Catch up on stories from the past week (and beyond) at the Slashdot story archive


Forgot your password?

Slashdot videos: Now with more Slashdot!

  • View

  • Discuss

  • Share

We've improved Slashdot's video section; now you can view our video interviews, product close-ups and site visits with all the usual Slashdot options to comment, share, etc. No more walled garden! It's a work in progress -- we hope you'll check it out (Learn more about the recent updates).


Comment: Frequency of Mathematica bug-fix updates (Score 2) 210

by iliketrash (#49076033) Attached to: Interviews: Ask Stephen Wolfram a Question

About 15 years ago I found a bug that affected all Fourier-like transforms in Mathematica. (It was related to how the constants can be “allocated” between the exponent and an overall scale factor—someone had tried to generalize this concept by being too clever by half, and made a mistake.) I did a sanity check with comp.lang.mathematica or whatever the group is called and then filed a bug report. I understand that the error was not corrected until a later major release of Mathematica.

A few months ago I returned to Mathematica with a medium-sized project which involves some probabability calculations (PDFs, characteristic functions, etc.) I quickly found that Mathematica failed to crack an integral because it did not do a simple, trivial, second-semester substitution. I also found an error in the way a special function (MeijerG) is calculated numerically. In all, after only about three weeks of returning to using Mathematica, I filed five bug reports (one of which was UI-related) and have two or three saved up for when I get more time. I have watched the Mathematica release cycle for some years including the “dot” releases, and I am not encouraged that any of my reported bugs will be addressed before the next major release. (I believe that would be version 11.)

I have finally drank enough Kook-Aide to appreciate Mathematica and indeed have rather quickly (after my recent return) found it indespensible in my work; I am no longer even tussling with whether to use Octave/Matlab or Python/NumPy/SciPy for numerical work.

So: Why does Wolfram respond so slowly to bug reports? There seem to be only one x.1 or x.0.1 release after each major release, if that. Why not release more-frequent bug fixes like most other software houses, rather than let bugs exist for years in some cases?

Erim Radcliff

Comment: Notes from an old hand (Score 1) 464

by iliketrash (#48722647) Attached to: Ask Slashdot: Are Progressive Glasses a Mistake For Computer Users?

I’m going to be as brief as I can. I could write pages on this topic. Buying progressive lenses twice over a few years has been the worst retail experience of my life. The dispensers are like clothes salespeople and other retail stores: get the customer out the door as fast as possible. Unfortunately that didn’t suit my style and so I had to become my own expert. I won’t write generally but will try to address your specific problem.

I have scanned the earlier comments and didn’t see all that much that is useful to you. (Sorry, other posters, I said _scanned_.)

First: There are _hundreds_ of progressive lens designs, and probably thousands of patents. Most dispensers have found that they have a decent success rate with a particular design; for all I know, they might have a bulk rate from a manufacturer. If the lens does not suit you, take them back and ask for another design. Any respectable dispenser will tolerate this—they won’t love you, but they will tolerate it. On my first go-round, I had four different designs, but then the lab screwed up at least one or two.

Some designs are quite crappy. There has been a lot of progress in lens design in the last 10 years. Some lenses are still designed by an old buy who remembers Zernike polynomials and does his work with a pencil and paper. You don’t want these. Get a modern lens design known as digital progressive lenses, high definition, HD, or more commonly, free-form. These designs are made using a computer and divide the lens up into “pixels” where a performance goal is met subject to an overall cost function. The in-between-points are probably interpolated with a cubic spline. You see, the problem with lens design is, if you try to make it better in one place, it gets worse in another place. This is why it is an interesting design problem. The modern, free-form designs are _far_ better than the old designs. Some designs actually allow precise measurments to be made on your frame _before_ the lens is made; this information is sent to the lab and incorporated into the lens design, essentially a custom-made design. Shamir comes to mind, but there are others. At least one company has a slick jig for an iPad and an optical splitter that fits over the iPad’s camera so that even a monkey could set this up without screwing things up. (FYI, be prepared to deal with monkeys.) The free-form designs can trade off near, intermediate, and distance regions that you can select depending on your lifestyle. Many new free-form designs excel at getting a wide channel, which is to say, the intermediate, which is great for most computer work. Free-form designs are much better at getting the near vision wider. My current lens is a free-form design and it is _far_ better in both intermediate width and near width than my old conventionally designed lenses. (FWIW, my current lens is the best Seiko lens, and I recommend them.) And you will also find that there are “computer glasses” which you should consider; these have giant intermediate and near regions and pretty much punt on distance.

