i wish there were a good, free 3D modeling program that somehow magically had a UI that was about getting artistic stuff done first, and less on the low-level vertex manipulation bleck type stuff.

sketchup seems to sorta try to be less bad that way, but then it suffers from another UI problem: you have to do it the way they want you to do it, rather than possibly a way that makes sense to you to do it. so that's really frustrating. like, it ends up feeling like i should just go back to using wings3d instead.

so you are better off just hiring really good hackity hackers? ;-)

what is particularly impressive about Flash + Firefox 3 (no beta) is that on this here Mac Book Pro it fails to read the keyboard properly. so games which use keys to move are all effed up, like i can only move in a couple of directions, and can't stop. this happens in more than one Flash program and it doesn't happen in Safari, so I guess it is some happy thing about Adobe and Mozilla.

'As Doug Lea put it more than 20 years ago “the problem with defining equals as class method is that there are too many senses of equals for one method to support. and the author of a class won’t have all of them in mind”.' (You have to scroll down a bit find that, it is interesting.)

you tell me who has won, now. hint: it ain't us.

speaking as a software person: when it comes to making a really great design, it is the exceptions which are a true test of mettle, not the rules.

would it make any sense to have a feature where you can mark a class (e.g in Java) as "synchronized" rather than having to add "synchronized" to each and every method?

i mean, sure, it is horrible style because you should be paying more attention than that, but when faced with some gnarly inherited code, i would be happy to be able to use such a thing at least as a first hack pass.

oh well.

it sure will be funny when we're less mortal. well, actually, only those of us who can afford to be, and who aren't dead by the time the tech works. in other words, not me.

(funny how the TED web site has a lot of pretty crappy ui issues.)

obviously my idea of "fun" is not what other people think of "fun"?! i want to pay money to be camper-griefed?

i mean, gosh, why would anybody want to be able to preview their comment? crappy blogger crap.

oh those funny JSON advocate kids, not having a decent JSON library for Java. Partially because, fundamentally, JSON is kinda crap!? "Hey, let's get away from something complicated and make something simple... which is simple because it doesn't really do what you want it to do. Ha ha." (Well, what I want it to do :-)

Guess that's why AMF is supposed to be nice and popular? But, there again, good luck getting a decent library for it - at least, if you are looking for open source. (Or did Adobe get around to OSing theirs?)

Java generics are sorta nutty. Check out the deal with Enum, in particular. I might actually have need to try to use something like that soon, even. Scary.

• Software task estimates should be probability distributions, not precise sounding single values.



• Wow, some genius! "You can use System.out.println from Java to print debug statements to the screen."



• Not, shall we say, Vista? "The finest pieces of software are those where one individual has a complete sense of exactly how the program works. To have that, you have to really love the program and concentrate on keeping it simple, to an incredible degree."



• Perhaps a good band name, or two. And: holy heck, so good it hurts. (why not rub it in, too?)



• weird how firefox puts up some very first url for url completion for a site i haven't visited in donkey's ages.



• seems like Bracha's newspeak might be a "middle way" language?



• i'd like to just burn through DoD money some day, too.

man. i sure wish i were that good.

my bet: pax will replace e3. and more power to 'em.

wow, here i thought i should like Dino Run from the sounds of it, and then it has to go and have a pathetic jump that is like negative fun, and pits you get trapped in. so that's a really great way to not have me ever want to play it again. sort of the opposite of what the first 5 minutes (i didn't even last that long) should be all about, non? weird.

the main effect Indiana Jones 4 left me with was: i feel old.

software is hard. just 'cause you have tests doesn't mean your tests don't suck, too.

oh, yeah, i love me a good naming!

blogger, to paraphrase, is the worst thing except for all others.

i sure am glad the green maximize window bar button in Mac OS X is such an excremently unhelpful piece of poo. yay, hallowed Apple usability. (like, usability which iirc didn't suck this hard in pre-OS X days.)

just say no to RPC; you'll be better off in the long run. (cf. 8+ fallacies of distributed computing.)

some day, it would be - gosh - kinda sorta maybe oh tickle me silly nice if eclipse would freaking show statics in the bloody freaking debugger.

or maybe it is a JVM thing?

