Wednesday, April 30, 2008

yes, usability is hard. but i still freaking hate gmail. of course, i don't hate it as much as anything Y! does since the ads aren't as bad, so i stick with it even so. just! can't! win!
every time i have to use the cygwin installer, i weep with sadness. what a horrible horrible horrible user experience. as if windows wasn't bad enough, somebody had to bring the wonderful usability one associates with unix over to it. so now you get the crap that is windows-usability-think multiplied by the crap that is unix-usability-think. not all that much of a win. (but better than having to really learn dos batch files, i guess.)
compared to having to do it on your own, maybe maven isn't so bad after all. the fact that people in this day and age have to thrash around at all to get things installed and building is pretty much utterly pathetic.
i think the claim that Scala ports to .Net is kind of bogus, given my experience trying to get it to interface with MSFT DLLs, too. so just don't plan to use it for anything other than a JVM. (or via something hacktackular like IKVM.)

Tuesday, April 29, 2008

i tried "cry of the infected" demo. it was big, slow, bloated, and just overall amateur. i guess the "fps creator" engine isn't that great and/or the developers don't know how to use it well. besides the horribly long loading times, really juddery frame rate, and overall graphical suck, it also had great things like requiring a mouse wheel to change weapons - but i'm playing it on a laptop with the trackpad.
the old amiga classic game Stunt Car Racer was good because it was so on-rails. modern stunt car games sorta suck because it is just plain hard to get such a physical 3d thing working via a narrow-view 2d flat display medium, i think; one spends more time out of control having no fun and just being frustrated.
knytt stories has a great look and feels. but the game itself? can you say boring shoe leather?! sheesh.
i never have, and i think i never ever ever ever will, understand why anybody thinks the Tower Defense games are even remotely like something one could possibly label as "fun".
i like indie games. well, some of them. funny how the ones which seem to be trying to be commercial sorta suck more often than the ones which are purely done for the love of it. like, the Zombie Shooter shareware demo was painfully bad (it didn't even have an option that showed you the keys or let you rebind them). and the Tank Universal demo sucked a lot in terms of loading screens taking for ever and in terms of not being able to rebind "fire" and in terms of the motion of the tank being, for me, utterly wrong (the tank should not keep moving when i let go of the motion key, it should slow down to a stop). maybe the commercial people think it is "close enough" or that the only thing which matters is graphics or something, i dunno.
i just had a lovely usability experience in Firefox 2.0.0.14. i was downloading a file, and 2 others were already done. i clicked on the "open" link for one of the ones that was already done, and instead of opening it, it was removed from the download list. i think what happened is that the progress bar for the newest download went away a split second before i clicked, because that download had finished, and so the window was redrawn with everything shifted up a few pixels, so that ff thought i had clicked on the "remove" link which is just below the open link. in other words, a better design might be to have the links laid out horizontally rather than vertically.

Monday, April 28, 2008

i still hate blogger.
any streaming video player which does not show you how much has been buffered is just utterly insanely wrong evil bad. hard to believe such things exist, but i've seen 'em.

[correction: i am wrong; it does show it. only it is really hard to see the difference, especially on a slow connection which is precisely when you need the difference to not be so hard to see.]
people seem to like to say that software is the worst thing on earth when it comes to success vs. failure stories. it is probably true. but other hobbies sometimes fare similarly poorly.
i would like to reiterate that Adobe Media Player is hardly the best advertisement for AIR. it is slow and kinda crappy. even on an only year-old dual-core laptop. maybe they got paid by Intel to make it painful to try to convince people to upgrade? bleck.

