Francis Avila

Clojure(script) developer. @datomic_team power-user. Coauthor of Clojure in Action 2nd Ed.

[Francis Avila @favila]

"As for me and my house, we will coerce the predicate function's value to boolean. And write 'else returns false' in the docstring."

[Francis Avila @favila]

I dug into Datomic's datasource internals seeking enlightenment and wrote some toy custom datasources along the way. https://t.co/It851sYKFG

[Datomic @datomic_team]

Jepsen Datomic Pro 1.0.7075

https://t.co/UlJ2G0XanU

#clojure #datomic

[Francis Avila @favila]

I wrote an article with everything I know about Datomic Entity Ids. There's more to them than you'd think. https://t.co/WwCA5RRDoj

[Francis Avila @favila]

@cemerick Why only a part?

[Francis Avila @favila]

Wrote an article about some things to watch out for when using a unique composite attribute in Datomic https://t.co/SXvj3IlJAj

[Francis Avila @favila]

@cemerick Your time went foom

[Francis Avila @favila]

Made a blog, wrote a first post about #datomic and a narrow question of schema design. HTH!
https://t.co/JO2DZRcw23

[Francis Avila @favila]

@MaxCountryman @aboodman IIRC motivation was that the magic properties in terrible dom and browser apis could not be implemented in js without a feature like proxies. The desire for immutable data structures became mainstream much more recently (post-React).

[Francis Avila @favila]

@cemerick I have the same itch. I am planning to try OpenSearch “trace analytics” in a docker container and sending it opentelemetry from my app.

[Francis Avila @favila]

@quoll @borkdude You might consider the ArraySeq. It’s a seq that wraps an object array and reuses it for subseqs. It’s the thing returned by sort.

[Francis Avila @favila]

@quoll Is this even possible on any of the hosts Clojure runs on? It could emulate it of course, but eg js or the Jvm AFAIK can’t return multiple values without some sort of box.

[Francis Avila @favila]

@mpenet @borkdude @ryrobes Yes, this is what 1.12 will do for empty? CLJ-1872 it won’t call seq if x is Counted

[Francis Avila @favila]

@nikitonsky @ryrobes There is not-empty, which is close, and also useful with if-some etc

[Francis Avila @favila]

@ryrobes @nikitonsky So not only will *more types* support empty? than seq (eg transients), but it will *also* be faster. What reason will be left to use seq CLJ-1872

[Francis Avila @favila]

@ryrobes @nikitonsky Clojure 1.12 is getting a fast-path in empty? for Counted objects that avoids the seq allocation. I hope this dumb seq idiom goes away after that lands

[Geoffrey Litt @geoffreylitt]

Big news in web app persistence: SQLite + Chrome are collaborating to create an official SQLite WASM build, backed by performant filesystem APIs!

I'm mostly very excited for this. But there is one catch that I'm still trying to understand... 1/

[Francis Avila @favila]

@cemerick When I run Jvm guis with this jre they instantly look 100x better https://t.co/PNiLXgcsAq

[Francis Avila @favila]

@cemerick Maybe still too kiddie, but my 8yo likes “Human Body!” from Smithsonian. It has big, detailed illustrations

[Francis Avila @favila]

@borkdude Bit me again too! https://t.co/YD1mVijxKb

[Francis Avila @favila]

@ericnormand But another interesting q: why is the hardware always the "bottom" and the software the "top"? Is this spacial metaphor consistent across disciplines and cultures? 4/3

[Francis Avila @favila]

@ericnormand Just listened to your "abstract at top or bottom" podcast. I think the paradox resolved by definition of "abstract": it means lit. "take away", so we need to ask "what are we taking away?" when making "abstraction" or "more abstract". 1/...

[Francis Avila @favila]

@ericnormand BTW, my "default view" has always been "top is more abstract," but is probably the InfoSys bias of my career. Engs probably to bias towards "the thing we make" is the more-abstract thing, because they put effort into removing a concern from their own problem to gain power 3/3