3500+ entries in 0.054s
gabriel_laddel: ^ this contains information about how to get a list of all hardware - the kernel can determine hardware -> driver mappings somehow. Find, document.
gabriel_laddel: Another useful thing someone could do is to figure out how we would go about generating a canonical hardware -> driver mapping
gabriel_laddel: something went wrong with imaxima and gnuplot in portage - it only works on one of my machines now
gabriel_laddel: Anyways, if someone were to move portage to CLOS detecting cycles is easy, hence we can have all the documentation we ever wanted
gabriel_laddel: Portage has a USE flag (what is that? Idk, some nonsense abstraction) for docs so you can tell it to build all docs, but it gets caught in cyclic dependencies atm.
☟︎ gabriel_laddel: Adlai: you're piping data back and forth from CL to elisp or...?
gabriel_laddel: If someone would like to make themselves useful and move us that much closer to a source-only #-assets distribution, a CL interface to portage is desperately needed.
gabriel_laddel: I'll note that I wrote a prototype for the RPC described above - ran into an issue with TCP or the library I was using it from. Messages were disappearing in flight.
☟︎ gabriel_laddel: if you don't you'll be fucked, because generating syntatically correct code means AST manipulations
gabriel_laddel: ^ anyway, all of the above are simple if you adknowledge that you've got to operate on an AST
gabriel_laddel: I should be able to have a cursor on either one of the (x 4) and hit a keystroke to extract it into a let binding
gabriel_laddel: Why isn't all this information part of the version control toolchain? I can't query over all commits to find those that changed `some-function'..? WHAT THE FUCK IS HTTP? All I want to do is expose a procedure to a network - how is this more difficult than selecting a list of procedures which are then exposed?
gabriel_laddel: l data structures to those appropriate for the information being pushed through it? If the type information exists, shouldn't it inform auto-completion? Shouldn't I be able to query over the type signature, known return types and lambda list of all procedures?
gabriel_laddel: is not be handled automatically? Why isn't there grammer/spellchecking for my comments? If I modify a package or system definition at runtime, shouldn't I be prompted to write that change to the defining expression? Why must I optimize my programs and add type annotations? Can't test data be used to add typing annotations in an automated manner? Can't this information inform modifications of the program's fundamenta
gabriel_laddel: 7.
http://log.bitcoin-assets.com/?date=28-04-2015#1113989 << When I change the argument list, rename a procedure, use an unnamed reference why are the holes not marked or filled? Why must I manually declare my local variables and functions? Can't the current top level expression be searched for occurrences then factored out on a keystroke? Why must I balance strings, whatever the level of nesting or escaping? Can th
☝︎ gabriel_laddel: mircea_popescu: are you ever going to post the Jerusalem follow-up to your Istanbul article?
gabriel_laddel B.A.D (bitcoin-assets distro) update #1, I just finished rewriting mozrepl in parenscript. Now to cut out extra crud.
gabriel_laddel: While we're here, I'll note that being tied to text (perl) is quite limiting. On CLIM today I can tell any object how to "present" itself, i.e., how to draw itself on the screen. Such a thing is unthinkable in a language based around regexes.
gabriel_laddel: way less mental fatigue looking at them, I've found << I suppose. One can learn to read most anything. The importance here is the triviality of generating "syntactically correct" program texts.
gabriel_laddel: s-expressions + read tables are the ultimate notational abstraction
gabriel_laddel: not only that - you /must/ have a simple parsing algorithm, else upgrades are a bitch
gabriel_laddel: we know how ALGOL and meta-programming (tree-traversals) interact
gabriel_laddel: <trinque> seems relevant to the lisp vs perl thing << ummm... the angle I'm attacking this from is that meta-programming is inescapable and sexprs are the only sane way to go unless you want to program forth or APL.
gabriel_laddel: decimation: well, now it is sitting on-site at the 100 reactor sites across the US
gabriel_laddel: I ended up trolling through gov. documents on the yucca mountain facility (I had previously quoted wikipedia)
gabriel_laddel: mircea_popescu: oh, btw thanks for your input on my document
gabriel_laddel: hence they'll be beaten in the marketplace by yours truly.
gabriel_laddel: this conversation is going in the direction of "heat death of the universe" so I'll note that companies that use e.g., scala have to hire a lot of people to regin in the complexity inherrent in ALGOL.
gabriel_laddel: I should note that "they" (algol programmers) try to add the tree traversal properties of lisp to their languages
gabriel_laddel: (I'm assuming you're saying "why don't you think they don't - do tree traversals")
gabriel_laddel: they can't do tree traversals on their "structured data"
gabriel_laddel: my first pass at a definition for a string: A string is a collection of characters from an alphabet.
gabriel_laddel: I wouldn't dare draw that equivalence without agreeing on a formal definition for regexes and the process in question.
gabriel_laddel: which means that as long as you live, you will not live in a clean world. << select portions of humanity have, to varying degrees created "clean" computing.
gabriel_laddel: It doesn't run on "perl" "perl" is a well specified computer program. Meat runs another separate well specified meat program.
gabriel_laddel: This conversation bothers me. We don't understand how to make dragonflies, nor how the reproductive system works. We know how to write computer programs that work - the failing to do the latter correctly is a human failing.
gabriel_laddel: we know how to write computer programs such that you don't need regexes.
gabriel_laddel: people are machines - but computer programs are not people, nor cells.
gabriel_laddel: the reason we build abstract machines is because people are not very good at repetitive tasks.