Also supplementary: system landscape, deployment.

Internal to some business unit. External systems may still be part of the same enterprise.

Could this be represented in XMI?

I wonder what it would be like to model this in Fibery.

Except if e.g. a component is used in different containers.

C4 understands something, but eschews knowledge of e.g. the exact same component being used within different containers.

You also deploy instances of them in specific environments.

Is it possible to model the intermediary, such that it shows upon a deployment diagram, but leave it out of this diagram?

Also enterprise to show external/internal?

Would be nice to track elements across those boundaries. I.e. this system in C4 is that business unit in ArchiMate. Kind of like context propagation in distibuted tracing.

No: include the entire system in a single model. Diagrams should be focused views.

How do I do that in the DSL?

These teams, however, also want models. They see inside of those software systems. This should be just another view. On one model that contains elements for everything in the system.

The OpenTelemetry Collector is a single binary but can be deployed in different roles in order to scale horizontally

How about internal architecture?

It occurs to me this is the second coming of https://xspf.org/ (web playlists).

In Python: https://hypothesis.readthedocs.io/en/latest/

As far as I can see, this isn't integrated with the current annotation client, at least not the bookmarklet.

Announced: May 8, 2015 Completed: July 10, 2015 Recipient: Ilya Kreymer Related Project(s): Hypothes.is

http://anno.fund/#portfolioModal2

Follow-up post: https://web.hypothes.is/blog/fund-on-demand-web-archiving-completion/

The video refers to the "Get Bookmarklet" button on the front page.<br> This has moved to the "Hypothesis Bookmarklet" button on the https://web.hypothes.is/start/ page.

Just what I was thinking ;-)

So hard to teach anything if you don't know about the person you imagine you're teaching.

Yeah but they don't snap together like lego. Compare this to William Burroughs's cut-up technique. Our brain makes juxtaposed prose snippets make sense in spite of absent intention. This is the danger of "remixing" modules without embedding them in a context that defines terms, contrasts shifting usages, fills in historical background, etc.

Well, because some of them are wrong and/or lack the nuances or information that you're trying to teach. I.e. you still need to contextualise your sources, which is what that chapter was probably doing.

I don't think this has been orthodoxy since at least Thomas Kuhn.

This seems like a red herring: the requirement is not that the entire study be reproduced, but that the data resulting from the study available, so that the conclusions based on parsing this data can be verified.

That's hardly a problem with reproducibility, though.

Do we do this?

Progress!

(To say nothing of a bookmarklet highjacking my text selection .. what could possibly go wrong?)