Many dispensers will bitch and moan if you get a short lens (frame); these became fashionable 10+ years ago and still are. That’s bullshit. They will try to put you in an “old fart” frame because they claim that it forces fewer compromises on the design. Then they will take your prescription and increase it so that it is more powerful at the bottom of the giant lens so that it is correct at the near vision region that you actually use. You simply can’t rotate your eyeballs far enough to see to the bottom edge of a giant lens.

Getting the frame adjusted (bent) is critical for progressives. The optical axis has to be aligned with the optical axis of your eyeballs. The lenses come from the lab with a fitting cross or dot on the lens. If you get your new lenses and do not see the fitting cross, your dispenser has removed it. Ask them to put it back on. They have a simple tool to do this. The fitting cross should be directly in front of your pupil when holding your head normally and looking at a distance. You won’t be able to see the cross yourself but they can; or you can look in a mirror to satisfy yourself that it is correct. THIS IS IMPORTANT! The nose pads can be adjusted to raise or lower the fitting cross. Two other parameters are important: the pantoscopic angle and the wrap angle. The pantoscopic angle is the angle that the mostly vertical lens makes with the front of your mostly vertical face—look at your profile using a double mirror. Most designs are intended for about 10 degrees. This is a little hard to measure yourself but you can eyeball 10 degrees. Your dispenser should bend the frame for you; in my case, I had to learn how to do this myself. Obviously if you have plastic frames this isn’t always an option, so word of advice: get metal frames. Best, get three-piece frames. The wrap angle is how much the frames wrap around your head. Imagine looking at your head from above and observing how they seem to wrap around, or match your head shape. Getting this set as much as you or your frames can tolerate will enormously help the global distortion (not focus) that you see. Demand that this be be adjusted or do it yourself. (Bending metal frames is not rocket engineering. Get some needle nose pliers and pad it with a thick wash cloth to avoid marring the frames. I have titanium frames—Marchon 720—and they are great to work with and indestructible.) Overall, get the frames as close to your eye as possible, as this also reduces global distortion. (Contacts are zero distance and have zero global distortion.) I set mine so that my eyelashes gently brush the lens when I’m laying on my back, face up.

Pay for optical obsorbing coating. This is a worthwhile add-on, as it greatly reduces all sorts of anamolous reflections. This is kind of cool—it is seven or nine layers of optical material of slightly different index of refraction, sputtered on the lens in a series of quarter-wavelength absorbers. Pretty neat, and really effective. And if you get rimless or monofilament-attached lenses, don’t get the edges of the lenses polished—polished edges reflect overhead lights and sunlight like crazy and you will see these annoying reflections all the time in such lighting. Get the dull finish. The dispenser will try to tell you that the polished edges is cosmetically better but screw that. I have pretty thick lenses and nobody has ever said my unpolished edges look uncool.

The vast majority of material used for lenses is polycarbonate. This is the hard, tough plastic that Compact Disks are made of. That’s fine if the only thing you’re going to shine through them is a laser (single wavelength) but for the octave or so of visible light, polycarbonate is the _worst_ material there is. Why? Because is it seriously dispersive. That means that the speed of light (and the index of refraction) is dependent on the wavelength of the light. And polycarbonate is the worst. Really. There is a crude but common and effective measure for this phenomonem called the Abbe number. Polycarb’s Abbe number is 30. A resin known as CR39 is outstanding with an Abbe number of 60, which I think is also the Abbe number of crown glass. The so-called high-index materials are pushed by many dispensers. They tend to have also crappy Abbe numbers and are not nearly as strong as polycarbonate but they are popular because the lens can be made very slightly thinner. So if you don’t care if you go blind when your car’s airbag goes off in your face, get a high-index lens. Otherwise, there is a good compromise: Trivex. Some manufacturers re-name it, but it’s still Trivex. It is as tough as polycarbonate but has has intermediate Abbe number of 45. I forgot to mention why dispersive materials are bad—they act like prisms for anything off the optical axis, and the further off, the worse. I mean, they separate colors into their spectral components. Not cool. This will bother you especially in high contrast scenes such as nighttime signs or reading in bright light. You will actually see the color fringing. Just get Trivex and be done with it.