[Update: it can, but for some asinine reason they aren't on by default?!]

also? inner / anonymous class member and local variables don't seem to show up worth diddly squat, either. yay!

if i were you, i'd start hoping that we really do run out of oil, and post-haste. and that nobody gets fusion working. so we can all just return to cave dwelling rather than cyber gulags.

'I ask Zheng whether China's surveillance boom has anything to do with the rise in strikes and demonstrations in recent years. Zheng's deputy, a 23-year veteran of the Chinese military wearing a black Mao suit, responds as if I had launched a direct attack on the Communist Party itself. "If you walk out of this building, you will be under surveillance in five to six different ways," he says, staring at me hard. He lets the implication of his words linger in the air like an unspoken threat. "If you are a law-abiding citizen, you shouldn't be afraid," he finally adds. "The criminals are the only ones who should be afraid."'

super duper nerd out time, baby.

you can't win child
you can't break even
you can't even quit the game.

(michael jackson can sing?! who knew!)

if there is an error on a line in eclipse, then you cannot get the javadoc on anything in that line as a tooltip any more - which might be exactly what you need to most expediently be able to fix what is causing the error. sweet.

if humanity makes it that far, i sure hope the XXX Century (as opposed to the recent XX Century) really lives up to its name!

nice how google reader's "mark all as read" button sometimes (a) is not disabled (2) is positioned over a long list of unread items and yet (iii) doesn't actually do any bloody thing about marking them as read when i click on it.

somebody make me an organic free trade no slaves no sir tshirt that says:

i love twitch.

no, wait, wait, maybe:

i <heart> twitch

?

moderators have a rough life; damned if you don't, damned if you do.

(as an aside, i love how blogger fails to post this then says in the error page something about how i should go to great lengths to report the error and tell them what i was trying to do when it happened. because (a) they can't figure that out from their own logs?! and (b) because "i was trying to blog" is such an out-of-this-world concept for them?! haaayyytttee.)

first, you don't have your cell phone on silent-vibrate. second, it plays some song that i might have liked if i hadn't first heard it a zillion times only as your ring tone, and is thus for ever ruined and only makes me see red.

i love technology! oh, and people.

i love it when a site uses captchas where you can't tell if certain letters are upper or lower case.

could the firefox upgrade process have more new dialog boxes coming up and taking focus away from what i'm trying to do otherwise? hayte!

almost as bad as your mind is able to believe; as if somebody making an FPS never played any FPS games before. i guess there is something about how 99.9% of everything is crap, i shouldn't let it get to me - there's on occasion the nanaca crashes of the world.

too! much! to! read!

hey, and i just watched that movie "V", neat!

nefarious heh heh!

"Greg that’s a good sentiment. Star Trek definitely has a core message of solidarity and using technology to advance the human race rather than destroy ourselves. That’s a good message and I hope it’s captured in the new movie."

every time i print something in OS X i have to tell the print dialog that i want it 4-up and double-sided.

why doesn't it remember? why does everything have to suck?

couldn't we all still learn a thing or two from Lisp?

funny how the very first example makes me think, "gosh, maybe that wouldn't be a problem in a typed language." well, ok, maybe it is more of the one word, multiple duties problem. something?

hindsight is 20/20, or just somebody slinging mud to get page views? you decide.

wow. now that's what i call sci fi.

wow. like, are there any games that don't suck?!

amen!

at the moment, i think i would prefer to be able to have the ability to replace some forms of function overloading with one function using optional arguments with default values and nametags. just seems like it would overall be less verbose and confusing and more maintainable. although then it might be harder for your IDE to show you if nobody is using a given form of it, which Eclipse can (mostly) do with current Java.

if you want to see something "funny" in Eclipse, pick a term that is used a lot and use refactoring to rename it, then before committing the change hit undo. it will laboriously undo the typing of each letter in each and every occurrence in the file, which can be kinda... slow.

i think these days my personal bent is that i loathe anything that isn't designed first and foremost with an eye towards flexibility. the YAGNI philosophy might be interpreted by folks to mean "hack the crap outta that thing and hard-code it all!!" in which case they should be taken out back and - let go. i'd rather that people somehow either ignore YAGNI, or think of it meaning that you aren't going to "need" the tight coupling, and that you should stay an order of magnitude more decoupled than you'd normally by knee-jerk.

