> I've noticed a decent improvement in how many monitor related headaches I get from switching to an LCD.
I'll 2nd that, so MOD up.
LCD + DVI = Fewer Headaches!
> Is there a scientific reason why looking at a monitor would decrease eyesight?
The focus & clarity is higher on a LCD, so your eyes don't have to work as hard. The "pixels" on a CRT are not really "one" R,G,B triad, but a "fractional" multiple. Result: Blurry Image, and Eye-Strain.
True, you loose a bit of gamut by tossing the CRT, but a softer image on the eyes is worth it.
I also found: Contacts = More Headaches. Glasses = Fewer Headaches.
Your eyes (like the rest of your body) need oxygen.
The wallpaper radio button is a very good idea. Simple, and effective.
You'd be surprised at wallpapers. The bigger the company, the more likely someone will have an interesting wallpaper, someone sees it, which sparks a discussion, and people get to know one another a little bit better. Who _doesn't like getting complimented? i.e. "Hey, thats a cool wallpaper you got there! Where did you get that?"
* The fact that it is impossile to distinguish between ARRAYS and POINTERS causes all sorts of problems. * There is no native NULL pointer * There is no way to DISABLE the automatic up-casting of floats to doubles. (This SUCKS for console programming.) * No way to prevent integers being promoted to FLOATS. (SUCKS again, for console programming)
Counter-examples:
void func( const int v[3] )// init from vector
void func( const int v[4] )// "already defined"// won't compile since arrays == ptr
void func( const int n )
void func( const int *aArray )// func( NULL);// doesn't call the pointer version!
I _really_ wish there was a way to declare/define arrays, and functions via keywords. If C was consistent in reading from Right-To-Left, you wouldn't need the typedef hackery to define function pointers, etc.
i.e. func( int ) *pFunc; int* aPointers[3]; int array[3] *pArray;
That said, I love C++, because it is a pragmatic language.
2 complaints: Flash, and no links to the images.
on
The Cosmic Year in Review
·
· Score: 3, Informative
Would help if they didn't use Flash, and actually had links to the Nasa site, so we could save them for our wallpapers, instead of clicking taking you to the next pic.
That question is ambiguous. Best at _what_? I would say, none, at the moment. They all have their strengths and weaknesses.
I'm tired of almost every MMORPG having XP = Monsters Killed. Games need to let go of the 1970's (classes, and a warped view of what eXperience Points are), and embrace the new millennium. i.e. I can't be a pure crafts{wo}man in WoW without being a fighter. Sadly, you don't get XP for crafting, only killing and exploring. UO had the right idea with trade skills. If you can live with this "restriction", WoW is a good, fun, combat game. It's a much better (streamlined) EverCrack.
The next "big thing" in MMORPGs is Interface. WoW's interface isn't revolutionary, but shows the incremental evolution.
The next "big thing" after that, is Game Scalability. UO had the right idea with people being able to own their own Homes, Towers, Castles, etc.
Think of a MMORPG Sci-Fi game where you can not only own your ship, but with enough credits, star-bases, and planets. A viritual world where not only combat, but trading, crafting, and entertaining are all viable ways of enjoying the game, giving the player interesting choices, not spoon-fed "advancement", is the "best" one.
-- "You can sell your values, but you can't buy morals."
-- _Michaelangelo_
> Make the best game and apparently 5 million people will come.
WoW is definately _NOT_ the best MMORPG. (I've been playing since the Beta.) What it is though, is better then 80% of the OTHER MMORPGs -- most of the annoying problems are addressed (fortunately mine-jacking isn't too common.)
To give a poor analogy:
If you have a few normal looking woman, and an better-then-average looking womans walks by, she looks beautiful in comparison. Wow is the better-then-average looking woman. It's all relative. While you can certainly have fun with the "normal looking" woman, most people will be naturally attracted to the eye candy.
WoW got the basics right; If you don't make the first 10 levels enjoyable, you're never going to have a customer base at lev 60 to support you. Something that EverCrack doesn't understand is that I _don't_ have 4 hrs to waste raiding.
And yes, WoW is definately inspired by Diablo II -- the Skill Tree, er Talent Tree has 3 tabs. Talents have dependencies, and level requirements.
