MiaApr 16

GLAM tech / metadata folk (but mostly probably Archives folk in practice), some help with a vibe check - if I say 'records in context', what's your response?

#AI4LAM #GLAMtech #RecordsInContext #archives

Show threadDorothea SaloApr 16
@mia RIC-O ontology. Also "kill it with fire."
Show threadMiaApr 16
@dsalo 🤣
Show threadDorothea Salo

@mia oh, if you think I'm harsh, talk to @platypus about RIC-O

possibly don armor first

Apr 16 at 1:52pmWeb
Show threadDorothea SaloApr 16

@mia @platypus slightly more seriously:

standards developers tend to go about things one of two ways

one is "the simplest way that could possibly work," e.g. BagIt, PCDM, OCFL, even .epub back in the day when I worked on it

one is "chisel it in stone like Moses, only we're gonna need a LOT more stone" e.g. FGDC (gah), FRBR, RIC-O, OAIS

I have seen too many of the latter category crash and burn to trust chiseled-stone standards architecture much

Show threadHector CorreaApr 16
@dsalo @mia @platypus no lies detected :D
Show threadDorothea SaloApr 16

@hectorjcorrea @mia @platypus I mean, there can be failure modes in both directions.

OAI-PMH was trying to be "the simplest thing," it's just that van de Sompel and Lagoze completely overestimated their target user population (scientists).

In fact, "not understanding either the problem space or the target user population" may be THE COMMONEST standards failure mode, now I think about it.

OAIS has never understood the whole of the problem space.

Show threadDorothea SaloApr 16

@hectorjcorrea @mia @platypus Whereas in the early days of .epub, we had key user populations represented on the working group:

* publishing production people (me, at the time)
* ebook hardware/software designers and coders
* markup/web nerds (also me)

and without any of those groups .epub would have failed HARD.

Show threadHector CorreaApr 16
@dsalo @mia @platypus I know designing these things is hard, but it does not help that they tend to be "IT driven" and as you said with little consideration about what the community really needs.
Show threadHector CorreaApr 16
@dsalo @mia @platypus and I would bet that most of them fail because they try to do too much (rarely they fail because they do too little).
Show threadDorothea SaloApr 16
@hectorjcorrea @mia @platypus Agree. "Does too little" is also fixable in a way "overbuilt to unusability" isn't.
Show threadBen O'SteenApr 16
@dsalo @mia @platypus agreed! IMO I'd go further - chiselled in stone standards are made to appear to work only under duress and pain, while the simplest standards are those that grow in use when the technical teams are not watched too closely...
Show threadMiaApr 16
@benosteen @dsalo @platypus I think it was @thisisaaronland who said 'if you can get two people to agree how to do something, you've got a standard', way back in the day
Show threadRuth [☕️ 👩🏻‍💻📚✍🏻🧵🪡🍵]Apr 16
@dsalo @mia the graphs of wrath