Friday, April 25, 2008

if yer a nerd, consider adding your wisdom for the greater good.
ctm in fifteen minutes.
sure, the world is going to hell in a hand-basket, and most everybody on earth is going to starve. but it could be worse, no? like, uh... uh... hm. dogged, maybe it could only get worse.
some leaders are actually willing to be leaders and pony up responsibility when it is time. others? yeah, well, others are ass hats who shrug off the miserable train wreck they rubber stamped all along, fire a couple of underlings, and continue on in life. and you wonder why people need to believe in heaven and hell?
it bugs the crap out of me that there are things like wikis which make it easier for people to put *words* on to the web, but there aren't great tools which let people make pictures. of course people probably would suck at making them anyway. and making pictures tends to require some nice application, and then takes up lots of visual space on the screen.

but anyway i was sorta wondering if there could be a way for people to have an oo drawing type program which supports both mouse interaction and also textual interaction almost like playing an infocom game. somebody smart & experienced could maybe come up with some simple vaguely natural language vocabulary for it for domains. e.g. if i'm describing a web system, i should be able to say "the FrontEnd connects to the Database via HTTP" and those get automagically turned into a picture. and then you can have vocabulary for moving things around adding removing sizing labeling whatever. and maybe it all also has some kind of underlying skeletal system with reverse kinematics so that you can have some things be linked and kept in relation and automatically laid out (yeah, as if that is easy) as you move 1 item around so things stick together nicely.

ok i'll step away from the crack pipe now.
it bugs me that Scala doesn't have an imperative for loop, and yet there are all these examples from supposedly official Scala types which use "var" + "while" + "increment". i mean, take a freaking stand already: either have imperative for loops (since you are already making crappier versions of them with the while), or get rid of that mutation type stuff entirely and e.g. only have tail recursion or whatever.
personally, i think extremes are bad when it comes to education. different students in the same class might learn best via different paths. in the case of the study, my conclusion would be that instead of throwing out all "concrete" examples, i would show 2 small concrete examples in sequence, side-by-side on the blackboard, and then follow up with a 3rd phrasing in more abstract terms, next to those 2. the point being that we have to learn how to bridge the gap between the abstract and the concrete problems, too, no? and if you could have the 2 concrete examples be given in textual and graphical forms, that would be good as well. the more ways to catch the insight into the problem the better, i think. of course you have to beware overloading people with information, so the problems need to be very concise at first.
if your programming language doesn't have:



  • concurrent GC

  • a REPL

  • a great debugger

  • great support for delegation (has-a)

  • a sane multiple inheritance story

  • support for overloading operators (really)

  • static typing with great automatic type inference

  • support for local functions

  • first-class functions

  • ideally, a great macro system (see PLT)

  • a way to support RAII/finalizers

  • a really good macro system

  • being able to write big numbers with commas/periods (and be locale-aware in some non-suckful manner)

  • light weight scalable concurrency (e.g. csp or actors or something)

  • hot code replacement

  • something better than try/catch/finally, maybe scope-exit

  • named arguments, to support optionality and reordering

  • good handling of non-null-ness, and Maybe/Options

  • multiple dispatch of some non-lame-in-the-long-run ilk

  • metadata support

  • i guess i forgot to mention macros ;-)

  • all on top of a nicely portable (e.g. written in C and then bootstrapped in itself later i guess) virtual machine

  • typedefs / type aliases

  • parameterizable modules, something like the PLT stuff




then it kinda bites. of course, that doesn't mean it is useless. far from it, especially since it is apparently nigh impossible to get all the semantics i want working together. i mean, i sure as heck don't know how to do it, and from what little i read it seems like just getting anything real up and running and robust and usable at all is bloody hard work even for the well educated.

(and if your blogging system can't freaking format text around bulleted lists nicely, but insists on having a mile-wide gutter between 'em, it sucks butt too.)

Thursday, April 24, 2008

i think if you tend to use logging instead of a debugger, it could perhaps indicate that your source code is too different from the runtime behaviour. so when you stop and look at things in the debugger you don't have the information you need to know what code path was taken. so then you think debuggers are useless, and you go back to logging for all debugging. check to see if the functions are long and have lots of conditionals in them. rewrite them to, basically, suck less. then debuggers will be more useful :-) to you. not that logging goes away entirely, but that you should be able to make use of more tools, and get more maintainable code in the process.
you can take the functional approach to crazy syntax limits as well, but every time i have to go back and look at my old Java code i'm like blinded by all the boilerplate and indentation and looping and just crap. calgon o'caml scala, take me away!
gosh, maybe Make can get somebody to teach them how to make an ecommerce site that doesn't suck stinky unwashed buttocks?
my brain!
is it me, or does perl 5's s// not have a "consider the source to be literals, e.g. don't see parentheses as being grouping" modifier?
shove fibonacci up your nose, although without always handling the negative case?