It's not rocket science that WoW is so popular -- just good gaming sense. WC3 is not revolutionary -- its evolutionary. It's taken the best (better?) ideas, and gone with them. Same with WoW. For every MMORPG failure, there's a lesson to learn on how NOT to do things.;-)
Anyone know if the compressed images leak has been fixed?
i.e.
Firefox (on Windows) can and will suck up an infinite amount of memory. This is because under some circumstances (Well, always, at least for me and many other users) it does NOT remove the uncompressed images from memory when a tab is closed.
You can get a ~ 50% speedup if you change one line in the Swap() section, which is kind of ironic, since you said "Bottom line, my register usage makes all the difference."
arg[i] = arg[point]; to
arg[i] = small;
If you got an email addr, I can send you your code cleaned up for readability, and a complete testing program.
Loved your insights, and matches my experiences. (I haven't seen any FF code either.)
i.e. There have been a few times (3 or so) that FF has completely stalled. Attaching the debugger and seeing that it is "stuck" inside a few functions, and forcing them to return, has brought the browser back to life, complete with my 20+ tabs all intact, and allowed me to save/bookmark them all.
For me the biggest complaints with FF have been... 1. Memory usage (bloat with pre 1.0.7 was terrible.) 2. No ability to start a new window in a different process/thread. I believe the WinMain code is using a named atomic, and the new thread terminates (to prevent memory usage duplication) if it sees the semaphore already exists. Wish there was a command line option to force it by-pass this check. 3. Un-responsive UI (also much improved, but still stalls occasionaly)
-- /. sucks @$$ for posting code snippets. At least let the registered users in good standing be allowed to bypass the lame "lameness filter."
> For the same reason we send robots into hazardous environments - it is too dangerous to justify sending humans.
That's a good point; you only hinted at the other half of the equation, IMHO.
Robots let us get the procedure predictable, so when it actually is time to send people to space objects, that we're not doing it for the first time. The ratio of Reward/Risk is much higher with robots then with humans.
i.e. What band puts on a concert without practising?
> How will we justify the use of this when so many people die very young from preventable causes that are beyond their control (as opposed to simply not taking care of oneself)?
You're under the delusion that their are accidents. There are none. Everything that happens, happens for a reason.
> How will we prevent the extreme accumulation of wealth that this would allow if it is not equally accessible to everyone?
The same way we prevented it today. Oh wait, we didn't.
-- /. sucks @$$ for posting code snippets "Your comment has too few characters per line" ?? /. eats extra whitespace in "code" -- any workarounds?
> I get really irritated when developers check in code with sections commented out. If the code is no longer relevant - delete it! If you really need to get it back, get it from a previous version in the source control system.
I assume you're not talking about #if... #endif sections.
How many times has code been cleaned up, and ended up broken? It's a PITA to: (1) rename / resave the current version as "new" (2) fetch the old one (3) diff When it is easier to read the "old" and "new" code, intermingled.
Sometimes the old code is left in, as a "fallback" -- since the new code is still in a state of flux, and hasn't been thorougly tested. A good example is optimizing a function, piece-wise. I found it best to: (1) have complete copy of the original function commented out (2) have a "working" copy (or is that "not-yet-working" ?;-) of the new function, with commented parts of the original that are in the state of being converted.
But yea, once the stability has settled down and the new code is working, I agree. Nuke the old sucker. Unfortunately, once code is checked in, people tend to forget about it. "It's working, right? Why clean up cruft?" lol.
-- /. sucks @$$ for posting code snippets "Your comment has too few characters per line" ?? /. eats extra whitespace in "code" -- any workarounds?
If you use proper descriptive variable names, you don't need comments 50% of the time! Why? Because if you can't name something properly, chances are, you probably don't understand the problem fully.
Another good idea, is that many people use "notes" in the code that you can search for, such as:
// TODO:... // BUG:... // FIXME:...
Here are few examples with commentary:
int nBitsPerPixel;// Look ma, no comment is even needed!
int bpp;// Bits Per Pixel -- plus now I need to remember what the hell "bpp" stands for. Sure this one isn't TOO bad, but this is NOT FUN when people invent their own non-standard acronyms.
