Posts Tagged ‘linkedin’

Filename Conflicts On Download

Thursday, March 4th, 2010

I was very satisfied when Apple revisited how the Finder handled some of its file-naming architecture. Some behaviors were bugs and absolutely had to be fixed. For example, prior to Leopard there existed a mixed behavior when renaming files under different view modes (Column View, Icon View, Detail View, etc.).

Finder's toolbar buttons for the different View Modes.

Normally if you start changing the name of a file and change your mind you could hit the Escape key, and the name would be restored to what you started with. However, under the Column View, hitting Escape was equivalent to hitting Return, and whatever text you currently had would be accepted. Very annoying! Obviously the first behavior is more intuitive, and it is inarguable that the renaming file behavior should be consistent throughout all View Modes.

The Good

Finder's intelligent filename selection which excludes the extension when renaming.The feature I most liked, however, was that if the extension of the file is visible and you started renaming it, the entire filename except for the extension was selected. This change made renaming much quicker and easier because not only did you no longer have to remember the extension, but less typing was required (on average four characters: period + three letter extension). At first these four extra characters might not seem like too much work, but if you rename a list of files, you would drastically notice the difference.

The Bad

There is at least one more thing that I see that needs to be changed. When Safari downloads a file and another file already exists with the same name, it begins appending sequential numbers (preceded by a dash) to the end of the filename in order to distinguish them (“-1”, “-2”, “-3”, etc.).

Two icons showing the intended behavior of Safari's renaming algorithm when downloading files with the same name.

But is the added suffix really appended to the end of the filename? Turns out it isn’t and I consider it a major bug. Say what you will, but I don’t see it as simply a behavior that isn’t necessarily right or wrong. I think it’s obvious what the intention is (refer to the screenshot above) and the screenshot below shows an example when the intention doesn’t carry through. From my observations, the Finder inserts the “suffix” immediately before the first period it encounters while searching from beginning to end of the filename. It should be inserted before the last period.

Two icons showing the malfunction in Safari's renaming algorithm.

The Solution

It would be very easy to iterate from back to front of the filename (my Page Capture widget does). And it happens to be more efficient since extensions are typically shorter than the actual filename. Hopefully someone over at Apple reads this post and fixes the problem. While they’re on it, they might as well fix the other problem I’ve mentioned with the TextEdit Autosave algorithm.

Uses of Polarizing Filters

Tuesday, January 19th, 2010

Just some examples of how having a polarization filter handy can improve your photos drastically! All of the following photos are straight out of the camera (no post processing except perhaps cropping).


Color Saturation & Definition

Drastic changes to the colors of the sky and trees with a polarizing filter.

“See Through” Water

A polarizing filter can help see 'through' water.

Saturate Hair Color

A polarizing filter can saturate hair color.

Sky Definition & Water Tones

Drastic changes to the colors of the ocean and sky definition with a polarizing filter.

Gross iPhone OS Update Typography

Monday, November 9th, 2009

Check out the first line of the update message in the screen shot that I took on my iPod Touch below. The mistake between ‘O’ and ‘0’ is too apparent to not write about… not to mention typographically gross!

Screenshot Someone’s finger must have slipped from the ‘O’ to the ‘0’ — bad form!

iTunes Sucks

Tuesday, October 20th, 2009

You know, it’s a shame that iTunes now sucks.

A few years ago in the midst of the chaotic world of digital jukeboxes, iTunes emerged as a simple application to manage and play music — offering the basic features that one would need. Although it wasn’t groundbreaking in most regards, it did manage to stand out among the rest due to its extremely simple interface and integration with Mac OS X*.

These days, iTunes has become the fat and greedy roommate of the OS — every time it is called upon, I can hear the squeak of the plastic sticking to it’s back as it readjusts it’s fat ass on the vinyl couch. Its shear size and its toll on the processor has grown immensely and we can thank the superfluous features that are bundled with it: an integrated music store, Quicktime (oh my God… Quicktime for heaven’s sake!), a synchronization manager for your events/contacts/ringtones, etc.

