them
[13:37:43] Oh Seokbeom_web_484 leaves the room
[13:37:45] <cabo> There is a standard for that :-)
[13:37:52] Qin Wu_web_787 leaves the room
[13:37:53] <Wouter van der Beek_web_766> note: i am not against mapping files, it may be usefull in other situations.
[13:38:38] <Michael McCool_web_673> anyhow, I guess I should properly create issues on some of the above against SDF...
[13:40:03] <Michael Koster_web_476> @wouter, yes, there are other ways that don't require mapping files, but if it's exposed and standardized it makes it really easy to implement an OCF endpoint using something like node-wot
[13:41:19] <Wouter van der Beek_web_766> well i can see that an action should have qualities to indicate an transaction, e.g. failure reverts everything.
[13:42:18] <cabo> cross-links
[13:42:38] <cabo> See core-dynlink
[13:43:09] <Wouter van der Beek_web_766> with relationships please, so that one can define state
[13:48:39] Michael Koster_web_476 leaves the room
[13:48:44] Michael Koster_web_123 joins the room
[13:50:27] <Michael McCool_web_673> right, if an action does affect state, it would be nice to know what it changes, so only those need to be invalidated in a cache
[13:51:15] <Michael McCool_web_673> I hope this chat is being saved so I can look at it later when writing up an issue
[13:51:34] <mcr> the chat is logged, yes.
[13:51:47] <Michael McCool_web_673> great
[13:52:10] <mcr> I can't recall where exactly, but ask on the list.
[13:53:11] Michael Koster_web_123 leaves the room
[13:53:16] Michael Koster_web_307 joins the room
[13:53:27] Oh Seokbeom_web_214 joins the room
[13:53:58] Paolo Saviano_web_411 joins the room
[13:56:24] <Wouter van der Beek_web_766> well UPnP had a specific naming convention for arguments and state variables.
[13:57:14] <Wouter van der Beek_web_766> A_ARGS prefix is used to indicate an function argument, without that prefix the variable in the argument list is being manipulated.
[13:57:28] <cabo> https://www.ietf.org/jabber/logs/asdf/2021-03-11.html
[13:57:41] <Wouter van der Beek_web_766> sdf or the org using sdf could use a simular scheme
[13:59:05] <Wouter van der Beek_web_766> onedm.org?
[13:59:06] Alan Soloway_web_929 leaves the room
[13:59:43] Julien Maisonneuve_web_904 leaves the room
[13:59:43] <Michael McCool_web_673> thx
[13:59:46] <francesca> thank you! bye
[14:00:02] Francesca Palombini_web_639 leaves the room
[14:00:08] Umberto Fattore_web_285 leaves the room
[14:00:09] francesca leaves the room
[14:00:12] Alex Huang Feng_web_169 leaves the room
[14:00:14] Dave Robin_web_721 leaves the room
[14:00:14] Andreas Ruest_web_612 leaves the room
[14:00:14] Oh Seokbeom_web_214 leaves the room
[14:00:15] Juan-Carlos Zúñiga_web_142 leaves the room
[14:00:15] Wouter van der Beek_web_766 leaves the room
[14:00:15] Eve Schooler_web_739 leaves the room
[14:00:16] Andre Bondi_web_307 leaves the room
[14:00:17] Emmanuel Baccelli_web_258 leaves the room
[14:00:17] Michael Koster_web_307 leaves the room
[14:00:18] Michael Richardson_web_435 leaves the room
[14:00:18] Michael McCool_web_673 leaves the room
[14:00:22] Carsten Bormann_web_726 leaves the room
[14:00:22] Steve Donovan_web_780 leaves the room
[14:00:25] Niklas Widell_web_925 leaves the room
[14:00:32] Ari Keränen_web_262 leaves the room
[14:00:55] Stan Srednyak_web_581 leaves the room
[14:01:25] Yoshifumi Atarashi_web_105 leaves the room
[14:01:25] Michael Breuer_web_985 leaves the room
[14:01:25] Christian Amsüss_web_732 leaves the room
[14:01:25] Alessandro Amirante_web_532 leaves the room
[14:01:25] Trent Adams_web_218 leaves the room
[14:01:25] Paolo Saviano_web_411 leaves the room
[14:01:25] Nigel Davis_web_722 leaves the room
[14:14:24] alex-meetecho leaves the room
Powered by ejabberd - robust, scalable and extensible XMPP server Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!