« first day (1329 days earlier)      last day (3660 days later) » 

user55340
8:00 PM
On the other hand, you've got some novels released nder a CC license: antipope.org/charlie/blog-static/fiction/accelerando/… (attribution, non-commerical, no-derivative) and another similar one though not specifically CC: scalzi.com/agent
 
user41796
Easier for the compiler to complain and throw an error. :-)
 
user55340
That said, I bet that @ThomasOwens hasn't read about the lobsters yet.
 
@MichaelT That was that ebook that you linked to? Yeah, I read it.
 
user55340
Best part of accelerando - each part follows from the previous, but you've got no idea where the next one will go.
 
user55340
And those lobsters keep showing up in strange places.
 
user55340
8:02 PM
(btw, there's a sequel of sorts to it.
 
user55340
Glasshouse is a science fiction novel by British author Charles Stross, first published in 2006. The novel is set in the twenty seventh century aboard a spacecraft adrift in interstellar space. Robin, the protagonist, has recently had his memory erased. He agrees to take part in an experiment, during which he is placed inside a model of a late twentieth/early twenty-first century Euroamerican society. Robin is given a new identity and body, specifically that of a woman named "Reeve". Major themes of this novel are identity, gender determinism, self-image- and conformity. Contrary to po...
 
There has to be a good polymorphic way to assomplish this.
 
psr
@MichaelT Though there is actually nothing preventing congress from only allowing software copyrights that are for the purpose of making the software more available to the public. Except, you know, congress.
 
user55340
Another neat book - the "what if the strossian singularity was averted"
 
user55340
Implied Spaces is a 2008 space opera novel by American author Walter Jon Williams. It explores themes of transhumanism, artificial intelligence and ontology. Setting Implied Spaces takes place approximately 1,500 to 2,000 years in the future. Humanity has entered a period of technological singularity fueled by a series of massive supercomputers orbiting the sun. These computers act as giant solar collectors and computational machine for running human society. While humanity numbers in the hundreds of billions, only perhaps a few million live in the solar system. Some have gone on slowe...
 
user55340
8:05 PM
(I think that glasshouse is kind of a prerequisite for reading implied spaces in that it explores an aspect of post-singuarltiy war and implied spaces explores another)
 
user41796
Pro tip: It helps to actually attach the debugger to the process prior to running the test you wanted to walk through.
 
and set a break point in the test
 
user55340
@GlenH7 not if you already have an omniscient debugger: lambdacs.com/debugger/debugger.html
 
user41796
don't be getting too fancy on me now
 
Anyone want to suggest a new design? Since this was supposed to be my brilliant solution to fix stupid ones?
 
user55340
8:11 PM
And I owe @ampt another $0.05 - blog.jetbrains.com/idea/2014/03/…
 
user41796
@ThomasOwens You need to call the parent's version of the method or the child's version?
 
@GlenH7 The child's version. The parent has all empty methods for each of the message types. Each child chooses the messages that it needs to accept and what to do with them via overloading.
 
user55340
 
user41796
Or is it that you have a parent reference to a child object and you need to poke through to the child's method?
 
user55340
(note the 'history of i' panel)
 
user15026
8:12 PM
@MichaelT I keep wanting to read and like Stross's books but I just can't ever get into them.
 
user41796
@MichaelT Can I get that in C#? :-)
 
@GlenH7 That. I have a Parent x = new Child().
 
user55340
@AshleyNunn accelerando is a fun one. But you've got to read it as a collection of short stories.
 
And then, I have methods for the handful of subclasses of Parent. Based on the specific type of Parent, I need to call the right method.
 
user41796
@ThomasOwens And I suspect you can't do the proper cast to child class in the code, right?
 
user55340
8:13 PM
Its the idea that technology (at that point) is moving too fast to try to keep up with - and its speeding up.
 
I could. But it doesn't matter, I don't think. x.getClass() = Child.
 
user15026
@MichaelT See, as a premise, that sounds awesome, but so do others of his books, and I try to read them, and something about his language or something just florps my brain and I can't do it
 
It's just that instead of calling the method for accept(Child x), it's calling accept(Parent x).
 
user41796
@ThomasOwens I would think that if you could cast the reference to the actual child then the compiler will know which is the "correct" child method to call
 
So there's no good polymorphic solution? That's going to bloat my code with a handful of casts...
 
user55340
8:15 PM
He's writing it like its perfectly normal what he's writing about... and its not normal.
 
I wanted to avoid instanceof and casting.
 
user41796
@ThomasOwens offhand, I can't think of any. But that's a bit outside of the type of things I normally write
 
user15026
@MichaelT and I think that is where my brain starts to hurt, because I can't quite grasp his concepts and world buidling
 
user55340
>
Welcome to the fourth decade. The thinking mass of the solar system now exceeds one MIPS per gram; it's still pretty dumb, but it's not dumb all over. The human population is near maximum overshoot, pushing nine billion, but its growth rate is tipping toward negative numbers, and bits of what used to be the first world are now facing a middle-aged average. Human cogitation provides about 1028 MIPS of the solar system's brainpower. The real thinking is mostly done by the halo of a thousand trillion processors that surround the meat machines with a haze of computation – individually a tenth
 
user55340
And then you get to the eigen parents... which makes one's head hurt.
 
user55340
8:17 PM
Just remember... the cat is a very important character.
 