Callback_Mouse();// Gee, I wonder, is this a Callback?? mouse();// who calls me? and what does it do? Set or Get mouse paramaters??
Any half-decent editor will have auto-completion, so typing in long names are a thing of the past. Plus you don't have to remember _cryptic_ abbreviations. Dealing with Unix's naming convention is bad enough, don't make me deal with it in code too. (What, someone couldn't think to use the names:/user/temp, instead of the inconsistent:/usr/var ??)
Coding Standards ================
Whatever standard you pick, PICK ONE. A bad standard is better then no standard at all. and The nice thing about standards, is that there are so many to choose from.
When dealing with Physics or Math calculations, it always very handy to keep track of the "units". On the HP calculators, they use "_{units}". i.e. 299792458_m/s. In code, this isn't practical -- as the code tends to become too verbose. It would also be great if we could see the "scope" -- local, global without extra clutter. What kind of system could we design/use that is practical?
Now note that when you take a standard to the logical extreme, you end up with something impractical. Applied to this domain of programming we find that absolute abonimation called Hungarian notation, micro-managing every little detail. (Do you really need to clutter up the variable naming detailing the internals if a string is zero-terminationed?? NO!!)
However, if you use something more PRACTICAL, you end up with a very nice, compact, naming schema.
Prefix
a arrays / tables
b bool / flag
i iterators, along with the object type. i.e. iWord
n number of, or size of. i.e. nWords, nBytes, nBits,
r reference (not really needed, but some people like them.)
p pointer (handy with pSrc, pDst, pTmp to indicate the direction of the data flow.)
g global (together, with '_', you can tell the "scope" of a variable, or function.)
s static (usually indicates singleton pattern.)
_ member variable. (some use m_, but that just clutters up the code with a redundant character. Postfix '_' is also in fashion these days. Use whatever floats your boat.)
Postfix
_ argument is modified. (Traditionaly, "Out" was used, but that clutters up the code. "In" is also redundant, as a) everyone should be using "const" correctness, and b) everything is "In" )
And since I'm dealing with math so much I usually add in:
v vector
m matrix
// along with operator overloading, this shows one doesn't need silly verbose
> I certainly regret making the decision to do 256x2 in my machine... especially since upgrading to gig and maintaining dual channel means I have to buy 2 sticks of 512. Bummer.
I'll second that. Back in 2003, I picked up the Corsair top of the line mem sticks 2x 256 TWIN512-3200LLPT -- wish I had picked a single 512 stick instead.
I'm really amazed how much prices have fallen, since then. i.e. (NewEgg) $70 ASRock 939Dual-SATA2 Socket 939 ULi M1695 ATX $139 AMD Athlon 64 3000+ Venice Socket 939 $153 eVGA GeForce 6600GT 128 Meg $80 Corsair ValueSelect 1 Gig $55 Case w/ 400 W ==== $497 not bad at all.
For gamers, I don't think its unreasonable to assume they have an extra monitor, keyboard, and hard drive laying around. I exclude cases, because why take apart a functioning case+mobo (especially jumpers / LEDs)? I found it to be a far more efficient use of my time, to get a new case with a new mobo.
>...you realise that only a poorly-coded game can be used to cheat in multiplayer, right? Any sane coder will keep the game state entirely in the server, so it's impossible to cheat by "hacking" the source at the client's end.
Slashdot Mirror
> I've noticed a decent improvement in how many monitor related headaches I get from switching to an LCD.
I'll 2nd that, so MOD up.
LCD + DVI = Fewer Headaches!
> Is there a scientific reason why looking at a monitor would decrease eyesight?
The focus & clarity is higher on a LCD, so your eyes don't have to work as hard. The "pixels" on a CRT are not really "one" R,G,B triad, but a "fractional" multiple. Result: Blurry Image, and Eye-Strain.
True, you loose a bit of gamut by tossing the CRT, but a softer image on the eyes is worth it.
I also found:
Contacts = More Headaches.
Glasses = Fewer Headaches.
Your eyes (like the rest of your body) need oxygen.