Wednesday, April 23, 2008

sure, Scala probably sucks less than Java. but it sure ain't perfect. (not that anything ever is, of course.) currently i'm frustrated with the fact that traits can't have parameters to them. blah. verily, C++ still has some things other people could learn from, no?
sure, we can try to make things like concurrency more simple, but things are always screw-up-able: computers are hard.
i think i've come to the conclusion that 9 times out of 10, if you make a function which takes a boolean parameter, rather than having 2 functions, you are kind of a jerk when it comes to understandability and maintenance.

well, unless you at least put in some typdefs/constants which are named well to represent what the true and false values might mean.
in java there are primitives and wrapper classes for them. and now there is auto un/boxing. so maybe the idea is that you could use them more interchangeably. but consider "==" with the 4 different possible types (primitive,primitive) (primitive,wrapper) (wrapper,primitive) (wrapper,wrapper) although two of those are redundant (they would not be if we were talking about equals()). so you really do have to know what the types are and worry about the semantics. oh well.
it isn't just me and this mac book pro. other people at work with similar machines have also been experiencing some really impressive wedges, crashes, hangs, and general freak-outting of their machine. so much for the bally-hoo'd feature of oh it runs on unix it must be perfect.

update: oh, great! that's number two for wedges today! both times while using the "cmd-tab" application switcher.
adobe should die and fail and die and death and stuff. that's how i feel after using just about any of their products, in particular when it comes to updating things. lots of windows coming and going, stealing focus from what i'm trying to do in the mean time, and preventing me from running my browser while it is updating. all a truly wonderful heart warming fuzzy fun user experience.
you are an ass-hat who deserves to go out of business if you advertise something and then have a broken link to it and then say oh "we are having minor issues with our content management system" as if not being able to get to the content which you are trying to 'sell' is a 'minor' issue. aaaaarrrrggggghhhhhh!

Tuesday, April 22, 2008

i love how adobe flash programs seem to often put up a window that is "always on top". e.g. loading splash screens and stuff. some even have the regular native ui chrome, but you can't get another window to show up "above" them. so that's a great thing in the world of usability.
i like disposing/RAII, so it bugs me when languages don't have them. i guess if you know what you are doing in a real language which lacks them you can add those features yourself. there are problems with all that, of course; for example, that you can end up with a bunch of different implementations of the same idea (witness the insanity that is php).
i'll mix up some blog posts i've been re(a)ding: "Red is the color of a lot of good things."

Monday, April 21, 2008

war is not the answer. shopping is.
working for The Man is all it is cracked up to be. fortunately, the coming apocalypse will sort us all out, no?
i'd like to read this, but eh, can't be bothered, really.
so, i guess trains have never run on time anywhere, after all. so sad.

Sunday, April 20, 2008

those wonderful words which tell you just how great and perfect a thing Cygwin integration is: "couldn't allocate heap".

Saturday, April 19, 2008

go go gadget embedded systems, with a free book an' 'at.
scala i guess is nice, but since it is predicated on the JVM i am not at all sure it has the power to let me do things really immutably the way haskell would. can the JVM handle that much garbage?
sorta like how you can lookup who uses a method in various IDEs, it would be neat if one's development system automatically maintained documentation which pointed back along a relationship. if functions are higher order and are put into lookup tables for event handling, for example. sorta like how i would like a system that has type inference but which inserts comments about what the types turned out to be, so that when one goes back to read the code the types are in there.

Thursday, April 17, 2008

java sucks: the Random class doesn't even easily let you create things e.g. in the complete range of Integer.MIN_VALUE to Integer.MAX_VALUE?!
a perhaps sometimes useful cheat sheet for jvm args.
macs would suck just a little bit less if java wasn't horribly broken and out of date in various ways on them. but no, i guess they need to continue to stuck?
i very much appreciate that other people are working hard and maybe not writing documentation is most efficient for them, but it sure sucks to be an end user who just wants some basic API docs (to answer basic questions like, "are those parameters floating point, or integers?"). to me when somebody says they won't document it properly but you can use the forums, that really kinda grosses me out. :-(
i hate firefox. if you got to foo-bar.org (i don't know what that really goes to, it is just an example) when you wanted to go to foo-bar.com, then the next time you type in foo-bar the completion will be .org and apparently you can't get rid of it?! maybe short of nuking all types of history?! argh!
i am continually stunned by what a piece of excrement word press is.

