Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    Stefano Bocconi
    @sbocconi
    Hello @fosterlynn and @bhaugen , thanks for your replies. @bhaugen I agree that the operation is typical, I used an example from one of our pilots but the goal is to establish a sort of library or templates if you will for the most common situations that happen in flows so that different pilots can use the same logic. I am not sure if I understand the web search point, we need to model this in Value Flows
    @fosterlynn your remark about whether the same lot gets in and out of the laundry is relevant, but for flexibility reasons I would like to model the case when the lots are different, so the laundry could make bigger or smaller lots as output.
    We are planning to use the trackingIdentifier for the gowns as this is a given to us and the way they already track the items in the system (meaning they scan the tags of the gowns). I am not sure about the lot, but I guess it is useful to have a lot tracking id there as well, as they might have some kind of recording of the transportation to the laundry since it is a different "agent"
    so in general we deal with external given tracking identifiers
    Stefano Bocconi
    @sbocconi
    @fosterlynn my first impression is that the consume produce actions for the lot make sense, as to introducing new actions such as pack unpack I am a bit hesitant as we (or to be precise another party in Reflow) have already implemented tools that would need to also implement the new actions, and I am not sure of the consequences (in terms of resources and time) there. I need to discuss this, I wonder if there could be a work-around using existing actions that we could try (without destroying the semantics).
    Lynn Foster
    @fosterlynn
    @sbocconi understood, back with you within a few hours
    Lynn Foster
    @fosterlynn

    @sbocconi - Are you decrementing and implementing the resources based on the action of the events? Assuming you are, and we are talking first about just putting the gowns into a lot in preparation for sending to the laundry, then what we want to happen is to keep the accounting quantity the same (the agent still owns the gowns) and decrement the onhand quantity of each gown because it is packed up and not available? Or maybe that happens when you send them to the laundry? I feel like I need to be more clear about what actually happens operationally here. Anyhow, the trouble with consume and produce is that the accounting quantity will be affected, so the "owned" gowns wouldn't show as "owned" in inventory or reporting while packed up and at the laundry.

    We can note that transfer-custody does what we want in general, and could be used when giving the gowns to the laundry, and then again when receiving them back. Would you be tracking in the bigger lots then? Or will they also be scanning the gowns for identifiers?

    Looking at what might work for just the packing into a container lot, and not finding anything that works very well. Really, we would want to define something that would know it will be creating and then removing containedIn relationships, to do it more modularly. But maybe that could be done in a specific packing process, maybe by ProcessSpecification?

    Also, there is always the question of how much is worth recording. There is no problem "skipping" things that aren't useful operationally, say for knowing about your inventory or actually coordination your processes, as long as the provenance is maintained as needed for your requirements. Or it is more like "burying" things into other things, not so much "skipping".

    And btw, thanks for these detailed use cases, they are helpful to us too. Is there a repo we can watch to see what you are doing? We'd also be happy to talk there about VF questions if that works better for you all. As you can see, there are pieces of VF that are pretty stable and have been implemented multiple times, and there are pieces that have not been implemented with VF vocabulary at all or enough to know they are stable for different use cases.

    Also, we'll continue to mull this over, but wanted to get something back to you sooner rather than later.
    Stefano Bocconi
    @sbocconi
    Hi @fosterlynn, we are definitely looking at containedIn and thinking of changing it for each gown. Could the process be that we produce a lot out of nothing, change all containedIn properties of the gowns, do a transfer-custody of the lot, and at destination possibly consume the lot (to produce nothing) and modify each containedIn properties of the gowns. Tricky is maybe that the lot does not have a contains property to get back to the gowns? That might be an implementation issue (to have a reversed index)
    maybe raise is then better than produce for the lot
    I wonder how the semantics is implemented, if A and B are "containedIn" C, and I transfer-custody of C to place X, is location of A and B also updated to X? Is this an implementation duty?
    Stefano Bocconi
    @sbocconi
    We can move the discussion to our github repositories, which would be nice for reference, but maybe here we have more community members?
    1 reply
    Lynn Foster
    @fosterlynn
    Hi @sbocconi , I don't see a problem with producing a lot out of nothing (in a process, right?). I'm thinking it wouldn't be exactly out of nothing, the individual gowns could come into the process and get their "containedIn" added or changed. Maybe each gown resource could come in as "accept", and go out as "modify". And the new lot resource output as a "produce".
    Raise is more for beginning balances when a system is installed, or for correcting inventory when the actual count doesn't match the computer - basically for when there isn't some known natural event to increment the resource.
    Thinking about at the laundry....
    Lynn Foster
    @fosterlynn

    I wonder how the semantics is implemented, if A and B are "containedIn" C, and I transfer-custody of C to place X, is location of A and B also updated to X? Is this an implementation duty?

    New territory. Thinking out loud, you would I assume want to report out the correct location of a particular gown if asked - so either the software would need to know to look at the container location to give that info, or you would have to update each gown when the container changes location. I'm thinking that changing each gown is safer?

    1 reply
    Robin Vobruba
    @hoijui:fabcity.hamburg
    [m]
    .. because it is possible for containedIn not to be of physical nature like members of a team, instead of gowns containedIn a packet)?
    Lynn Foster
    @fosterlynn

    .. because it is possible for containedIn not to be of physical nature like members of a team, instead of gowns containedIn a packet)?

    Hi @hoijui:fabcity.hamburg ! Good question. ContainedIn is a property of EconomicResource, which we try to think of as "not people" as a general principle. For transporting people, it would be interesting to detail out the semantics where location matters more than what we have, which is the primaryLocation, for a use case.

