Skip to content
Nick Ruest edited this page Feb 8, 2017 · 13 revisions

Time/Place

This meeting is a hybrid teleconference and IRC chat. Anyone is welcome to join. Here is the info:

Attendees

  • Nick Ruest
  • Bryan Brown
  • Diego Pino
  • Jared Whiklo ⭐
  • Amanda Lehman
  • Natkeeran Ledchumykanthan
  • Danny Lamb
  • John Yobb
  • Kim Pham
  • Melissa Anez
  • Martin Dow
  • Jonathan Green

Agenda

  1. Discuss the relationship between bundles and JSON-LD contexts and the limitations that Drupal puts on it.
  2. CLAW mentorship meeting at Islandoracon?
  3. Vector Clock or Version Vector, open question here https://github.com/Islandora-CLAW/CLAW/issues/526
  4. ... (feel free to add agenda items)

Minutes

irc log

  1. Discuss relationship...

    Adam not in attendance. Perhaps next week. 🤷‍♂️

  2. IslandoraCon CLAW Mentor Meeting

    Melissa: Already mulling over having a CLAW meeting on the Friday. What would this group like to see at this "meeting". Get together, meet (in person) and cover some topics they would like covered.

    Nick: This will be at the end of the conference, so some people may not be able to attend or be tired.

    Diego: Is the CLAW workshop part of the hackfest?

    Melissa: No, it is independent of that.

    Diego: Perhaps something around Drupal 8, mentorship during the hackfest? So its at the start of the conference instead of the end. Share what we have discovered in working with CLAW.

    Melissa: There would have to be more discussions around adding it to the hackfest.

    Nick: Be realistic about what we can do on that day. But in the upcoming months we can throw around ideas of what we have done.

    Melissa: Also allow for people with all level of skills and abilities to participate.

    Diego: Find a space during the hackfest for people that don't want to work on something and instead participate in some "training". This leaves space for people to hack around 7.x-1.x.

    Jon: U of S is thinking about "learning sessions" for the upcoming Access hackfest to allow people to participate without having to produce something.

    Melissa: Don't want to pass up the opportunity to have a whole bunch of people in one place talking about CLAW.

    EDIT TO ADD: Google Doc to brainstorm ideas

  3. Vector clocks vs Version vectors

    https://github.com/Islandora-CLAW/CLAW/issues/526

    Diego: Looking at how we are implementing consistency/sync and the parts that handle what happens when part of the system falls down.

    We are creating version vectors but we seem to be calling them vector clocks. My suggestion is if we are dealing with synchronicity of data then version vectors are the better options. This seems like what we are already implementing this, but calling them vector clocks.

    Danny: We are not implementing either yet. Parts are missing. What we really have is a lamport timestamp, this is also a holdover from the previous bi-directional sync. So it needs some changes. If we were going with full bi-directional sync then yes it should be a version vector. Right now for MVP we are using the lamport timestamp for version numbers and do "sort of" similar to eTag matching on updates.

If we are flushing from Drupal to Fedora all the time, this helps with the flow of data back to Drupal from Fedora.

Diego: One of the components of the version vector that is missing is the Agent. The use of API-X is also unclear to me. Somehow API-X will kick off some derivative process that are in some sort of registry. That created derivative needs to be added back into Drupal and Fedora. This is keyed from a predicate which is how I got into this.

Danny: That is the pipeline. API-X will provide the services. Fedora will hold your binaries and representations. If you want other kinds of content, you can have these services generate those other representations and (possibly) store them back into Fedora. The pipeline will ensure that these services will be run in order for each process. So you have asynchronous between processes, but each process is a serial list of processes.

Diego: I need to run two derivatives, the first creates a TN and the second creates a smaller version of the file. Then it will report back.

Danny: Each piece will report back when they are complete.

One of the big things is we are not passing state along with the JMS messages. I get the resource with a timestamp vector, do my changes and update with our timestamp. If the state has changed, then the process will fail and re-start. But part of the restart is to grab a new copy of the original content to derive from.

Nick: What would be the resolution of the ticket?

Diego: I think we need to not name it a vector clock if it is not actually a vector clock. Instead we should use the appropriate name for whatever we are using as our version tracking "thing".

Consensus is "version counter"

  1. Newbie tickets

    There are tickets available to try stuff out, grab one and people will help you.

    https://github.com/Islandora-CLAW/CLAW/issues?q=is%3Aissue+is%3Aopen+label%3Anewbie

This is an archive. For new Tech Call notes, click here

⚠️ ARCHIVED Islandora Tech Calls

⚠️ ARCHIVED Islandora User Calls

Clone this wiki locally