(don't get me wrong, i freaking hate blogger too. but it seems to not be as utterly obviously broken in all sorts of ways the way word press is.)
i've pretty much always thought that road kill was a pretty simple and clear indictment of so-called progress. not that i don't drive to work.

Wednesday, April 16, 2008

i tried Adobe Media Player on my not-completely-lame laptop (TravelMate 8210-6245) and it mostly worked (although I didn't actually watch any video), but it was a bit sluggish. I don't like it when computer imagery has artificial motion blur at 2 frames a second, that's just freaking stupid lame dumb sad.

i then also tried Joost Beta 1.1.4, and it brought up a window that was just utterly black. if i moved the mouse around, the cursor would change so presumably there were widgets being drawn and all, just that everything was drawn in black. way to go! so i uninstalled it and will i guess never want to use anything by them ever again, really.
if you are stuck with shared-mutable-concurrency (which is basically a no-win situation, as far as i can tell) maybe you could have your system do some basic automatic locking: in a method which touches a member variable, that method is automatically locked around that variable.

might not be much different than just making every method synchronized by default? so maybe just being able to mark a class as synchronized and have that cause all methods to be synchronized? i dunno.

i'm just thinking that in any system that is complicated, there should be an easy way to start off with the safest possible thing, and gradually tweak it for performance after it is really working.

the point is to avoid races. the fly in the ointment of all of this is that locking is of course dangerous, and you can end up in deadlock. which is back to what i said at the start: concurrency probably ought to be weilded via some other approach in the first place.

so forget all the hot air above, and just use Erlang / Scala Actors / JCSP / STM.
it is too complicated and proprietary to post the details, but i think i just saw both the Eclipse and Java compilers utterly fail to pay attention to the concrete type in the use of a generic Collection. e.g. it was supposed to only hold Foos and there was a line of code calling remove( Bar ) and there were no compilation or runtime errors?!
Eclipse colors variable names differently so you can see what is e.g. a local variable. But it doesn't (apparently) distinguish in the case of having a locally defined anonymous instance which uses 'parent' variables. Er, if you get what I mean.
you can tell i'm a big fat nerd because this makes me excited as a leetle girl on SNL. especially because Scala-MSIL doesn't really work for me.
i especially like the feature of this mac book pro where the menu bar stops working. completely. clicking on it does nothing. no menus. for any application. w00t! what is interesting is that the menus on the left of the title bar are hosed, but the ones on the right (e.g. clock, spotlight, adium) still work. wtf?!
wow. buzzword.com is totally the kind of UI experience i hate. guess i'm old.
i like how easy it is to be logged into Google apps with more than one account at the same time. oh. wait.
oh those nutty nutty soviets and their screw-driven vehicles.
things get better when they become higher order. e.g. in Flash CS4, animations are now editable entities themselves.

Tuesday, April 15, 2008

i use sonic.net for my home DSL. they don't totally suck; their customer service on the phone is good (i didn't like speakeasy very much that way). but their web site is kind of a joke when it comes to usability. oh well.
mmm, aircraft. i loved having books with pictures like that as a kid. (still have those books, somewhere.)
i wouldn't want to oppress anybody, preventing them from expressing themselves, but personally anything described as having syntax "based on the Pascal programming language but allows for a BASIC-like syntax to be used as well" is going to be kept at very arms length by moi.
i dig through the various "healthy" "energy" bars. most have some form of partially or fully hydrogenated oil, so forget that stuff. then i find one that seems ok, says it is mostly organic and all that jazz. of course when i go to actually eat the thing it turns out that it tastes like somebody's lemon zesty bathroom air-spray cum anti-perspirant. puke!
how can somebody who always sounds so sane be involved in such insanity? i guess it is just he ultimate expression of a pragmatic heart. at least, if somebody else had been in his place we might have ended up with much worse?
and you thought single sign on was a good idea. yeah, well, the bottom line is that usability and security/safety seemingly only rarely go hand-in-hand.
oh the love that is marketing weasel words.
how to make software (not just games) as if people actually mattered - gosh, what a novel idea!
i love using financial web sites which tell me i have $X available for e.g. buying stocks and then i try to buy stock and it tells me "your limit it $Y" where $Y << $X. what? the? eff?