too bad none of this philosophizing will help me with the current code base i'm looking at. :-(

yeah, Virginia, software is hard. especially programming languages. you'd think by now it would all be worked out? nope. what are the underlying problems? basically: security + maintainability (in some ways) vs. succinctness + flexibility + featuritis + maintainability (in some other ways) perhaps.

oh yeah, i'm sure it is a great feature of open office that it can't (as far as i can find) tell me what number the column i'm on is in a spreadsheet. it just shows it as e.g. "EB" with no numerical count of how many columns along it is. the math shouldn't be hard but i shouldn't have to wonder.

so in other words higher order doesn't really help you?

there's an interface. several classes implement it. Eclipse only shows me some of them in the type hierarchy. so i was mightily confused for a while there. until i realized that Eclipse, verily, has issues. oh well, software is hard.

did somebody finally make a backup system that isn't pure ass?

it always surprises me when somebody has a clue, and makes me sad i don't work with them.

it is great! except for when it is a cluster <expletive>, ahem.

it's not the tile, it's the grout i utterly loathe.

and who can honestly say they haven't had oral sex with the pillsurby dough boy?

i know i have.

i can maybe see how it is good that windows resets from entering japanese to entering english when i start a new application that i haven't used the language bar in before, but it drives me nuts that when i then switch to japanese it doesn't bother to remember that the japanese i was doing last time was hiragana.

windows is killing me. i just want to be able to type in hiragana, but it insists on either converting everything to kanji, or freaking out and turning things into messed-up extended ascii, like german double-esses!?

moral: hiragana entry works only just enough to be really confusing in Notepad2. so instead use like MS Word, it seems to work better there.

it sucks when the Win XP Taskbar tooltips get stuck rendering behind the Taskbar itself so they are chopped of or completely invisible.

boy oh boy, open source rocks! (joke!)

some web apps have an autosave feature. e.g. one wiki i used says when the page was last auto-saved. that doesn't really help me since i don't off the top of my head know what the precise current time is. it would be better if it said how long ago the save was done, or what percentage of the document is backed up, or some such thing.

so now you can wire together a bunch of questionable cruft (as all software is questionable cruft) into some gianter cruft? ... profit!

computers suck. people try to make them "usable" and "friendly" by e.g. having a (Windows) "My Pictures" folder, or an (Mac) "Applications" folder. of course, if you want to move or rename those, you are probably just begging for serious pain, right?

oh that wily christopher alexander (not, not the director of pr0n, you silly), turning up everywhere.

why not just use a prototype based language instead?

funny how firefox's spell checker doesn't know "firefox's".

while the smallest effective difference is very useful in protecting against heinous/designs, sometimes layering and separation really help transmit the crucial (ymmv) info.

maybe things like maven aren't that bad when compared to the bad old ways? well, maybe maven isn't so bad in some ways but then lulls you into other ways which are heckdom?

maybe Monad has it, dunno, but it bugs me that things like the unix shells do not have a standard way to integrate word completion with programs. so e.g. if i type "svn sta" and hit tab, it would be great if it could expand to "svn status".

all i can say is, anybody who isn't surprised that The Man really really sucks either isn't very smart or is a carefully buffered exec? interesting to follow the links to other tales of funny-to-an-outsider woe.

666 items in my gmail spam folder. dare i go look inside?

i find it hard to understand how Adobe can put out Flex Builder and not have any way to do a global file search. this is what the hallowed world of Flash development has to offer? zoiks.

i am not a fan of things like facebook, but it is nice to see them putting out some possibly neat software tools.

nice to know that actionscript screws the pooch in more ways than one, i guess just like any other language.

"alpha:Number (default = 1.0) — A number that indicates the alpha value of the color of the line; valid values are 0 to 1. If a value is not indicated, the default is 1 (solid). If the value is less than 0, the default is 0. If the value is greater than 1, the default is 1."

i read that (okay, skimmed) like 5 or more times getting ever more frustrated that they didn't say which was opaque vs. transparent, and only just now finally saw the one relevant-for-that-issue word out of all of them: "(solid)". way to bury important details in poorly written prose. oh well.

hooray actionscript.

documentation is hard.

well, maybe it is some form of job security; make the Scala syntax a little bit arcane (read: barfy) so not just anybody can use it?

i can't believe i'm going to give away my freaking genius game idea. although the real icing on the cake was through up by a friend.