    Lynn Foster
    @fosterlynn

    do a transfer-custody of the lot, and at destination possibly consume the lot (to produce nothing) and modify each containedIn properties of the gowns. Tricky is maybe that the lot does not have a contains property to get back to the gowns? That might be an implementation issue (to have a reversed index)

    Do you want to include a transportation process for the lot to the laundry? And do you maybe have access to a laundry to see what their operational processes might be? Or will your modeling exercise not include what happens at the laundry? I'm hearing that coming back from the laundry, the gowns might be in a different lot than they left in? Which seems fine, when you do an unpack process maybe they could lose that containedIn lot?

    @sbocconi apologies, back in an hour....
    Lynn Foster
    @fosterlynn

    at destination possibly consume the lot (to produce nothing) and modify each containedIn properties of the gowns.

    Thinking about this... and wondering what should be the most clear standard logic to consuming a container resource. It seems to me that in general consuming the lot might be expected to consume all the containedIn resources, thinking about use cases that actually do want to do that without detailing out everything in the container. Same for "move" and maybe other actions. But I understand for your use case. Still thinking about what existing alternative could possibly work better, if any....

    Lynn Foster
    @fosterlynn
    Started an issue here https://lab.allmende.io/valueflows/valueflows/-/issues/664, but I expect that to result in new actions.
    Lynn Foster
    @fosterlynn
    Regarding new actions, btw, it might make a difference if the current Reflow code uses the properties of the action to drive the behavior, for example incrementing/decrementing resources. If they are not referencing the action label, maybe adding new actions would be fairly seamless?
    1 reply
    Robin Vobruba
    @hoijui:fabcity.hamburg
    [m]
    (side node: maybe it makes sense to have an RDF property to link similar properties together? they could at the same time explain the differences)
    4 replies
    .. though it could also just be put into the comment (aka description)
    Bob Haugen
    @bhaugen
    @hoijui:fabcity.hamburg what code repo(s) is fabcity.amsterdam using?
    5 replies
    Robin Vobruba
    @hoijui:fabcity.hamburg
    [m]
    did you find an interesting project of them, something particular?
    Robin Vobruba
    @hoijui:fabcity.hamburg
    [m]
    My part in it. is to develop an assortment of tiny tools, like SVG/XML cleanup for git, RDF/Turtle cleanup, small KiCad related tools, ...
    so... I think there is really nothing to see as of now. I imagine, most interesting for you, would be what dyne does within the project.
    many others will interface with it. but they might be the ones you should go to for looking around.
    Adam, Jaromil and Henry, I think.
    Bob Haugen
    @bhaugen
    Thanks again. We're getting connected better in those matrix chats...
    Robin Vobruba
    @hoijui:fabcity.hamburg
    [m]
    so maybe to give a very broad overview:
    FabCity Hamburg was founded.. now, basically (a few months ago/last year).
    it won a big pot of funding money (big for me at least), which will be used, to a big part, to build its digital infrastructure - basically that which you/learndeep want to use - over the course of the next ~2 years.
    so.. that software will start to be .. usable/production ready/something like that, in about 2 years from now (at least in my simple mind).
    I am not sure how accurate that is, and the others in the project would surely say things more ... carefully.
    Lynn Foster
    @fosterlynn
    Thanks, and yeah we assume it will be awhile. :)
    Bob Haugen
    @bhaugen

    I am not sure how accurate that is, and the others in the project would surely say things more ... carefully.

    Don't worry about it. We have worked out how LearnDeep should start with some help from the Bonfire project team. They might or might not connect with FabCity later, so your time estimates seem good in either case.

    The LearnDeeps will start by developing their own software which will be compatible if FabCity continues to use Valueflows and a fork of Bonfire.
    Robin Vobruba
    @hoijui:fabcity.hamburg
    [m]
    ahhh great! :-)
    Jeremy Boom
    @jeremyboom8_twitter

    And please visit the gitter chat if you have difficulty accessing the gitlab repositories!
    There have been closures due to spam that may affect your ability to register.

    hi! Lynn and Bob please could one of you try to help me to sign into gitlab? it's saying i don't have a gitlab account when i do 😅

    3 replies

    FabCity Hamburg was founded.. now, basically (a few months ago/last year).
    it won a big pot of funding money (big for me at least)

    awesome to see you here @hoijui:fabcity.hamburg. congrats on the funding!

    happysalada
    @happysalada:matrix.org
    [m]
    I'm not saying you should make the switch, but if it ever turns out to be a problem, I run my own gitea instance (which is gitlab but lighter). Just saying if you ever need an alternative
    I don't know how to migrate issues though
    3 replies
    Robin Vobruba
    @hoijui:fabcity.hamburg
    [m]
    It is a distributed issue management tool, tha has import and export capabilities for the big platforms.
    .. aww just saw that it does not support gitea yet (according ot the README) :/
    Robin Vobruba
    @hoijui:fabcity.hamburg
    [m]
    :point_up: Edit: It is a distributed issue management tool, that has import and export capabilities for the big platforms.
    .. aww just saw that it does not support gitea yet (according to the README) :/
    happysalada
    @happysalada:matrix.org
    [m]
    Interesting tool, thank you for sharing ! I’ll keep an eye out for releases
    Lynn Foster
    @fosterlynn
    :+1:
    Lynn Foster
    @fosterlynn
    @sbocconi I just wanted to let you know that we did add "pack" and "unpack" to the vocab, thanks very much for the discussion! I know you are hesitant to use something new because of the code already written for your project. But just in case....
    https://valueflo.ws/introduction/flows.html#actions
    https://valueflo.ws/examples/ex-production.html#pack-unpack
    Stefano Bocconi
    @sbocconi
    Hi @fosterlynn, thanks for the update! We had suspended the work for some time due to holidays, we should be resuming this week and I will come back to this discussion group for any significant update