Monday, April 14, 2008

"hell is other people." especially at work.
seems lame to me that you can't move a tab into a window or a window into a tab in Firefox.
i expect people not to "get" how to do email because it is new. but what i'm seeing is people who apparently do not "get" email because it is too old; they are doing IRC and IM instead. oh, Alvin, you ol' prankster, you.

Sunday, April 13, 2008

word!
software is hard. and doing it right is somewhat laborious. but that's the level of effort we have to commit to if we want to really not suck. right?
people use comments to help them understand unfamiliar code, as well as the other reasons they comment things. i think it would be neat if people could have their own local comments that other people don't have to see. that way i could really explain things to myself, and not have to bug others. also, i could put in more swearing.
i hate java. there's WeakHashMap, but on WeakHashSet. and trying to make your own by wrapping a WeakHashMap sucks ass because java sucks ass when it comes to has-a (instead of is-a). argh! i need a day job using Scala instead.
you can't please all the people all the time, and software is hard, but i wonder if we'll ever get enough of a handle on complexity and usability to make the ultimately comfortable programming language?

Friday, April 11, 2008

for those of you (ha ha) who might have thought i got over my hating of Eclipse... grn.
where good intentions and experience are not enough; sounds accurately like entirely too much of the world to me. i think part of it is that people do not like to question the status quo, and if you do you find yourself getting signals that you are seen as kinda crazy.
basically, since all code sucks, we need the ability to sandbox at any level; it is nice to be able to kill off some (perhaps arbitrary?) portion of your system and know that there aren't any leaks. of course that can completely screw up your internal state, so you'd have to write your system expecting that from the get go. hm, now where have i heard something like this before?

some tricky things are also getting the granularity of the sandboxes right. and being able to set them up / change them in flexible ways.
when virtualization goes bad.
i have to say, i am constantly impressed with how much of an utter EPIC FAIL most technically related pr is. even stuff which i guess is supposed to tell you "what does it do? how does it do it?"

whatever.

especially ass-hats who require you to register to read their docs. sorry, but given the choice, i'm just going to go with some open source solution that isn't so close-lipped and apparently hostile to me being interested in evaluating their stuff.
some day, we will be. it has already begun.
if you don't let people benchmark you (Microsoft, IBM, etc.) then you deserve to be spat upon and go bust.

Thursday, April 10, 2008

hey, it is my bloody tax dollars at work, i should at least be able to get a laugh (and probably more than just a chill) from it all now and then.
because we wouldn't want to actually use the type system to make actual types for the things we need types for, no, instead we'd rather make lots of enums. yeah, that's great.
i like how it is apparently mostly impossible to go from a url for java docs for a given version on the sun site to any other version via any obvious editing of the url itself.
hot, hot, steamy... financial collapse, baby! coming your way.
the current mac thunderbird dmg installer graphics are stupid because the applications folder is not a link to your Applications folder. i mean, duh!?
shyeah, right, as if that's real. (actually, it probably is, and it sorta gives me a heart-attack.)
whoever it was that did the whole system for upgrading efi firmware on macbook pros must not have worked for apple. the experience was more like something from the unix or microsoft world. including the really loud horrible beep from hell. bad enough for a single person, but also great in an open office cubicle environment.