I’m a programmer so I know that little glitches occur here and there. But a stable release of software, especially when developed by a professional corporation such as Apple shouldn’t have the problems I’ve noticed with iTunes.

Erroneous progression bar in iTunes.20 seconds have elapsed and there is more than three minutes remaining. Why is the progress indicator at half?

Moving on… the application allows for adding media into its library by simply dragging and dropping files into the iTunes window. This feature is nice but unfortunately also very unreliable with large quantities of files. When I attempted to add many files by this method, I noticed that not all the songs I imported had made it. In fact, only about 80% of them were successfully copied. Yes, I can prove it:

Screenshot showing the differences between two iTunes Libraries.A diff of the original and imported librarys’ contents. The gaps on the right show files that never made it into the new library.

Why does fundamental functionality not work? Could it be because the application is so big now that there are more avenues for errors to be produced? Maybe errors like these emerged because Apple decided to add features like mad men to an application that wasn’t theirs to begin with.

What?

I know, but Iit’s true. iTunes evolved from an application called SoundJam which was originally developed by 3 people at Casady & Greene. Apple purchased from the rights to this bit of software and after a quick face lift, released it as their own “iTunes.” John Gruber talks a bit about related issues on Daring Fireball and John Knack responds with some insight. And if you’d like to know more about the early days of music players on the Mac platform, there is a very informative story written by the guys at Panic.

Anyway, company acquisitions and inheriting code is normal these days, but I fear that iTunes never had the proper foundation for the features that it now supports. The amount of functionalities in iTunes can be deceiving since it has managed to maintain the same basic interface structure over the years — a commendable design achievement considering the growth of iTunes in the last few years. But integrating new functionality and features under the same simplistic interface more often than not calls for shortcuts that can become problematic.

The LCD-like interface of iTunes which functionally is expected to show current track information is used for displaying the progress of a download.

For example, whenever the LCD-like display is used for anything but current track information, comprehension for the user becomes a problem. Applications mimic real-life objects in order achieve intuitive functionality, but when the expected functional behavior of an interface element is modified, confusion arises.

With the release of iTunes 9, other inconsistencies have shown up. Take the volume slider for example. The part to the left of the dial is now “filled in” like a progress bar. But the slider on the heads up display when watching videos remains like the previous version (which I personally prefer) which is just a simple dial on top of a slider bar.

Volume slider and progress bar inconsistencies in iTunes 9.

And wait a minute… isn’t watching videos in a music player an oxymoron?

I think the only reason I’m still able to use iTunes is because I disable everything that I can in the preferences. This way, I’m able to make iTunes appear like the simple music player it used to be; like the simple music player it should be. Even so, it’s still disturbing to know that all those frivolous features that I hate are still tucked in underneath and weighing my application down.

Trying to keep the interface of iTunes simple by disabling all the frivolous features.

I’ve disabled as many of the features I can in order to keep iTunes looking like a simple music player. But even with the features hidden, they are just itching to jump out at me because despite having disabled the iTunes Store, I see a “Contacting iTunes Store” message when I launch the application. How frustrating.

*Yes, soon after its initial release, iTunes did of course have the ability to synchronize music to a portable player, but this feature wasn’t new for jukeboxes. The problem (that iTunes “solved”) was that you couldn’t buy a portable music player from Company A and use a digital jukebox from Company B to efficiently synchronize your music. The best way to manage music on a device from Company A was to use Company A’s own proprietary jukebox software. It’s no different with Apple. iTunes synchronization only works with Apple’s line of portable MP3 players — uh, I think they’re called iPod — so the way that iTunes “solved” this problem was to gain dominance in the market so that everyone uses an iPod and iTunes anway.

If only there had been a way to quickly synchronize the contents of a music library from Company X with a portable music player from any other company. Hmmmmmm.

Apple’s TextEdit App. Can Erase Your Files

Tuesday, May 12th, 2009

Icon of the TextEdit Application.Apple’s TextEdit application has a massive design flaw that could potentially erase other files on your computer. Weird, right? Ironically, it’s TextEdit’s safeguard against loss of data that is the culprit of the defect. And the corruption of files isn’t a randomly occurring glitch either — it is caused by a shortcoming of the algorithm used in the Autosave feature.

