IRC logs for #trustable for Monday, 2019-01-28

*** brocka has joined #trustable00:03
*** brocka has quit IRC00:08
*** brocka has joined #trustable00:37
*** brocka has quit IRC00:42
*** brocka has joined #trustable02:39
*** brocka has quit IRC02:44
*** brocka has joined #trustable04:40
*** brocka has quit IRC04:44
*** brocka has joined #trustable08:29
*** brocka has quit IRC08:34
*** toscalix has joined #trustable08:41
*** sambishop has joined #trustable09:25
persiareiterative: My thought was to simply elide them and then bring them back later in the next layer of the model, wherein comparative policy is the target.  I've no issues with adding them as illustrative examples if you like, which examples might be superceded by a future change that actually includes the descriptions of appropriate policy and process.09:47
*** traveltissues has joined #trustable11:47
*** sambishop has quit IRC13:02
*** sambishop has joined #trustable13:22
paulsherwoodare we clear about how best to express 'intents' yet?13:54
persiaWe assert that they must be kept in artifacts (so we can track changes)13:57
*** sambishop has quit IRC13:57
persiaI don't believe we've gotten to any comparative definitions yet.13:57
persiaPersonally, I think it would be good to give each intent an identifer, but I don't think it reasonable to expect most definers of intent to generate and track UUIDs.13:58
reiterativeWe expect them to be defined in business / stakeholder language and don't require much of them at the moment - accepting that they may take a variety of forms, many of them not ideal13:59
persiaI'm also not quite sure how to handle the case where an intent changes: we can definitely track the change, but I'm not sure if the identifier should change and/or if we understand any practices beyond change-ID dependencies to constrain intent changes being considered actionable until they have been analysed to ensure the constraints are updated.13:59
persiaOne clean way to do that might be to have intents and constraints in the same store, such that any change to that store must be approved by rules that mean that if constraints change, analysts need to approve, and if intents change, both drivers and analysts need to agree (with the assumption that the analysts have verified that the modification to intents is reflected in a modification to constraints within the same change).14:01
*** sambishop has joined #trustable14:01
persiaNote that such a model (with intents and constraints in the same store) means an assertion that a single change is safely atomic, so there is no need for tracking how constraints map to intents formally.  For other models, it makes sense to have a mechanism to do that.14:02
persiareiterative: in terms of being able to begin to consider comparatives, where do you think we are with the merge proposal?  Do you think this is final draft (and then to the mailing list for criticism), or do we need to do another review round to define the base vocabulary?14:20
persiaI am starting to get interested in thinking about attributes of some of our terms that might be measurable (e.g. Can an intent be identified?), but don't want to start digging into that until we believe we have a sufficient and consisent base on which to start to build those arguments.14:21
reiterativeI have a pending commit that I want to push, but we're pretty close to a final draft, I think14:31
persiaCool.  Let me know when you want me to take another look, and maybe we can look at comparatives after (although I suspect we'll need input from lots of other folk there, as defining "better/worse" ends up being something that is hard to do).14:36
reiterativeNew commit pushed to pa-nomenclature14:42
*** sambishop has quit IRC14:46
*** sambishop has joined #trustable15:01
*** brocka has joined #trustable15:48
*** brocka has quit IRC15:51
*** brocka has joined #trustable15:53
*** sambishop has quit IRC16:45
*** sambishop has joined #trustable16:46
*** toscalix has quit IRC17:15
*** sambishop has quit IRC17:18
*** traveltissues has quit IRC19:05

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!