AMEN! That damn football pre-emptive scheduling pissed me off too.
Way to "promote" a new series, FOX, never let it be on 90% of the time!
The wallpaper radio button is a very good idea. Simple, and effective.
You'd be surprised at wallpapers. The bigger the company, the more likely someone will have an interesting wallpaper, someone sees it, which sparks a discussion, and people get to know one another a little bit better. Who _doesn't like getting complimented?
i.e. "Hey, thats a cool wallpaper you got there! Where did you get that?"
Peace
> C++ is a tremendously type safe language,
// init from vector
// "already defined" // won't compile since arrays == ptr
// func( NULL); // doesn't call the pointer version!
Eh?
* The fact that it is impossile to distinguish between ARRAYS and POINTERS causes all sorts of problems.
* There is no native NULL pointer
* There is no way to DISABLE the automatic up-casting of floats to doubles. (This SUCKS for console programming.)
* No way to prevent integers being promoted to FLOATS. (SUCKS again, for console programming)
Counter-examples:
void func( const int v[3] )
void func( const int v[4] )
void func( const int n )
void func( const int *aArray )
I _really_ wish there was a way to declare/define arrays, and functions via keywords. If C was consistent in reading from Right-To-Left, you wouldn't need the typedef hackery to define function pointers, etc.
i.e.
func( int ) *pFunc;
int* aPointers[3];
int array[3] *pArray;
That said, I love C++, because it is a pragmatic language.
--
Forgotten Christian History
Would help if they didn't use Flash, and actually had links to the Nasa site, so we could save them for our wallpapers, instead of clicking taking you to the next pic.
i.e.
http://antwrp.gsfc.nasa.gov/apod/ap040305.html
> So, which one is the best?
That question is ambiguous. Best at _what_? I would say, none, at the moment. They all have their strengths and weaknesses.
I'm tired of almost every MMORPG having XP = Monsters Killed. Games need to let go of the 1970's (classes, and a warped view of what eXperience Points are), and embrace the new millennium.
i.e.
I can't be a pure crafts{wo}man in WoW without being a fighter. Sadly, you don't get XP for crafting, only killing and exploring. UO had the right idea with trade skills. If you can live with this "restriction", WoW is a good, fun, combat game. It's a much better (streamlined) EverCrack.
The next "big thing" in MMORPGs is Interface. WoW's interface isn't revolutionary, but shows the incremental evolution.
The next "big thing" after that, is Game Scalability. UO had the right idea with people being able to own their own Homes, Towers, Castles, etc.
Think of a MMORPG Sci-Fi game where you can not only own your ship, but with enough credits, star-bases, and planets. A viritual world where not only combat, but trading, crafting, and entertaining are all viable ways of enjoying the game, giving the player interesting choices, not spoon-fed "advancement", is the "best" one.
--
"You can sell your values, but you can't buy morals."
-- _Michaelangelo_
> Make the best game and apparently 5 million people will come.
;-)
WoW is definately _NOT_ the best MMORPG. (I've been playing since the Beta.) What it is though, is better then 80% of the OTHER MMORPGs -- most of the annoying problems are addressed (fortunately mine-jacking isn't too common.)
To give a poor analogy:
If you have a few normal looking woman, and an better-then-average looking womans walks by, she looks beautiful in comparison. Wow is the better-then-average looking woman. It's all relative. While you can certainly have fun with the "normal looking" woman, most people will be naturally attracted to the eye candy.
WoW got the basics right; If you don't make the first 10 levels enjoyable, you're never going to have a customer base at lev 60 to support you. Something that EverCrack doesn't understand is that I _don't_ have 4 hrs to waste raiding.
And yes, WoW is definately inspired by Diablo II -- the Skill Tree, er Talent Tree has 3 tabs. Talents have dependencies, and level requirements.
It's not rocket science that WoW is so popular -- just good gaming sense. WC3 is not revolutionary -- its evolutionary. It's taken the best (better?) ideas, and gone with them. Same with WoW. For every MMORPG failure, there's a lesson to learn on how NOT to do things.
> What would be my ideal MMO? World of Warcraft with Ultima Online's gameplay and EVE's economy and skill gain.
