IRC channel logs

<jar286>I'm uncomfortable saying "data model" when talking about OCapN, as many people are doing now. It's a protocol, not a design for a client. I would prefer "abstract syntax". Am I way off base in saying this?

<Zarutian_iPad>datum types?

<jar286>For me "datum" or "data" has a connotation of "inside the address space" or "inside the client/server", while "syntax" connotes what is transmitted or written down

<jar286>and it's not a protocol spec's business to talk about what goes on inside a node (client, server, endpoint etc)

<Zarutian_iPad>for me a datum is a bitpattern that has spefic semantics, like a 32 bit signed integer 2s compliment big endian

<Zarutian_iPad>regardless of where that data is an address space file or on graph paper

<jar286>So for example in a math book you would say the book's 'data model' is the choice of numbers used (integer, rational, complex etc)?

<Zarutian_iPad>yup

<jar286>When we used 'datum' in one of the Scheme reports it was tongue in cheek. The word was not in use in computer science in that way.

<Zarutian_iPad>really? I seen it in rather CS papers

<Zarutian_iPad>like those before the Von Horn paper or such

<jar286>Would be interested in seeing those... but it's sort of off topic. Any time two people are exposed to different "cultures" (maybe in this case, eras!) words can have conflicting connotation.

<Zarutian_iPad>gagnstak is the Icelandic word for datum btw.

<jar286>it's not just "data" that strikes me as inappropriate but "model" as well. "Model" implies that we are forming theory of what's going on inside a node. But again a protocol spec should be behavioral, not 'cognitive' or model-theoretic

<jar286>don't know "Von Horn paper"

<Zarutian_iPad>the paper that introduces the concept of capabilities

<Zarutian_iPad>Van Horn might be the right spelling

IRC channel logs

2023-05-19.log