     New! From VisionActy: the sequel to its classic

     Probably Wrong Verdict: SoCal

     the long awaited

      Probably Wrong Verdict: NoCal

while i don't want anybody to diss the red-green color blind folks, on the other hand i'd suggest that red is not for "good" marks, ok? sheesh.

i sure am glad somebody out there sucks less; especially sucks less than i, who do nothing, do.

now if only it were for the JVM :-(. maybe one can sorta make due with (some quite different) alternative tools like Guice, Ashcroft, j/Easy-Mock, and such.

i think it is funny that there are towns called Effington. i would like to have been born in some British town that was near a river called something like Effington-on-Thyme (get it?) or some such.

i've probably written about this before, but Eclipse's Open Type dialog is sorta nice, but also sorta an experience in crack-smoking effed-up UI. there are 2 classes with similar names: FooBar and Foo. i'm trying to get to Foo. so i type Foo and hit enter.

and it opens FooBar.

my theory: anybody who has a "how's my driving?" sticker on the vehicle with some phone number is not actually interested in getting reports if they fail to meet any of these requirements:

• a toll-free number


• a phone number that has a mnemonic word, and is shown thus, like 1-800-ARE-SAFE or whatever


• a very succinct vehicle identifier


• vehicle identifiers that are redundant so that even if you only remember part of it you can still get the right vehicle

anything less means, to me: "hey, we're ass-hats!"

at least around where i live, american highways are rather poorly sign-posted. like, there are a few exits going one way that don't exist going the other way on the same highway. at the very least they could put up some bloody signs saying "for RoadX (for which an exit exists on the other side) take ExitY" or something?

it always bugs me that deleting files takes any time at all, especially when killing a directory tree with a root - couldn't it just mark the root with a "deleted" bit and be done?

i know, then writes would have to pay the price of figuring things out, rather than spending all the energy during delete so that writes are faster, and most of the time people are more worried about write performance.

still, seems like there could be some kind of tricky optimization that could make deleting suck less. hey you phd candidates, get right on that.

it's what's for dinner.

oh, the truth stings, my programmer friends.

any artist who refers to Q*bert has got to be way cool.

if we knew how to define what we expected and allowed, debugging would possibly be easy. well, ok, maybe testing would be, but not debugging. but i think (a) we don't have great ways of saying what should be allowed and (b) even if we did (like some amazing natural language system that somehow didn't allow for horrible inconsistencies and vagueness) i think that we just don't often know precisely in our own heads what should / not be allowed.

?!

entirely too much finding of utter bloody genius.

fundamentally, we aren't wired up to really be able to easily debug much of anything, let alone concurrency.

pod coffee is evil.

eclipse suck: Quick Type Hierarchy works looking from an interface on down, but is not working for me when in an implementation and trying to look up the hierarchy; it doesn't show the parent interface at all.

even with the most rudimentary and static languages there is a gulf between the ASCII and what happens at runtime. and they can be really hard to understand and debug. which means i fear going whole hog and making everything late-bound and dynamic and relying on emergent behaviour. if there were good static checks and proofs which constrained the runtime statespace and helped match the specs etc., then i'd feel a lot safer going in that direction.

another reminder that it isn't (solely) the syntax, it is also (highly) the semantics.

so there are lots of interesting takes on concurrency (erlang, clojure, stm, etc.) and yet i wonder how well tested any of them are. getting concurrency right is not easy, so jumping over to something like scala's actors is a little bit scary. like, if i take a random 3rd party single-threaded library i have a better chance of (a) being able to recognize incorrect behaviour and (b) be able to fix it, whereas with concurrency systems i'm a little bit more clueless and thus vulnerable.

which all sorta boils down to thinking that while clojure and scala's actors are fine and dandy, it would be completely insane to use anything other than erlang for an actually shipping product. at least for like the next 5 to 10 years.

either MSFT finally had to cave, or they have some really nefarious tricks ready... like, uh, patents maybe?

no worse than ones from Sun, for example, i guess.

uh, ok. so firefox 3.0b5 doesn't save the .dmg to disk, but opens it as a giant file in the browser. and the ui thread apparently doesn't have priority over the reading-off-the-socket thread because i can't get it to respond to the back button or the stop button for like 5 or 10 seconds.

and yet when i click on the same link in Safari, it downloads to a .dmg file just fine.

do not go gently into that macro good night.

