Madelaine McMasters

Done. Enjoy spring break, Heather!

Qwalyphi Korpov wrote: I mention this because today I look at my keyboard. It has a home key. That don't take me to any of those. This reminds me of the reason my Father tossed his first PC and never went back. It had a word processor that, upon exiting, displayed this confirmation request... "Press Enter to exit, Escape to return".

Jaser Newell wrote: I understand what you're saying, but I still consider it a bug. I admit that I didn't understand the nature of how floats worked in lsl in the beginning, and in fact did make the mistake of thinking the ASCII representation would be an accurate description, and I thank those of you that posted for pointing out my error. Yet as you pointed out, base 10 can be done by emulation. SL chose not to do that, which even if this choice was necessary can still problems. As you also pointed out, lsl uses the same representation for assignment and comparison. That's why converting it to a string and back to a float after the arithmetic operations allows it to work without any of the difficulties referenced thus far. This seems to be a perfectly suitable solution that gets around that pesky binary problem. Well, even emulating decimal doesn't rid you of problems, and it would add considerable burden to the SL servers. You can't represent 1/3 in a finite number of decimal digits, so the errors are still with you. Math systems like Maple get around such problems by representing 1/3 as a ratio, not as a single number. The need for such things in SL is virtually non-existant, so tried and true floating point, which is built into the Intel hardware, is the standard. Pesky problems abound, Jaser!

Jaser Newell wrote: As far as the issue not being a bug, I would disagree. If something as simple as 3.1 + 0.03 creates any other value than 3.13, then I consider that a bug no matter what the reasoning is behind it. The "bug" that's bugging you is an artifact of the binary nature of computers. To accurately perform the decimal calculation you give as an example, you'd really need a computer doing math in base 10 (either in hardware, or via emulation as in the case of Mathematica, Maple or other math environments). As Dora explained, a number like 0.1 cannot be expressed in a finite number of binary weighted fractions. In decimal it's easy, it's just zero units and one "tenth". In binary it would be... 1/16 + 1/32 + 1/256 + an endless series of smaller binary fractions Note the "endless". There is NO finite series of binary fractions which totals 0.1. It requires an infinite number of bits. That's not a bug, that's the way math is. Now, if you only assign constants to variables and then test those variables against constants, doing no math in between, there's a good chance it will work as the compiler uses the same floating point representation both for assignment and comparison (provided you typecast as necessary, as I explained earlier). But the moment you perform an arithmetic operation on the variable, all bets are off. The result of the computation will contain the unavoidable discrepances that arise from trying to represent decimal fractions on a binary computer.

Jaser Newell wrote: After further investigation I think I figured out what the issue was. My original script made use of "+=", creating the number as the script ran, as follows: Task = 3; Task += 0.1; Task += 0.03; This seems to sometimes create a slightly different value for "Task" than would be expected that is undetectable when it is converted into a string. While "(string)Task" will produce "3.130000", "Task" does not equal exactly 3.13. Since I have a user defined function that is always ran right after "Task" is changed, I just threw "Task = (float)((string)Task);" into it. Since the slight difference is undetectable when converted to a string, when the resulting string is converted back into a float it produces the value originally expected and allows the script to work. Right, you can not always count on the result of a floating point math operation to produce an exact result, as the decimal fractions you express as constants may have no exact equivalent in floating point. It's also common for math operations to result in rounding/truncation errors. For this reason, scientific calculations are often performed at floating point sizes well beyond the required precision for the result so that the accumulation of errors over the course of the algorithm does not corrupt the result. The FPU in your computer does internal calculations at 80 bits, even though a double float is stored in only 64 bits. I was sure I'd seen the error you describe in my own test with your code, but as I saw it only once, it 's just as likely I was in error. A simple assignment of a float constant value to a variable, followed by a test for equivalency, should work so long as the precision of the constant is declared to be the same as that of the variable. As I said earlier, LSL has only one size of float, so your original example should work. ETA: If you were attempting to code multiple sets of states a single variable, say assigning one set of states to the units digit, another to the tenths, another to the hundredths, etc, you could achieve the same result using integers and chunking your sets into little spans of bits. You could encode eight states in three bits, 32 in five, etc. But I think breaking the groups out into separate variables would be both easier to understand and easier to debug.

