@databasecultures I suggest:
@timnitGebru and @alex
@[email protected] and @[email protected] (posts infrequently but normally with links to longer articles) might be of interest
Explore local activities
@databasecultures I suggest:
@timnitGebru and @alex
@[email protected] and @[email protected] (posts infrequently but normally with links to longer articles) might be of interest
Most people either need to pick a level of trust and use a device as it is, or not use the tech. This isn't an elitism argument; I know something about software but if it comes to medical or legal matters, I'm in that same boat. Yes, we can educate people about the threat vectors they don't understand exist if and when they're relevant, but trying to make individuals responsible for the security of their mobile phones via verifying their software stack is like making me responsible for the safety of the drinking water coming out of my tap.
I'm not saying we shouldn't make things better, and I'm not saying there aren't a lot of low hanging purely technical fruit we shouldn't try and improve (I'm looking at you, most package repositories for most languages) but its worth remembering that improving things for everybody has to include everybody actually being able to use the improvements.
Watching @[email protected] post about trust in computing and being reminded just how little most of "us" (developers, open source advocates, even just anyone with a senior administrative job) understand the amount we've learned about tech over the years. No, people aren't going to check the hash of their downloaded open source binary blob. You don't because you're lazy, or you trust the mirror server you got it from. Most people don't because they don't know there is a binary blob, and you're about 3 layers of understanding they don't want and shouldn't need away from even telling them it exists.
Most people either need to pick a level of trust and use a device as it is, or not use the tech. This isn't an elitism argument; I know something about software but if it comes to medical or legal matters, I'm in that same boat. Yes, we can educate people about the threat vectors they don't understand exist if and when they're relevant, but trying to make individuals responsible for the security of their mobile phones via verifying their software stack is like making me responsible for the safety of the drinking water coming out of my tap.
Watching @[email protected] post about trust in computing and being reminded just how little most of "us" (developers, open source advocates, even just anyone with a senior administrative job) understand the amount we've learned about tech over the years. No, people aren't going to check the hash of their downloaded open source binary blob. You don't because you're lazy, or you trust the mirror server you got it from. Most people don't because they don't know there is a binary blob, and you're about 3 layers of understanding they don't want and shouldn't need away from even telling them it exists.
@Fishercat He's disreputable, and I may take them off the list soon on morality grounds, but it's Card's _Ender's Game_ and _Speaker for the Dead_.
@[email protected] @[email protected] In this time of "this is not a both sides issue, stop it" Card is a refreshing reminder of a simpler age where I could look at someone's views and go: "you know what, I disagree, and I think that your views are problematic in ways you aren't seeing, but I can respect both you and the thought you've put into them."
(this assumes that I haven't missed anything personally reprehensible about Card, in which case I'll just sigh and go back to my doom scrolling)
@mavnn that is very fair tbqh
I will reword a bit
@[email protected] thanks! I'm aware you didn't mean it that way, but... triggers be triggers, and I guessed you'd want to know 🤷♂️
@[email protected] as the son of and father of people with dysgraphia I agree with the sentiment of the post (see also learning to draw) but that 'everyone' cuts a bit. They both got hurt pretty badly by people assuming they could 'just fix that', and with my son for a while I was one of the people doing the hurting
@[email protected] @[email protected] ah, good news. I've just deployed 1.0.1, where I tried setting the COMPILE_ALL_LOCALES but then spotted it was overridden by the :prod flag regardless. My last attempt was to directly edit config/bonfire_common.exs with static values:
--- a/config/bonfire_common.exs
+++ b/config/bonfire_common.exs
@@ -4,8 +4,8 @@ default_locale = "en"
## Localisation & internationalisation
# Only compile additional locales in prod or when explicitly requested
-compile_all_locales? =
- config_env() == :prod or System.get_env("COMPILE_ALL_LOCALES") in ["true", "1"]
+compile_all_locales? = false
+# config_env() == :prod or System.get_env("COMPILE_ALL_LOCALES") in ["true", "1"]
locales = if compile_all_locales?, do: [default_locale, "fr", "es", "it"], else: [default_locale]
@@ -18,7 +18,7 @@ config :bonfire_common,
config :bonfire_common, Bonfire.Common.Localise.Cldr,
default_locale: default_locale,
# locales that will be made available on top of those for which gettext localisation files are available
- locales: locales,
+ locales: [],
providers: [
Cldr.Language,
Cldr.DateTime,
@@ -33,10 +33,10 @@ config :bonfire_common, Bonfire.Common.Localise.Cldr,
gettext: Bonfire.Common.Localise.Gettext,
extra_gettext: [Timex.Gettext],
data_dir: "./priv/cldr",
- add_fallback_locales: compile_all_locales?,
+ add_fallback_locales: false,
# precompile_number_formats: ["¤¤#,##0.##"],
# precompile_transliterations: [{:latn, :arab}, {:thai, :latn}]
- force_locale_download: Mix.env() == :prod,
+ force_locale_download: false, # Mix.env() == :prod,
generate_docs: true
@[email protected] @[email protected] I managed to get the compile to complete by shutting down everything else on the machine, and the compiled result only lists "en" as an available language, but it still seemed to want to generate 126 locales while building bonfire_common which was what made me think I was missing something.
