IRC channel logs


back to list of logs

<nalaginrut>afternoon guilers~
***siel_ is now known as siel
***ozzloy_ is now known as ozzloy
***darthlukan_ is now known as darthlukan
***w1gz_ is now known as w1gz
***sbp_ is now known as sbp
<amz3>I have the tuple querying notation inpired from datomic working :D
<amz3>I use global for the time being, I need to tweak minikanren to avoid the global, but so far so good :)
<amz3>I use a global
<jeffrin>hello all
<jeffrin>iam planning to learn htdp
<jeffrin>may be htdp uses scheme
***cataska_ is now known as cataska
***rlb`` is now known as rlb
<amz3>TIL: Scheme is hairy and CaML is simple
<amz3>I mean some one told me that
<amz3>I did some CaML/OCaml in the past! It wasn't programming from my point of view! I hated so much
<amz3>probably today things will be different
<csed>how2PAIP 101
<amz3>what is it?
<csed>amz3: The second chapter in PAIP deals with creating uh, a thing that generates sentences based on pre-defined rules. So for instance:
<csed>One of the exercises was to create a custom grammar. So I actually created a grammar for C at first, and it did generate valid C code (sort of). But then I was like "eh" and replaced everything with uppercase nonsense.
<csed>And now giant lizards.
<zacts>I love the new guile homepage theme
<zacts>and artwork
***mario-go` is now known as mario-goulart`
<amz3>that is netsplit!
***michaniskin_ is now known as michaniskin
<amz3>I have an issue with a macro. I think I need to know whether a datum/syntax is bound before doing something with it
<amz3>at L34 name should take its value from the lexical scope
<amz3>but because the macro mangle the content of uav in make-vars, it gets transformed to a "free variable" (a mk variable that is resolved later)
<amz3>I found a trick
<amz3>since mk variables SHOULD end with a "?" char in the pattern, I check in the macro that the datum/syntax ends with that char and actually only make those datum/syntax mk variables
<amz3>so now the convention to use "?" and "??" is a MUST
<amz3>or everything is broken
<amz3>*funny thing*
<amz3>mk has its own variables. So guess how to model variables on top of scheme?
<amz3>It's so simple that... I did not guess it ;)
<amz3>it use a vector with a single cell!
<amz3>IIUC it can use a single cell list too
<amz3>the use of vector make the output easier to read i guess
<amz3>here is the result:
<amz3>the code based on query* (green) looks definitly nicer (I cheated a little by spliting the procedure)
<amz3>now, let's conquer the world!
<amz3>I tried to load both conceptnet and wikidata
<amz3>both failed