And if I don't have an accept(Parent x) method, I get a not applicable for the arguments.
 
user41796
@ThomasOwens I have generally always worked the reverse where I have a child reference and need to get back to the parent method. Sorry I can't think of any other ways to approach the problem.
 
gist.github.com/ThomasOwens/3ecb7ad44109d13e3070 demonstrates my problems. I suppose I could ask on Programmers or Stack Overflow.
I think it's small enough to post somewhere.
Would you all agree that this is a design question? Or should I make a simple class diagram for Programmers?
 
user55340
Looks designish. Borderline question... gota watch out to make sure that Robert guy doesn't try to close and migrate a good question away though.
 
you want to be able to pass random messages and have them handled correctly?
then I'd say only have a handle(Message) in the super and override that and use if(msg instanceof SpecificMessage1) as needed
 
8:27 PM
@ThomasOwens eek, composition over inheritance is screaming itself from the rooftops at this. Decompose the problem, and then don't recompose it into a single whole. Rather, take the components identify all relations between them and figure out how to make them communicate as independent pieces rather than bound pieces such as inheritance forces.
 
you need a runtime check either way
 
user41796
I really wish I had more close votes today
 
@GlenH7 here, have mine
 
user55340
@GlenH7 got one left.
 
@ThomasOwens s/then/than/f
 
8:30 PM
@JimmyHoffa I really can't. Each child is a unique message/event representation. The parent represents the common things that each message has, while each child has its own unique information.
 
user41796
 
@ThomasOwens header.
 
@JimmyHoffa Yes, each has a common header and footer.
 
common = header, specific = body. This is how I've handled precisely that situation before. HeaderHandler = common. BodyHandler = chosen based on info in the header. Not chosen based on type overload
 
or a getter in Message to get the type
 
8:32 PM
So the problem becomes generation. Right now, I have a method that takes a byte[], returns type Parent, but actually returns a specific implementation. An Abstract Factory, really.
 
@ratchetfreak tons of ways to do it, avoiding a switch is best, but push comes to shove... it beats using inheritance as such
 
If there's no parent superclass, how do I generate an arbitrary message from a byte[]?
 
factory
 
@ThomasOwens WholeMessage<T> { CommonHeader Header { get; set; } T Body { get; set; } CommonFooter Footer { get; set; } }
 
Let me think this through...
 
8:34 PM
where T : IBody, and WholeMessage<T> : IMessage which is {CommonHeader Header { get; set; } IBody Body { get; set; } CommonFooter Footer { get; set; } }
 
Hmm.
 
(this is all just spaghetti thoughts; obviously far from accurate, but hopefully conveys some concepts that might make sense)
 
@JimmyHoffa why not combine header and footer
 
Actually, I did combine header and footer already.
 
they are both common and the footer is likely just a CRC
 
8:35 PM
@ratchetfreak that's fine too, I'm just throwing it out there cuz he mentioned footer. Though to be sure, if you want to avoid readahead it may be easier/more performant to have them separate
 
user55340
@GlenH7 there's my last with a too broad. I still think it would be a good chat question... I could give a fair bit of answers to the person's question about my experiences.
 
@ThomasOwens you're working on a parsing problem. Just don't write an NCube. What the hell is with everyone doing parsing today?
@ThomasOwens jparsec.codehaus.org
 
I didn't parse I just wrote a parser, I last touched that code a year ago anyway
 
user41796
@MichaelT same here; but it's a bad question for main
 
@JimmyHoffa Considering it's a binary message format, I think that may not be the best tool.
 
user41796
8:37 PM
@ratchetfreak is that like saying you smoked but didn't inhale?
 
@ThomasOwens parser combinators don't care what the format is.
parser combinators just let you define the rules for parsing whatever value types you're dealing with
 
nothing stopping you from emitting the proper tokens either way
 
user55340
@GlenH7 I opened a haskell file in an editor but didn't write a monad.
 
they're general - that's why they're so good for parsing problems
@MichaelT bullshit.
 
user55340
ed foo.h
567
q
 
user55340
8:38 PM
@JimmyHoffa see? Didn't write a monad.
 
@MichaelT oh, you were using ed. It doesn't support monads, try emacs next time.
 
Yeah. Inheritance was a dumb idea.
 
yay another convert
 
Because it's not even an is-a relationship. I don't know what I was thinking when I wrote this code.
 
you were writing a parser
no one thinks clearly when writing a parser
 
8:41 PM
Luckily, I haven't released this library yet.
So the only one using it was me, dogfooding it.
And this is why you dogfood. To find your stupid problems before releasing it to the world and needing to make breaking API changes to fix them.
 
