geoffo, ShadowKyogre, [tantek], mblaney, treora, TeeKay, [chrisaldrich], [jeremycherfas] and KartikPrabhu joined the channel; ShadowKyogre and mblaney left the channel
schmudde, [Murray], swentel, [KevinMarks], joshproehl, Loqi, ethanyoo, gxt__, [jeremycherfas], KartikPrabhu, [Rose], geoffo, [jgmac1106] and [Raphael_Luckom] joined the channel
#[Raphael_Luckom]Anyone here like making DSLs as an approach to system building? I was thinking about the conversation yesterday about plain html / css and minimal js vs big web frameworks...one of the differences between those approaches for me is that the big frameworks have too many builtin assumptions to be useful in building DSLs (something like react is kinda a dsl already, but for a very broad domain). While the simpler tools are much more
#[Raphael_Luckom]regular building blocks for more useful, narrowly-targeted DSLs.
#[Raphael_Luckom]also gotta admit that I'm a little fuzzy on where "framework" ends and "dsl" begins...like if you have a set of known primitives (react, angular components and such) and you build your system out of those primitives, to me that fits the definition of a dsl but I'm not sure it would fit every definition of dsl.
gRegorLove, KartikPrabhu, [chrisaldrich], shoesNsocks, [schmarty] and ShadowKyogre joined the channel; ShadowKyogre left the channel
#[Raphael_Luckom]it doesn't seem like dsl vs framework is _really_ settled. A conservative definition is given here ( https://www2.dmst.aueb.gr/dds/pubs/jrnl/2000-JSS-DSLPatterns/html/dslpat.html ) -- basically requiring that a DSL have its own syntax, separate from that of the language implementing its interpreter / compiler. Martin Fowler allows a more expansive definition in a discussion of "internal DSLs" (
#[Raphael_Luckom]was looking for--you can see how something like JavaScript makes sense for implementing a DSL, but once you've tied yourself to a specific DSL like react, you no longer get much benefit from the dynamic-language features of JS. In that context a migration towards typescript, which gives up dynamic features in exchange for static analysis, starts to make sense.
#[Raphael_Luckom]https://www.martinfowler.com/articles/languageWorkbench.html#InternalDsl ), which use the syntax of a host language. Fowler seems to recognize frameworks like Rails as also being, or at least including, internal DSLs. Fowler _also_ notes an affinity between internal DSLs and _dynamic languages_, observing that dynamic languages tend to have lightweight syntax rules that make internal DSLs easier to implement. _That's_ the connection I
#[snarfed]it’s data markup, not code, but otherwise similar
#[Raphael_Luckom]HTML is commonly described as a DSL--personally I'd be fine with calling mf2 and html both dsls. I don't think either definition I found would exclude them.
#[snarfed]turing completeness is one key razor. data formats are fundamentally different than programming langs. but otherwise, yup, similar
#[Raphael_Luckom]that's true, it's a useful distinction
#[snarfed]apart from HTML/mf2, the CMSes and services we build here generally aren’t so complex internally and don’t need so much internal domain-specific programmability that they’d warrant turing complete DSLs
#aaronpk🙉 <-- me when i am reading this conversation :P
#aaronpktrying to remember if the latest changes would have affected that
ShadowKyogre joined the channel; ShadowKyogre left the channel
#gRegorLoveHi ShadowKyogre, were you able to get Bridgy Publish working?
#ShadowKyogreNot yet (for normal Bridgy). But I've made some progress with finding the tools that I'd need to fix my redirect rules (for fed.brid.gy). Found an .htaccess tester and a redirect manager module earlier this morning.
#ShadowKyogreOkay, turns out I did just need to set up the redirect manager on PW
ShadowKyogre, schmudde, gxt__, [jeremycherfas], [KevinMarks], [tantek], [Raphael_Luckom] and wagle joined the channel; ShadowKyogre left the channel
#ShadowKyogreOkay, good news. Can discover myself on Mastodon, but still can't get my stream to update properly. Publishing to twitter via bridgy is working fine now (I see a big response body in the logs).
ShadowKyogre joined the channel; ShadowKyogre left the channel