Wednesday, April 09, 2008

i <lurf> debugging via Eclipse / on the JVM. if you are in a constructor, sometimes (often?) it can't show you values of things which have already been done in the constructor, because you aren't done with the constructor. e.g. in Eclipse i'm getting "org.eclipse.debug.core.DebugException: com.sun.jdi.ClassNotLoadedException: Type has not been loaded occurred while retrieving component type of array."
no, i don't do a good job of this myself. but maybe i'm gonna start trying: if you don't have proper type-checked types for unit (e.g. seconds, inches, whatever), you should at least put the unit name into variable names. yessir.
there's a gamut of quality of error checking in programs. one of the influences is who wrote the code. some people are more aware of and/or worried about errors and how to deal with them. some... aren't. even when one does pay attention, it is hard to get it all right. it requires, you know, thinking. and testing. and stuff. eh, whatever; just ship it.
if the main way you judge your code is by if you yourself understand it, then you are not doing the best job you could. ditto if your group doesn't give you the opportunity on a regular basis to find out if anybody else understands your code.
my new little song i sing to myself: zimbra sucks, zimbra sucks, zimbra sucks, zimbra sucks!

not that it sucks that more than any other product on earth; everything sucks, ya know?

[although zimbra does do a good job of ignoring the features of other similar products, like, it doesn't seem to have a decent document auto-save, from my recent empirical experience. nice!]
if it doesn't get reported, it doesn't exist. if one person reports it, it still probably doesn't exist. we certainly wouldn't want to follow up on any leads towards other people who might have noticed it!

Tuesday, April 08, 2008

i will never condone starting (0,0) at the top left of the screen. i think that is only done because that is historically where the first memory for the screen was located? the one true way is to have the origin be at the bottom left. really!
if a given floating point implementation accumulates errors, then why doesn't the given language support setting an epsilon for all floating point comparisons? hayte.
you know the world has become a hell on earth when you go looking for help on how to disable tooltips in a program and all you can find are people talking about how to enable them.
ok, i completely freaking hate java. the fact that you can't access static fields in enum constructors is utterly hateful; it completely breaks any claims of "oh enums are just classes and will make perfect sense". i guess i would have at least told people up front when introducing enums that they are created/initialized before statics - that makes sense when you think about it, but if you've never heard that it just sucks buttocks. grn.
yeah, it sucks because software is hard. and because, well, people suck.

Monday, April 07, 2008

"favor delegation/composition over inheritance." and yet a lot of the supposedly OO languages make it a completely freaking nightmare of suck to do anything even remotely like delegation. having to manually write all the wrappers is not my idea of fun or a good approach to maintainable software. hello, java, c#, c++, scala?! it makes me mad that languages are so full of crap.
so since most programmers suck (and some really really really suck), i think we should as a society and industry be requiring people to use pure/strict languages. it should, i'd hope, lead to less crappy code because people would have a harder time going nuts with side effects. well, ok, maybe they would just learn to hack monads into everything? why isn't Concepts, Techniques, and Methods de rigeur? (er, and, why haven't i found enough free time to finish reading my copy?!)
more fun for kids of any age.
how crazy will it be when we first find signs of et life?
just so you know, zimbra sucks. just like everything (e.g. gmail) sucks. only in new and special ways.

Sunday, April 06, 2008

pfSense was nice and small and fast to boot. the installation tripped me up until i noticed the one letter difference between "WAN" and "LAN" when it was asking me about ports. then, it turns out their web UI is fixed width and does not work well when you've set your browser font sizes to be anything other than the defaults. and then for some strange reason while it could route from inside to www.yahoo.com, it failed to do so for search.yahoo.com.

so then i switched to untangle. the machine i am running it on is less than half the required specs in every way! well, except for having 2 ethernet ports, that is sort of a non-negotiable thing. so it is something of a nightmare to start up the system, and it is something of a nightmare to run the java ui on it. but once it is running if i just do all the administration remotely from my newer laptop then everything is pretty copasetic. it even routes successfully to search.yahoo.com.

so untangle wins.

i do have to note, however, that the untangle ui is a bit on the discombobulated and just plain silly-forced-metaphor side.
look, i've done this myself, so i know that most geeks are idiots when it comes to explaining things, with some kind of probably worth formally studying for somebody looking for a masters in psychology project kind of problem. not only are they idiots, they are jerks to boot. i think it is sort of related to the geeky cluelessness of usability, too.