You declared your use very clearly. Your choices are progressive including “computer glasses", lined bifocal including so-called executive or Franklin bifocal which extends the “line” across the entire width of the lens which I think you should seriously consider, or a single-vision lens that you wear only for computer work, but if you need correction for computer work then you also need it for reading and other close stuff. Your choice depends on if you are OK with removing and replacing your glasses during the day and carrying them to restaurants etc. versus how you tolerate multifocal, progressive or conventional. But—a single-vision lens will work just fine for your use (but your dispenser won’t make as much money off of you). And you should know that adapting to progressives can take a while. Give it at least three weeks. You will be surprised what your brain can adapt to especially if you have never worn glasses before. But seriously, a free-form design will be easier to adapt to compared to an old-school progressive design.

Oh yes—you will not find _anyone_ who is an expert on comparing all of the available designs, but some dispensers are more concerned about learning this information than others. I don’t know how you find them. I found tons of useful information on the discussion lists at

So there. I believe I started out saying that I would be brief. :-/

+ - "Uncalibrated touch-screen" votes Chicago-style->

Submitted by iliketrash
iliketrash (624051) writes "A touch-screen voting machine in Chicago suburb Schaumburg has been found to repeatedly convert votes for several Republican candidates into votes for their Democrat counterparts. The problem has been blamed on a "calibration error of the touch-screen on the machine.” Of course, Chicago has long been notorious for voting hijinks, mostly of the Democrat flavor.

Do touch-screens need such a gross calibration, or for that matter, any calibration at all? If so, what are the chances that an uncalibrated screen would reliably convert all or several votes for one party into votes for the other party?"

Link to Original Source

Comment: Flamebait (Score -1, Troll) 223

by iliketrash (#45250165) Attached to: Ask Slashdot: Best Cross-Platform (Linux-Only) Audio Software?

Why are you such a Linux retard? You're killing yourself creatively by asking for this kind of of software for Linux. If you really want to explore your musical creativity you need to get the hell off such a limited platform. OS X and probably Windows platforms are far, far better as far as choices in this field. Yea, there is Adaucity which you're already found is a piece of crap, or would have discovered had you not limited your world to Linux. And Ardour—nice work, but that's about it. Do yourself a favor and let your creativity flow by looking outside the extremely limited options offered by Linux.

Comment: Calligra web site is uninformative (Score 1) 30

by iliketrash (#44462751) Attached to: KDE Releases Calligra 2.7

The Calligra web site is incredibly bad—there is almost no information about any of the programs, just a few superficial paragraphs. Just awful, an embarrassment to the development team, I'm sure.

This is an example of the "airport signs" problem: The people who build the airport already know how to get around the airport and so the signs that they put up are not helpful to those who rarely visit the airport.

On another crappy note, the link to the OS X installer just goes 404.

Oh well.

Comment: Then three days later, it's back to normal (Score 1) 173

by iliketrash (#44454523) Attached to: Camping Helps Set Circadian Clocks Straight

Right. Then the person gets back from his/her happy little camping trip and back into his/her normal life and three days later he/she is back on the same stupid schedule. I've done this many times and camping is not necessary—any outside influence that causes one to rise earlier will do. I didn't read the article but it sounds really stupid to me, so I won't.

Comment: Re:Google Docs CAN DO LaTEX (Score 2) 160

Well, this does look interesting. I played with it for 15 minutes and: I can't think of anything worse than writing anything of any significant length inside a browser. This bypasses all of the hard work that my OS provider (Apple) has spent for decades polishing a decent user interface. As far as I can tell, everything has to be done using the mouse/trackpad—no keystroke shortcuts.

Also, compiling even the short sample document is excruciatingly slow. There is an option to use my local TeXLive installation but the radio button to select it was disabled. If one really wants an easier-to-use LaTeX editor, there are free ones that also provide one with menu-selectable math items.

Comment: Why doesn't Linux use sub-pixel rendering on text? (Score 1) 169

by iliketrash (#43930649) Attached to: One Week With GNOME 3 Classic

The screen shots from show that this interface is not using sub-pixel font rendering. I have noticed this on most if not all other Linux-type screen shots. Apparently the favored font rendering method on Linux is the old-fashioned "treat every pixel as some shade between the font color and the background color". The characters so rendered are substantially less well-formed and harder to read. And this surely isn't a matter of intellectual property:

Comment: Woohoo! Multiple windows side by side! (Score 1) 800

by iliketrash (#43867109) Attached to: First Looks At Windows 8.1, Complete With 'Start' Button

"It'll also be possible to have multiple windows from a single app so that, for example, two browser windows can be opened side-by-side."

WFT? Is this for real? Was this not possible before? Surely the implication of this statement that this was not possible before is wrong.

I do not fear computers. I fear the lack of them. -- Isaac Asimov