Could you list specific details? (I've been a UO and WoW fan since they came out, and would love to hear your ideas.)
--
"A fine is a tax you pay for doing wrong, and
a tax is a fine you pay for doing all right."
-- Unknown
Anyone know if the compressed images leak has been fixed?
i.e.
Compressed Images Leak
Mod parent up.
Care to provide a (one line) summary of what each extension does?
I tested your code on VC60 (Release)
You can get a ~ 50% speedup if you change one line in the Swap() section, which is kind of ironic, since you said "Bottom line, my register usage makes all the difference."
arg[i] = arg[point];
to
arg[i] = small;
If you got an email addr, I can send you your code cleaned up for readability, and a complete testing program.
Cheers
Loved your insights, and matches my experiences. (I haven't seen any FF code either.)
...
/. sucks @$$ for posting code snippets. At least let the registered users in good standing be allowed to bypass the lame "lameness filter."
i.e.
There have been a few times (3 or so) that FF has completely stalled. Attaching the debugger and seeing that it is "stuck" inside a few functions, and forcing them to return, has brought the browser back to life, complete with my 20+ tabs all intact, and allowed me to save/bookmark them all.
For me the biggest complaints with FF have been
1. Memory usage (bloat with pre 1.0.7 was terrible.)
2. No ability to start a new window in a different process/thread. I believe the WinMain code is using a named atomic, and the new thread terminates (to prevent memory usage duplication) if it sees the semaphore already exists. Wish there was a command line option to force it by-pass this check.
3. Un-responsive UI (also much improved, but still stalls occasionaly)
--
> For the same reason we send robots into hazardous environments - it is too dangerous to justify sending humans.
That's a good point; you only hinted at the other half of the equation, IMHO.
Robots let us get the procedure predictable, so when it actually is time to send people to space objects, that we're not doing it for the first time. The ratio of Reward/Risk is much higher with robots then with humans.
i.e.
What band puts on a concert without practising?
Peace
> How will we justify the use of this when so many people die very young from preventable causes that are beyond their control (as opposed to simply not taking care of oneself)?
/. sucks @$$ for posting code snippets "Your comment has too few characters per line" ??
/. eats extra whitespace in "code" -- any workarounds?
You're under the delusion that their are accidents. There are none. Everything that happens, happens for a reason.
> How will we prevent the extreme accumulation of wealth that this would allow if it is not equally accessible to everyone?
The same way we prevented it today. Oh wait, we didn't.
--
> I get really irritated when developers check in code with sections commented out. If the code is no longer relevant - delete it! If you really need to get it back, get it from a previous version in the source control system.
... #endif sections.
;-) of the new function, with commented parts of the original that are in the state of being converted.
/. sucks @$$ for posting code snippets "Your comment has too few characters per line" ??
/. eats extra whitespace in "code" -- any workarounds?
I assume you're not talking about #if
How many times has code been cleaned up, and ended up broken? It's a PITA to:
(1) rename / resave the current version as "new"
(2) fetch the old one
(3) diff
When it is easier to read the "old" and "new" code, intermingled.
Sometimes the old code is left in, as a "fallback" -- since the new code is still in a state of flux, and hasn't been thorougly tested. A good example is optimizing a function, piece-wise. I found it best to:
(1) have complete copy of the original function commented out
(2) have a "working" copy (or is that "not-yet-working" ?
But yea, once the stability has settled down and the new code is working, I agree. Nuke the old sucker. Unfortunately, once code is checked in, people tend to forget about it. "It's working, right? Why clean up cruft?" lol.
--
==========
If you use proper descriptive variable names, you don't need comments 50% of the time! Why? Because if you can't name something properly, chances are, you probably don't understand the problem fully.
Another good idea, is that many people use "notes" in the code that you can search for, such as:
Here are few examples with commentary: // Look ma, no comment is even needed! // Bits Per Pixel -- plus now I need to remember what the hell "bpp" stands for. Sure this one isn't TOO bad, but this is NOT FUN when people invent their own non-standard acronyms.
// Gee, I wonder, is this a Callback?? // who calls me? and what does it do? Set or Get mouse paramaters??