oh gosh, lots to learn. maybe the JVM won't quite die off just yet?

shared mutable state concurrency sucks. but message passing scares me a lot when it comes to gathering debugging information: i'm assuming there won't be a "stack" across all the messages i can examine so easily, which seems problematic?

yay Java. i want to enforce order of locking by passing a closure to something which takes care of the order of locking. but that's a nightmare suckfest in Java. so that's great.

ha ha this will make 1,667 posts.

remind me to get around to sucking less, like by doing my own flash card learning system that has all the bells and whistles i think it should have. not just for gratuitous reasons, but to really help teach and learn.

of course, even with a great software system that isn't enough. you have to have somebody get together a good set of data to use. (for example, with learning a language or a foreign alphabet you might want to have really simple words as examples of the individual letters.)

education is hard.

user interfaces are hard to get Just Right.

i have wondered in the past if it was just me not groking spreadsheets, but i guess they really are kinda borked and could stand some fixin'. right! somebody get cracking on that one, eh?

so i broke down and installed Steam. which pretty much makes me like one of those people complicit to something utterly immoral.

the kicker is - not that it shouldn't be too unexpected - that after the install went for a while it said "The Steam servers are too busy to handle your request. Please try again in a few minutes." and it appears that most of the games off of the orange box edition i was trying to install were not installed. and it doesn't seem to have anything that lets me simply try again to install them anywhere. totally and utterly ungrokable.

that along with drm suckage pretty much sums up why i've always detested the whole Valve and Steam thing. i seriously would rather live in a world that had no Steam and no Portal because of it, because people realized how evil and wrong it all is inherently and never bought it, consigning Valve to destitution.

i suck. but they suck worse. but i still suck for not having stuck to my principles. oh well.

blogger's ui for moderating comments... pains me.

i'm a usability weenie (or is it wiener?). so unconferences, apparently, freaking kill me. talk about a cluster <expletive>. just because you know something doesn't mean you know how to present it. among a zillion other basic problems with the entire approach.

okay, i'm speaking from an N of one, i'm a jerk. but it sure doesn't make me feel like i want to rush out to another one.

i think i've been saying something like that for a while now. (just more incoherently?)

blogger just said i couldn't log in because i had cookies disabled. so i hit the back button and went in again and it worked. so i don't think cookies are disabled. yay, blogger.

bah, humbug! kids these days have it so good compared to my school experience. check out all the game design challenges.

hm. maybe my vanity language would only have 1 fixed size integral type (64 bits) and one fixed size floating point type (64 bits? 128 bits?). and then some BigNum types. (and utf8, as an aside). so that you don't have to get into all the suck of "was that an int that just overflowed even though i wanted a long?" or "was that an ell at the end of the number, or the digit one?" kind of stuff.

maybe if it were cool, it could let you define your own types of arbitrary size so if for some reason you thought it was really so important to have e.g. a giant array of single byte values, you could.

but, really, i don't know what i'm talking about.

you have to break a few eggs to make an omlette.

i'd buy that for a dollar like to write a server with some tools like that. call me a masochist, i guess.

yes, making a non-mostly-suckful programming language is really freaking hard!

oh those silly twitter kids, with their humour and funnies.

hm... personally, i think that if the main format for your resume is the MS Word DOC format, you are kinda someone i don't want to hire.

indictment of OOP? in Java, you have to cast Graphics to Graphics2D to be able to set the pen stroke.

uh, you mean like shared mutable state in general, and with locks for concurrency?

sorta funny to then come across, "Although coroutines do not provide an advantage over a well-designed multithreaded application for typical GUI scenarios..."

don't get me wrong: I'd rather be doing Java than C/++.

so blogger sucks and often gives me an error when i try to post and i click the back button then publish again and then it works. or i have to do that a few times in a row before it "takes". now-a-days i see something where it will flash a page that has the red error banner on it and then that is quickly replaced with a success message. oy veh. better, but still, yuck!

i click on the close box of an application on OS X (Terminal, to be specific) which doesn't currently have focus. it puts up a dialog box asking me "are you sure?" and i hit the return/enter key. the key press doesn't cause the dialog box to go away, and instead the press goes to the application that had focus before i clicked on the close box. wtf? i would think that clicking on a non-focused application anywhere should first cause it to get focus?!

the more i think about it, the more i think the songs of Hall'n'Oates really are total genius.

