Where communities thrive

  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
    Joshua MacDonald


    I'll start this discussion with a bold position. The term "resource" should be avoided entirely.
    In English usage outside of modern computing platforms, the term is more commonly used in plural form.
    What is a single "resource"? I believe "static scope" best describes this concept in the way we are using it.

    In the spec meeting today, there was a question about how this relates to distributed context--I would call that dynamic scope.
    There's a connection with programming languages, which use "lexical scope" and "dynamic scope" to describe these two kinds of scope.

    If a "resource" can be a single cluster or a single rack or a single computer, then conceptually we shouldn't stop generalizing at the single process.
    A resource can be a single shard, a single peer, a single vhost, ...

    Tristan Sloughter
    I don't like the term "scope" for them if that is what you mean. I sort of like "resource", but would be fine with a change, I just think it should be something closer to "resource", like "component" or "service"
    Joshua MacDonald
    I've promised to produce some pseudo-code to explain this idea in more depth, but I am thinking of "Scope" as referring to the container or context that helps locate current resources. Resources would be ordinary key:value pairs, but they would not be required to be set once per process, they would be contextual. <--- not sure this helps. I'll write some pseudo-code.
    Tristan Sloughter
    ah, think I get what you mean
    Ted Young
    Thanks @jmacd. I’d like to start with some basic examples and use cases. What data are we putting in here, how is it used, etc. The more concrete the better.