/user /temp, instead of the inconsistent: /usr /var ??)
int nBitsPerPixel;
int bpp;
Callback_Mouse();
mouse();
Any half-decent editor will have auto-completion, so typing in long names are a thing of the past. Plus you don't have to remember _cryptic_ abbreviations. Dealing with Unix's naming convention is bad enough, don't make me deal with it in code too. (What, someone couldn't think to use the names:
Coding Standards
================
Whatever standard you pick, PICK ONE. A bad standard is better then no standard at all. and The nice thing about standards, is that there are so many to choose from.
When dealing with Physics or Math calculations, it always very handy to keep track of the "units". On the HP calculators, they use "_{units}". i.e. 299792458_m/s. In code, this isn't practical -- as the code tends to become too verbose. It would also be great if we could see the "scope" -- local, global without extra clutter. What kind of system could we design/use that is practical?
Now note that when you take a standard to the logical extreme, you end up with something impractical. Applied to this domain of programming we find that absolute abonimation called Hungarian notation, micro-managing every little detail. (Do you really need to clutter up the variable naming detailing the internals if a string is zero-terminationed?? NO!!)
However, if you use something more PRACTICAL, you end up with a very nice, compact, naming schema.
Prefix
a arrays / tables
b bool / flag
i iterators, along with the object type. i.e. iWord
n number of, or size of. i.e. nWords, nBytes, nBits,
r reference (not really needed, but some people like them.)
p pointer (handy with pSrc, pDst, pTmp to indicate the direction of the data flow.)
g global (together, with '_', you can tell the "scope" of a variable, or function.)
s static (usually indicates singleton pattern.)
_ member variable. (some use m_, but that just clutters up the code with a redundant character. Postfix '_' is also in fashion these days. Use whatever floats your boat.)
Postfix
_ argument is modified. (Traditionaly, "Out" was used, but that clutters up the code. "In" is also redundant, as a) everyone should be using "const" correctness, and b) everything is "In" )
And since I'm dealing with math so much I usually add in:
v vector
m matrix
> I certainly regret making the decision to do 256x2 in my machine... especially since upgrading to gig and maintaining dual channel means I have to buy 2 sticks of 512. Bummer.
I'll second that. Back in 2003, I picked up the Corsair top of the line mem sticks 2x 256 TWIN512-3200LLPT -- wish I had picked a single 512 stick instead.
I'm really amazed how much prices have fallen, since then.
i.e. (NewEgg)
$70 ASRock 939Dual-SATA2 Socket 939 ULi M1695 ATX
$139 AMD Athlon 64 3000+ Venice Socket 939
$153 eVGA GeForce 6600GT 128 Meg
$80 Corsair ValueSelect 1 Gig
$55 Case w/ 400 W
====
$497 not bad at all.
For gamers, I don't think its unreasonable to assume they have an extra monitor, keyboard, and hard drive laying around. I exclude cases, because why take apart a functioning case+mobo (especially jumpers / LEDs)? I found it to be a far more efficient use of my time, to get a new case with a new mobo.
Cheers
Interestingly that Rhino 3D is picking up users.
Does anyone know how the big 3 modellers compare? (I know a lot of game dev studios use Max and Maya.) Also, what about Blender?
TIA.
Depends on your defintion of "good"
Tomb Raider was half decent.
Forgot this one ...
.NET 2003 in Batch Mode
3. Upgrading Visual C++ 6.0 Projects to Visual Studio
] if 6.0 to 2003 is any indication Get ready for a lot of manual header adding.
.NET 7 <--> .NET 7.1
Work smarter, not harder.
i.e. There are utils/macros to do that, specfically, you'll want to check out...
1. Project Converter
2.
--
If the Bible is infallible, which version would that be?
Sweet links!
Mod Parent UP!
> ...you realise that only a poorly-coded game can be used to cheat in multiplayer, right? Any sane coder will keep the game state entirely in the server, so it's impossible to cheat by "hacking" the source at the client's end.
How many games have you actually shipped?
> My friends first opengl programming gaem was a pushing crates game with an ascii file map format, good stuff!
You thinking of Sokoban ?