just because your technology is flexible, do not expect your people to not screw it all up nevertheless.

hey, yeah! why doesn't the ATM remember my language choice?! stupid usability user interfaces crappy systems deprioritize the end user.

hm. this almost makes sense: "Governance is about decision making principles to elicit the desired behavior. The governance policies should help an organization decide what things are horizontal, what things are vertical, and then assign people to work on those efforts within those architectural boundaries. Right now, many organizations are letting project definitions establish architectural boundaries rather than having architectural boundaries first, and then projects within those boundaries. Project boundaries are artificial, and end when the project ends. Architectural boundaries, while they may change over time, should not be tied to the lifecycle of a project."

words like "epidural", "spinal", and "catheter" are pretty much guaranteed to make me feel queasy. i do just writing those just now, even. so how about we change the terminology? get some marketing people in to call them something with "happy" or "fun" or "tickle" or "smooth" or whatever in the names. "Happy Tap"? i dunno.

oh, hey, games!! remind me to find the time to make something so i can get on lists like that. before i die / heat death of the universe. :-P

i always found the first VR Racing to be a wonderful experience, especially when the force-feedback was working. it just really felt good that the wheel would resist, and even when later VR Racing games came out with more polygons and texture mapping they never felt as good to me as the first one. i played an 8 machine setup at Disney Land with some friends once and it was a toe-to-toe race (i lost by a wheel) and that was one of the most exhilarating arcade experiences i've had.

i am so glad this MacBook Pro has the CD/DVD eject key right in the main keyboard area, so i can hit it accidentally a lot and have it put the giant "eject" overlay icon on the screen and make the painful computer vomiting sound.

nice how Preview and Skim both lack a rectangular-region-text-selection tool so i can copy stuff out of PDFs with more than one column of text easily. instead i get all the text across the whole width of the page. not so ideal.

plus ca change, plus ca sucer. c'est vrai! mon dieu und mein leapen luger wienerschnitzel blinkenlights gott!

even software engineers need religion. and dogmatic fights.

fun: converting List<Integer> to int[].

(i'd rather just not use arrays at all, but on the other hand there isn't a good fixed-size list in java.util.* as far as i know? (duh, especially for primitives (and yes i know about the apache stuff) ;-))

is it just me, or is it completely freaking insane that Java doesn't have typedefs e.g. for the primitives? sure, i guess we are supposed to make a full class instead, but since they have primitives, they should probably have real support for using them in sane ways.

(and even if you were to make a class, you'd potentially get into the problem of how Java doesn't support delegation well at all. so then you use inheritance, which could screw up the whole point of having a typedef in the first place, namely to prevent the original type and the new type from ever being used in place of each other.)

some day either the tools will be simple enough, or labor will be cheap enough, or i will be rich enough to be able to have somebody/something create vanity custom programming languages. i really want something that works the way i think it should, and i know i'm not educated enough to actually accomplish that in any near future on my own.

trying to come up with good english naming for getters and setters in light of plurals and things sorta sucks in the long run / large / real work of rubber meeting road. i think maybe perhaps kinda sorta on the whole i'd prefer to use the Scheme-y thing of name-p for predicates, or be able to use "name?" and "name!".

those characters are not so usable in Java or Scala, however. suck!

i dunno how well it could be implemented, but i want a language or tools which support complexity analysis. that way when i inherit somebody else's code, or use a new 3rd party library, or even want to better understand my own crap, i can get some insight into how expensive things are.

it seems like there could be all sorts of interesting tools and tests and proofs and annotations and stuff one could bring to bear on the overall idea. wonder why it hasn't happened yet?

i guess the most related thing is proof carrying code. seems like that never got much out beyond academia?

notice: new-age ides will help easily check spelling. which will help anybody else who has to use your code.

java generics, boon or bane? i mean, when you have to start thinking like a compiler to grok something, i worry a bit. really great things should be understandable as abstractions and not require one to know "oh what is happening under the covers".

mmm, motion.

firefox kills me. i accidentally hit quit and it brings up the "clear private data" dialog box and i click on "cancel" and then restart firefox... and everything is gone, i have to re-log-back-in to everything. so i'm not at all clear on what the <expletive> point of the cancel button is.

