484600+ entries in 0.302s
BingoBoingo: I doubt severely shared languages are a
thing
BingoBoingo: gabriel_laddel: What makes you sure we have a shared platform, or
that individuals here choose platforms for
themselves singularly?
gabriel_laddel: Without a shared language
the logs will end up in endless cycles of
the above.
gabriel_laddel: I'm simply of
the opinion
that our current platform (irc) is
too barbaric and doesn't force enough shared context upon us
to do anything interesting. Any sort of shared vision or whatever gets watered down into discussions like
the above.
☟︎ BingoBoingo: Hopefully
the DB doesn't have
the rights for it
to run, but...
BingoBoingo: trinque:
The problem with
taking any extant, deployabru RDBMS is
that unsanitzed inputs could
throw out rm -rf /
BingoBoingo: gabriel_laddel: Dunno how useful I can be. Drunk on vodka
trying
to protect my vital fluids from
triclopyr. Watched a Disney movie...
BingoBoingo: gabriel_laddel: Not dissing
the project, just questioning
the
target.
trinque: this assurance
that
the data means anything is why databases were created.
trinque: you could
totally be wrong
trinque: yeah, if you represent anything correctly in relational
tables you wont have false statements *according
to your system*
gabriel_laddel: Having
the same keybindings for
the web browser + editor gets me all hot and bothered.
BingoBoingo: <trinque> if I represent a grammar in SQL
tables you cannot fuck it up << You want
to
throw a stake down on
that???
gabriel_laddel: BingoBoingo: Making it production ready is going
to
take some
time, but as is, it is better
than any other distro I've used.
BingoBoingo still wonders about
the search for a singular #b-a distribution when x86 sucks as much as it does...
gabriel_laddel: Also, you don't have
to parse anything, or have "generator rules" or whatever, which cuts out a lot of
the complexity
that you get when doing something like what you've done.
BingoBoingo: Quick question...
There's no macguffin
the plot can use
to out BTC BTC is
there...
gabriel_laddel: So, in your model of computing you get
to work with incomplete languages in
the "data model" and when generating "views". While
this does provide defaults, when hacking lisp you always have
the full language at your disposal.
trinque: which amounts
to building a database in
to your
thing, yeah?
gabriel_laddel: you just end up writing a predicate
to see
that
the information you're manipulating is (every #'string ...) or whatever
trinque: the database if
told as much as is known about
the data you want
to represent, wont let you fuck it up
trinque: whereas I assume everything's hell and hopefully
the database will keep it all coherent
trinque: the idea is
that your lisp code is always right, hence
the data is right
trinque: and if one of your hacks fucks up
the logic of what it is
to be a "whatever's in
that list" ?
gabriel_laddel: "how in lips could you
take
the ast of a view against one
table" <<
This ends up being just hacking at sexprs with
the full language at your disposal.
gabriel_laddel: btw, I'm still putting
together a model of what exactly it is you've done so
that I can discuss
this with you using your vocabulary...
gabriel_laddel: because parsing is a
terribly boring (and
totally unnecessary)
task.
gabriel_laddel: anyone /can/ add meta-programming
to whatever language
they want.
gabriel_laddel: everything is a "list of lists" or a "tree" when you get down
to it (in
the compiler -
though yes, you can go directly stack machine).
trinque: pop
the result into one new list
trinque: there you go, now fucking iterate over
two of
them and filter by respective conditional expressions
trinque: I'm not seeing
the incompatibility of
the
things
trinque: so
then from what position do you proclaim lisp?
trinque: so like, why do you
think we put ASTs in
the database?
trinque: if I discovered a way
to do something lispy in a non-lisp environment, do you
think
this'd be
the first
time I've done
this?
trinque: well grow
the fuck up, and discuss ideas like a man.
trinque: gabriel_laddel: asumming
the AST is represented as a set of
tables with foreign key relationships, each
table representing a production rule in
the grammar
trinque: which amounts
to an iteration over one or more lists with an if statement
gabriel_laddel: trinque: could you rephrase "take
the ast of a view against one
table"
Adlai: people became lazy... everything changed once
the bitrot nation attacked
Adlai: back in
the old days,
there was no asdf:load-system... everything consisted of passing magic dispatch
tokens
to asdf:operate
gabriel_laddel: I have
to check
that I know what
these
terms mean before I respond. one sec
trinque: being able
to reflect and consider
things as sets, not
the
tree, is valuable
trinque: even if you pivoted one into
two
trinque: and derive
the view
that will apply
to
the new
table
trinque: consider
the ast of an operation
to change
the
table
trinque: how in lips could you
take
the ast of a view against one
table
gabriel_laddel: Sure, even in sexprs you have a 'grammer' you might want
to check
trinque: that's not
the purpose of it at all.
gabriel_laddel: I'd like
to clarify
that what I'm finding appaling here is
the huge amount of effort spent generating syntatically correct strings.
Adlai: "fetch me your
trainedest monkeys!"
trinque: I
think it's code for "not mine but his"
Adlai always found "business logic" reminiscent of "military logic" in
the necessity of distinguishing it from plain simple old "logic"
☟︎ trinque: tool for representing business rules, piles of data, and
trying
to make money, please.
Adlai: this seems
to be more "store
the AST in a db rather
than
text"
trinque: so give me
the enforcement of logical structure in lisp
Adlai: I'm not sure
there's any parsing involved
Adlai: (beyond
the hours forever lost
turning into weeks you'll never get back)
Adlai: what's wrong with extending code-is-data
to another data format?
trinque: I'm surprised you find
the relational model controversial.
trinque: so you can pull lexer
tokens out of
the former for example
that aren't in
the AST itself
trinque: gabriel_laddel: it's just a join between
the
tables
that represent
the grammar and
the AST
tables
trinque: SQL's a highly functional language, easy
to
think about
trinque: SQL is just a
terrible syntax.
trinque: along with automatic
to-string for
the ASTs
trinque: so one of
the projects I built with
that is a schema generator given some ebnf representation of a grammar
trinque: so yeah, you can just insert into
that and alter
the database arbitrarily
Adlai: can you stick some row into some
table of constraint specifications, which lets future-you insert previously-invalid rows
Adlai: this is starting
to sound like unfeature #4
trinque: I can bitch slap your mistakes with constraints and
the very structure of my schema
trinque: any atomic alteration of
the state of it cannot be wrong
trinque: if I represent a grammar in SQL
tables you cannot fuck it up
trinque: you can have an invalid AST in lisp, in
terms of your grammar; no one stops you from making any pile of cons cells
gabriel_laddel: like, say
that I want
to find all javascript functions with 3 arguments
trinque: it's fine for pulling levers on my database, but
that's it
trinque: you lispfolk can rip
the relational database from my dead hands
assbot: Logged on 29-04-2015 00:40:05; Adlai:
trinque: what, you want a
table with
two columns, CAR and CDR?
trinque: fwiw
the best relational system is probably yet
to be built, and is probably made of lisp
Adlai: trinque: make
them all equal!
tables of cons cells! cdr-coded self-joins!