@[email protected] In version 1.0.2-alpha.6 you should be able to set COMPILE_ALL_LOCALES=no in env
@[email protected] @[email protected] ah, good news. I've just deployed 1.0.1, where I tried setting the COMPILE_ALL_LOCALES but then spotted it was overridden by the :prod flag regardless. My last attempt was to directly edit config/bonfire_common.exs with static values:
--- a/config/bonfire_common.exs
+++ b/config/bonfire_common.exs
@@ -4,8 +4,8 @@ default_locale = "en"
## Localisation & internationalisation
# Only compile additional locales in prod or when explicitly requested
-compile_all_locales? =
- config_env() == :prod or System.get_env("COMPILE_ALL_LOCALES") in ["true", "1"]
+compile_all_locales? = false
+# config_env() == :prod or System.get_env("COMPILE_ALL_LOCALES") in ["true", "1"]
locales = if compile_all_locales?, do: [default_locale, "fr", "es", "it"], else: [default_locale]
@@ -18,7 +18,7 @@ config :bonfire_common,
config :bonfire_common, Bonfire.Common.Localise.Cldr,
default_locale: default_locale,
# locales that will be made available on top of those for which gettext localisation files are available
- locales: locales,
+ locales: [],
providers: [
Cldr.Language,
Cldr.DateTime,
@@ -33,10 +33,10 @@ config :bonfire_common, Bonfire.Common.Localise.Cldr,
gettext: Bonfire.Common.Localise.Gettext,
extra_gettext: [Timex.Gettext],
data_dir: "./priv/cldr",
- add_fallback_locales: compile_all_locales?,
+ add_fallback_locales: false,
# precompile_number_formats: ["¤¤#,##0.##"],
# precompile_transliterations: [{:latn, :arab}, {:thai, :latn}]
- force_locale_download: Mix.env() == :prod,
+ force_locale_download: false, # Mix.env() == :prod,
generate_docs: true
@[email protected] "I'm not sure that we're quite ready to tackle this area. Let's move on for now and let it come back up when you're ready. And you're less likely to pull my arms off."
@[email protected] dang. Now I have a perverse desire to write a counsellor for monsters visual novel.
"The frog seems mildly upset"
"The giant spider seems to be masking with anger"
mastodon.social/users/GeePaw... It's a painful realization that I missed two programming languages I've absolutely written code in while replying to @[email protected] here. XML and yaml. Do better, industry, do better...
@mavnn Oh, again, a very different path than I, tho god love the Spectrum forever.
@[email protected] my just about to go to university son has deep dived into 3d graphics and physics simulations as a hobby over the last few years, and it threw him a while back when I answered a straight forward question on memory allocation with 'no idea, I've never written anything in an unmanaged language'.
Aight, give me your programming language chain of development.
Use a prefix, - for just a little, + for a whole lot, and you can repeat.
For myself, the kind of reply I'm looking for, my own programming language chain, in the next post.
@[email protected] hmm. I suspect I'll miss some stuff, but:
-Spectrum BASIC
+html (first paid work ever, well before becoming a 'developer')
+sql (I was a data analyst before accidentally transitioning to coding by getting so bored of my data analyst job that I automated it with...)
-vbscript in excel
+c#
-Python
+f#
-clojure
+elm
-ruby (a language I thought I'd enjoy and really didn't 😞 )
+haskell
+typescript
-ink
-gdscript
@mavnn ex_cldr was recently re-implemented: https://elixirforum.com/t/localize-next-generation-localisation-ex-cldr-v3-0/73933 tried that one?
@[email protected] Oh thanks, I'll definitely keep an eye for the future. It looks like it won't support all the features from ex_cldr this project uses until at least its 0.3 release though, so I'd need to make much larger changes than I was hoping from the parent project to go that route at the moment.
Today I was going to do ... something?
What I actually did was update one piece of software that I accidentally misconfigured, which took down the database, which then turned out to be too old for the new version of the software (after I fixed the misconfiguration) which I then upgraded that broke the database backup scripts.
And then I washed my hair.
@[email protected] revel in your glorious clean hair, your enemies will fall in wave before you!
@[email protected] Unfortunately my hair is twenty years too late for "who can clear the largest moshpit space" competative head banging. It was fun while it lasted though.
Who is the He-Man movie for?
I watched the cartoon as a kid (and loved it) but I am 47 years old and am not compelled to see it. Maybe on a plane, I guess.
It's also a product of its time and I'm not compelled to share it with my kid.
Like, what's the market for it?
@[email protected] Such a weird franchise to reboot. Especially because stylistically it looks very similar to the last movie version, which while not high art was really good at nailing that style.
Question for anyone who's used ex_cldr and/or gettext in #elixir - I'm trying to compile something in a memory constrained environment and the gettext back end has started finding 130 locales it wants to generate on every build. I think I've configured cldr to have only one allowed language, but the locales still get generated (and cause an out of memory error)
Anyone know the next place to look for what might be requesting these, or force them to be generated in turn rather than (it appears) concurrently?
Context: I don't know elixir, but for reasons I'm running my own @[email protected] instance on bare metal and I'm happy to mess in source code if needed.
Question for anyone who's used ex_cldr and/or gettext in #elixir - I'm trying to compile something in a memory constrained environment and the gettext back end has started finding 130 locales it wants to generate on every build. I think I've configured cldr to have only one allowed language, but the locales still get generated (and cause an out of memory error)