mark my words, if you want to remain sane in the zillion-core world, you are going to have to not use (A) threaded shared mutable state, nor (B) software or hardware supported transactions, and (C) pretty much anything that doesn't have purity as a very top design approach. In other words, all those people writing regular Java are screwing themselves in the long run. Even people who are using "just" the JVM are doomed. Especially those who also try to use Java libraries, since they aren't pure.

i! am! sick! ("and tired!") of! amnesia! as! story! conceit! so i think we should make a game of self-trepanation, so we can have the catch line be "because you remember too damned much!"

mmm, visualization. tasty.

did somebody make a good paradroid clone, finally? hard for me to tell since it utterly screws the pooch when it comes to multiple monitors on my Mac Book Pro. 'cause you know nobody has multiple monitors in this day and age. nosiree. bleck.

well, finally somehow managed to get it seeing the mouse. overall the game lacks a lot of the je ne sais qois that was the fun of paradroid. guess i still need to do my remake (or convince Andrew to make it for XNA).

usability of programming languages is hard. dial it one way for metaprogramming and you screw up the basic readability and understandability of it at times. are languages which have a heavier-weight syntax than Lisp better, if they also have a way to do AST reflection and manipulation?

eh, screw all that noise, just ship some new vaguely-kinda-working features and score!

yeah, concurency is hard, and most everything sucks. it just bugs me that most folks don't seem to really grok that.

Quote:

Best answer: don't write concurrency bugs! ::: Use the 'immutable' object pattern / Use private data / Use well-tested java.util.concurrency.* )

Don't Go There But When You Must ::: Admit to self: Here Be Dragons / Design API around concurrent access / It's not a bolt-on after-the-fact check-list kind of feature / Think Before You Write, and / Document, Document, Document! )

My Take:

Maybe use something other than shared-mutable-state-concurrency?!

it's not the syntax, it's the semantics.

i'll reiterate: i don't think i've ever used a Y! product that didn't drive me crazy within 5 minutes.

and, while i'm on the subject of flash, what is it with flash library people (like papervision3d, sandy 3d) that they don't seem to want to put out compiled swfs, but instead expect you to install svn or whatever and download tons of crap? how can people be so weird, and hateful of their supposed customers?

wow. i will pretty much do my damndest to never ever use papervision3d simply because their web site (?!) is apparently designed to make it as slow as possible for me to get into the actual data. like, they have some hateful flash animation thing up front that i can't click through to, you know, maybe get to the downloads, or documentation?! and once you get through the first thing all you get is some hateful undersea thing that is just rotating like a drunken freak. nothing useful her at all?! haaaaaaaayte!

just say no sir i don't like it, and use Sandy 3D, please.

i don't like Thunderbird. but i guess i don't like Zimbra more. so i'm back to using Thunderbird. actually, i should probably give up on Thunderbird and just use Outlook. that's how much i think Tbird and Zimbra suck. can i even get Outlook for the Mac?

currently Thunderbird is killing me because - among other things - it doesn't scroll when new mail comes in so it always shows up off the screen; and it doesn't have a way for new threads to be kept expanded so i have to manually open them every time; and that sucks because there isn't a way to nuke a whole thread at once. just a nice mix of suck.

only three days to go before skyNet/neoconNet takes over.

wow. word. i mean, anybody who thinks their code isn't doing broken stuff under the covers even when things look mostly ok is obviously naive.

mmm, feedback.

hm. maybe things like forth could be good in that maybe they are so low level that you quickly learn to write your own words, rather than writing in raw forth. in slightly higher-level languages like, say, Java, i have seen code where the author didn't make a vocabulary, and instead made giant methods that repeat lots of complicated code. maybe they thought that individual statements of Java were already pretty high level so they were already working at a sufficient level of abstraction? i dunno. but it hurts.

blogger sucks. word press sucks. typepad? typepad sucks, too. i go to leave a comment and it takes for ever for it to bring up the captcha, and then it takes for ever to get back to the post with the comment. a horrible user experience all 'round, every time i've tried to use it.

wow. word!

i don't really fit into this era because bad things drive me nuts, and yet i don't have the time to do everything let alone do it right, so i have to hope that other people who make things will do them right, and then they don't, and so i'm stuck with nothing being right. (sure, doing things right is hard.) basically anything that has any kind of user interface: software; interactive voice systems; the dmv; modern architecture and product design a la design of every day things; etc.