Hi Jaser, and welcome to the forums: I've seen this behavior before in other compilers, but after experimenting with your code in-world, I'm inclined to think there is some kind of bug in the system. I hand typed your code into a test object and it worked for me. I then pasted your code into the same script and it did NOT work. I thought there might be a hidden unicode character messing things up, so I reformatted your code and it began working. I was hopeful I'd figured out the bug, but wanted to confirm I could break it again by repasting from your post. That subsequent paste worked! I've been unable to reproduce the error since, even in new objects and even under my alt account. For future reference, you must be aware of how the compiler treats constants which do not have a declared type. In a system which has both floats (32 bit) and doubles (64 bit) a constant containing a decimal point would usually be treated as a double. Assigning that constant to a float variable will, depending on the compiler, either truncate or round the constant to the lower precision. Subsequent comparisons between the constant and variables may fail as a result. In such a system you'd avoid this problem by typecasting the constant: if (Task == (float)3.13){ llOwnerSay("Works"); } LSL has only one kind of float, and that's "float", so this should not be an issue here. As I am no longer able to make your code fail, I can't investigate further. Good luck!

LadySue's dress looks short enough to be safe, but if the bride's dress has a train you better be careful, Hippie. I've seen you walk and I fear the worst! ;-)

PudgyPaddy wrote: Candor Philipp wrote: the best option is to be between the extremes of what your target audience consider acceptable, and to try to preserve what is more important to them. What? This is not communist Russia! The "best" opinions are founded in reasoning and logic. I don't write opinion(s) to placate readers with flowery words so as to not offend misplaced sensibilities or to not cause emotional instability. I don't write opinions to win friends or win a popularity contest either. Going forward, whenever I read *your* opinion I will know that you didn't mean what you wrote--you just want to look good to the rest of us or the one. This is perhaps the worst statement that could have been written on an open and public forum. PS The OP didn't offer an opinion. The opening post asked a direct and specific question. As Canoro was discussing a "target audience" it's pretty clear to me he was not talking about opinions. He was talking about sales. Sellers do indeed placate buyers with stories that are likely to induce them to purchase their goods and/or services. There is a term for this sort of placation, it's called "marketing". Going forward, whenever I read your opinion, shall I make greater allowance for the possibility that you didn't understand the subject and that you just want to look good to "the one"?

I've never seen anything like that before, Nwyn. It looks like a nutty DOF setting. Do you have Shadows/DOF enabled? If so, try turning it off.

Monti Messmer wrote: At work i have an 18" CRT. Boss "Bigger screens dont speed up work u just want to watch ......" lol Sorry for that excurse back to topic. Monti Monti Messmer wrote: At work i have an 18" CRT. Boss "Bigger screens dont speed up work u just want to watch ......" lol Sorry for that excurse back to topic. Monti I had this discussion with my colleagues. Some of them thought that getting the fastest CPU was important. I argued that having a bigger screen was more important, except under certain circumstances. I continue to believe that. The time I waste moving stuff around on a cramped desktop can't be made up by a fast CPU. Graphics or compute intensive things like SL or gene sequencing argue for speedy chips, but for everything else, give me an expansive display.

Dillon Levenque wrote: I was going to ask, "But what if one is only a paramecium in a puddle even in SL?", but I probably don't wanna know.:smileywink: As one who is only an amoeba in SL, I can comfort myself with this...

Monti Messmer wrote: 2 or 3 monitors ? Good lord posh folks One 24" monitor - widescreen what makes a difference to old 16:9 Monti It helps to use those two or three monitors for work, then you can rationalize almost anything. ;-)

I checked 27" but that only tells part of the story. My main 27" iMac (2560x1440) uses my old 27" iMac (2560x 1440) as a monitor on the left unless I'm also using that iMac for SL. I also have a 28" (1920x1200) display to the right. I'll also agree that pixels are more important than inches. I'd love to have a single 8Kx4K 32" monitor. But even with all my screen real estate, I truly dislike both the large size of the SL UI elements and the inability to drag them out of the app window. When running other apps, I often drag lesser used control windows out to the side monitors. SL is one of the few apps I use which is confined to a single window. ETA: I just did a quick check of applications I use often. SL is the ONLY one that is confined to a single window.