Friday, April 04, 2008

my theory: when cats yawn, you yawn. but when you yawn, cats don't yawn. obviously, they have the upper paw.
mmm, penguins.
who am i to judge, really, but verisoft sounds freaking cool to me. i mean, all software sucks, so any tools we can use (we'll ignore the who's watching the watcher turtles all the way down aspect of the fact that the tools are written in software) to get a better handle on all this shyte are good. uh, gosh, well, i mean, if everybody (anybody?) actually used them, of course. ahem.
you know: for kids.
i just keep returning to that paper. one of those things which makes entirely too much sense, and yet it we have super popular main-stream languages which flub the pooch, so to speak, over the whole thing.
ok! it turns out that while you might have liked his game show, in actual fact Ben Stein is a Butt-Hole.
ok, one can learn a lot about psychology from filthy rich successful people, it turns out.
if what you are doing is trying to solve things as they are now with an eye out for say (at best) one year, then you are going to end up 15 years later in the same-but-worse situation. you really do need to find a way to go back to first principles!
in my dictatorship, there is a circle of hell specially reserved for people who make products that go overboard on the bloody tool-tippage.

Thursday, April 03, 2008

i've said it before and i gotta say it again: if you have closely related terms in your program which differ by one letter, you are asking for hurt, and frankly you deserve whatever you get.
complete and utter nerd genius!
could i be more schooled? "'The fox has many tricks. The hedgehog has but one. But that is the best of all.' However, Dimitris Vardoulakis has pointed out to me that Erasmus himself quoted this remark from Archilocus (c. 650 BCE). So here is something appropriate that Erasmus actually said: 'If you keep thinking about what you want to do or what you hope will happen, you don't do it, and it won't happen.'"
i! gave! up! and! went! to! google reader. (which actually still! really! kind! of! sucks! only a little bit less i guess.)

whimper.
web apps, shmeb apps: if you can't support right-click, then it is a piece of excrement.
i guess i completely fail to appreciate feeds from blogs where the feed is not the whole post. that doesn't make me want to click-through, that makes me want to unsubscribe and just hate you for ever. yup.
well, i'm going to give bloglines (their new beta) a try for a while, anyway. the main problem with blog reading is that there is entirely too much stuff to actually read. let alone then go and do based on what one reads. i mean, cooking alone could be an entire life. i think we really need to be able to clone ourselves and have the clones+original all wired up in a hive mind.
i am trying bloglines. good thing they make it hard to actually find and add feeds. yup. great first user experience, there.

(i've been trying to avoid just going with google reader in the hopes that somebody else doesn't suck, but at this rate i'm going to probably have to just cave in and try them out and hope they aren't completely dumb-asses with ui.)
just like netvibes, newsgator utterly fails in the usability department, especially when it comes to my custom font sizes. great. i guess this is why i never try anything 'new' on the web: i have years of experience that tell me i am just going to freaking hate it all within 5 minutes of trying it out.
i'm trying netvibes ginger. already they've managed to cock-up the user interface even just compared to what they previously had: when you delete a widget, you now have to click through 2 dialog boxes, not just 1.

gosh i hate everybody.
"at alltop, we go the extra mile to put stupid hateful ui on the page, making sure it covers up stuff you might want to actually see. yup, that's our contribution!"
if your web site has those pop up things when i hover over a url there (e.g. word press snap shots), then you are a dirty butt-hole of usability as far as i can see.
pandora sucks. their db thinks that since i like kmfdm i must therefore like all sorts of utter gothic-metal trash. (i know, maybe everybody thinks kmfdm is trash, but i don't. so they are just pissing me off as a potential customer; not a good way to do business.)
"As Frederick Brooks wrote over 30 years ago, 'Show me your flowchart and conceal your tables, and I shall continue to be mystified. Show me your tables, and I won’t usually need your flowchart; it’ll be obvious.' Or: data structures live forever, code changes."

Does that mean OO sucks?

Tuesday, April 01, 2008

could we be any more doomed? i think, almost, not.
the usa is no longer anything remotely like a land of the free or brave. just not even remotely.
where and when and how will i ever get enough reading time to really be even remotely up to date?
all i'm saying is, your demo shouldn't start off by showing required verbosity; that's a pretty glaringly big turn-off!
pretty sweet when non-Sun people make tools which Sun, if they were actually responsible folks, should have done themselves from the get-go.
wow. people with a clue! nice.
you know you need to learn forth.