I test my apps by dogfooding
because I'm too lazy to write test classes
 
I just print my code out and feed it to my dog, if she get's sick I rewrite and try again. She seems to like LISP...
 
I don't have a dog to feed it to...
 
user55340
Ohh, academia graduated!
 
user55340
 
user55340
8:54 PM
I blame caching.
 
It's always caching.
 
user55340
42
Q: Site graduated! New design launched

Stéphane MartinAs you can see the new design just went live. Which means this site has been officially launched! Congratulations! Thank you for your valuable design feedback. If you're still seeing the old favicons, please load the follow urls and do a hard browser refresh. http://cdn.sstatic.net/academia/img...

 
user55340
They've got a nice site design.
 
user55340
Heh - and the top voted advert:
 
user55340
11
Q: Community Promotion Ads - 2014

Grace NoteYou guys have gotten your own design and graduated, hurrah! And with that, you now have the opportunity to setup Community Promotion Ads for your own site, not just to apply to other sites. What are Community Promotion Ads? Community Promotion Ads are community-vetted advertisements that will s...

 
user55340
8:57 PM
 
user55340
One of the neat bits is how the icon is based off of Stack Overflows:
 
user55340
 
user55340
(well, I see it that way)
 
it's a bunch of books on a shelf...
 
user55340
@ratchetfreak yep, but its also got that 'stack' side that's leaning... rotate it 90 degrees and think of SO's logo.
 
9:00 PM
@ratchetfreak nope, it's a StackFellover
Mainframe developer learning hadoop... I think his tapes would get tied up if he tried to connect so many of them together...
 
user55340
@JimmyHoffa check the comments:
 
user55340
0
Q: Mainframe Developer planning to learn hadoop. How long will it take, and what is the best way?

mankand007I don't have any java experience, although I can pick it up fairly quickly. I can only work during my off hours, around 1-2 hours daily. What is the best way to go about learning hadoop (and java if needed)? How long does it take to get a good confidence in hadoop, enough to face interviews? ...

 
@MichaelT I heard "Hello Andy." in my head when I read the title alone... don't need to check the comments...
 
user55340
@JimmyHoffa nah... not andy. Not that MO at all.
 
I think this would rather suit programmers.stackexchange.com. — Axel 2 hours ago
 
user55340
9:06 PM
The comments are "ask this on P.SE" and me saying "No no no no"
 
^-- flagged rude/offensive.
 
user55340
> Please delete this as "not good advice" - people posting the "belongs on P.SE" reinforce the "post these types of questions on P.SE" which when upvoted shows up prominently. Kill it with fire.
 
@Axel please make it a habit to stop yourself from suggesting people ask Q's on other sites when you aren't an active participant on those other sites as you can't know those other sites acceptable scope without spending time on them first. — Jimmy Hoffa 40 secs ago
 
user15026
@MichaelT Oh, I see it now!
 
user55340
@AshleyNunn what is seen cannot be unseen.
 
9:13 PM
@MichaelT that's a pretty strong statement on the nature and existence of time travel. Just saying.
 
user55340
@JimmyHoffa Btw, I like the lead in for Timemaster by Robert Forward.
 
@MichaelT at least his name isn't Richard Forward... just saying.
 
user55340
> There exist semieducated but obstinate people who have raised the concept of strict local causality to godhead, and attempt to use such words as "obviously" and "it only makes sense that..." in an attempt to "prove" that their version of causality cannot be violated, and that *any* sort of time machine is logically impossible. From my reading of scientific literature, they are wrong.
If I receive a letter from this sort of person complaining about the "impossibility" of the time machines in this novel, I will throw the letter in the nearest wastebasket... *unless* the letter is accompani
 
whoever placed the ulna where they did should be shot. That dude's an asshole.
 
quick question, does std::map<int,int>[k] init the value to zero if k does not exist?
 
9:25 PM
@Zeroth .NET?
 
C++
 
shoot, go ask in the lounge (if they don't get too busy eating you alive first). None of us want that dirty language around here...
 
:P
 
@Zeroth half the problem with C++ is - that probably depends on which of the numerous runtimes you're living in.
 
Ugh, yeah
delphi/vcl...
 
9:27 PM
I'd write delphi over C++ every time...
You want a systems driver? Maybe a kernel? My choices are Delphi or C++? Yeah, still going with Delphi...
 
 
1 hour later…
10:34 PM
ack JScript.NET deprecated? Noezzzz
I always liked it...
^--- cool, new Haxe can directly naturally interop with .NET assemblies as if they're one of it's own.
also cool, it has real macros; not homoiconic but AST access pre-code-generation
 
10:56 PM
@JimmyHoffa TypeScript is the new awesome.
Do you use Haxe?
 

« first day (1329 days earlier)      last day (3660 days later) »