Folds up her easel and takes a seat in the audience to watch the beginning of SL as we'll know it. ;-)

Janelle Darkstone wrote: ( Okay, I'll stop. Just extremely bored and with way too much time on my hands today... :smileywink: )

Sephina Frostbite wrote: NOOB SOUNDING ALERT!! Whats linking? Most of the things you see in SL are made of more than one prim. To keep those prims together in one single object, they must be linked. If you've rezzed multiple prims, you shift-click on them while in edit mode. Each prim you click on while the shift key is down gets added to the growing collection. The last prim you click on this way will be the root-prim when you finally select "Link" from the edit window. The root prim is like the foundation of the linkset. When you later wish to move the linkset, the root-prim's center is where you'll find the editing axes. Building is great fun, Sephina. If you'd like help learning, Hippie and his friends offer classes.

FIFY = Fuzzy Innie? Freaking Yucky! Who are you going to believe?

16 wrote: Madelaine McMasters wrote: ...sits on your front porch and looks all humble. you can come inside and stay bc you must be really poor but .... !! you have to learn to get up really early ok. bc we even more poor. so the usual rules ok. like first up. best dressed but .... more !! Janelle moved in already ok. so have to get even earlier now than before. bc since she move in then is now: first up. best dressed. and fresher breakfast jejejjjejejejje (: You go ahead and believe that the first to rise does best while I make sure I'm the last to fall asleep. ;-)

Dillon Levenque wrote: Madelaine McMasters wrote: Dillon Levenque wrote: 16 wrote: + I think I better invade your lands before you serf start breeding. bc my massive army of one is not much outnumbered at this time Egad. I shall alert the Border guards. Just as soon as I have some. And you have more than one border. It's that damn road: they can come at me from either direction! I tried putting up barricades but the Lab keeps sending them back to me. I don't need a road.

Dillon Levenque wrote: 16 wrote: + I think I better invade your lands before you serf start breeding. bc my massive army of one is not much outnumbered at this time Egad. I shall alert the Border guards. Just as soon as I have some. And you have more than one border.

Strawberry Singh wrote: Since you guys were wondering, I thought I'd share a raw shot of the picture, with no photoshop: http://grabilla.com/03314-8eeb085f-a48e-4ba3-aa94-ce9562793a12.png Also yes, those are the SLink mesh feet, all of my style credits to that image are available on my blog here: http://strawberrysingh.com/2013/02/20/secondlife-com-spring And I'm not always half naked, sometimes I wear more clothes. Thanks to everyone for appreciating the pic! <3 Sometimes I wear less, Strawberry. Like two bird's nests and a pocketwatch. And I am forever thankful to you for finding the hair I've been wearing for nearly two years. I don't look as quite as good as you did modeling it, but that was never my intention. It's as me as I've ever seen. ;-)

I will not swoon. I will not swoon. I will not...

Perrie Juran wrote: Phil Deakins wrote: Just like I said - I do look good in pjnk Do you do fashion consultations? Sometimes I think I could really use some help. Stop it you two, I'm going to get the vapors!

Treasure Ballinger wrote: That was surely taken at the Cartel hangout, see the sofas and table with the kitty, in the background. You were lovely, Phil. :matte-motes-sunglasses-3: ... swoons. Yep, that's the Cartel Hangout. I recognize the railing, which left an indelible impression in my derriere (from sitting on it!).

Orca Flotta wrote: Impossible. That's not even a SL avie. The depicted girl in the graphic has no skin and painted hair, so what do you expect to achieve? A badly silk-screen printed avatar? My RL self is not an SL avie, nor is the Peanuts' character Lucy Van Pelt. I've mimicked both in SL as well as a cloud (not intentionally), a mouse, a ghost, a transparent hippie and Ms. Pac Man. I'm beaming over having done the impossible not only once, but many times!