The Autosave Flaw

When editing a document in TextEdit, a copy of your work is automatically saved every 30 seconds to the hard drive*. This behavior is common in software as it provides a convenient means to recover some of your work should the application unexpectedly quit or crash.

Unfortunately, the means by which TextEdit saves a copy of your work is awfully rudimentary. It simply writes your data to a regular file and gives it the same name as your TextEdit document but with “ (Autosaved)” appended as a suffix (without an extension). And since no verification is performed to see if a file with that particular name already exists, it will overwrite anything that gets in its way with no confirmation or warning!

Example

To better illustrate this flaw, take the following scenario. Suppose, for whatever reason, that I have a file named Craziness (Autosaved) and I create a text document called Craziness.txt in the same directory. In the screenshot below Craziness (Autosaved) is an image file (with the extension removed in order to illustrate my point):

Screenshot of my 2 original files.

When I start editing my Craziness.txt file in TextEdit, the application autosaves my work (as it should), but since my image document has the same name as what TextEdit would call its autosaved file, my image file is overwritten:

Screenshot of TextEdit's autosaved file.

When I’m done editing my Craziness.txt document, TextEdit removes the autosaved copy (as one would expect). However, now my original image file is gone with no real way to recover it (since it’s not moved to the Trash but actually overwritten):

Screenshot showing loss of data caused by TextEdit.

Solutions

Accounting for this file naming issue is so programmatically simple that it’s astounding the defect even exists. The simplest improvement would be to prefix the filename with a period (as in .Craziness (Autosaved)) in order to hide it from the Finder since the chances of having a naming conflict with a hidden file are greatly reduced.

But hiding the file from the user still allows for potential name collisions and as Mac OS X’s default text editor, TextEdit’s naming convention should be even more robust. To start, TextEdit could include either a timestamp or a sequence of random numbers to help make its autosaved filename unique. Most importantly, however, should be to verify if a file would be overwritten and if so, generate a different random number or append an incremental counter. Heck, even my Page Capture widget won’t overwrite files since it uses the same naming convention as Apple’s screencapturing application (File 1, File 2, File 3, etc.)

The Rant

One might argue that the possibility of having a file end with “ (Autosaved)” and not have an extension is pretty slim. So what? My argument is that the possibility of an application deleting other files blindly is a completely unacceptable use case scenario, no matter how rarely it may occur. I think it is more reasonable to expect that a corporation as large as Apple Inc. would produce software that doesn’t delete unrelated files from my hard drive without my knowledge. Especially since OS X is — as Apple claims — the “most advanced operating system in the world.”

*30 seconds is the default. The time interval is configurable and the user is allowed to disable the Autosave feature entirely.

Balancing On The Ethical Line

Monday, January 19th, 2009

As in any engineering profession, ethics is a key component of Software Engineering. In fact my college software development courses at Cal Poly emphasized engineering ethics as much as any of the other topics because although developers need to know about many technical disciplines (such as versioning, the software life cycle, software prototyping, etc.), they also have expectations to meet and responsibilities towards those who use their software — and their ethical foundation is what will define how they face up to these responsibilities.

Knowing that ethics is such an essential and basic rule of engineering, I’m disappointed when I encounter software products (large or small) that don’t adhere to software engineering ethics. Or that companies even hire software developers when they don’t enforce an ethical background. I’ve stumbled upon some recent examples that have bothered me enough to write about.


Forcing By Confusion

When customizing the installation of Mac OS X 10.5 (Leopard), I am given a list of optional applications to install. Some options expand to reveal dependent sub-items, who’s relationship is clearly indicated by the grouping of which they are a part.

Deceptively forcing to install Microsoft Office

As I clicked on several of the check boxes I noticed that the Office 2004 Test Drive option enabled itself automatically (without me having directly clicked on it). This behavior is not what is expected of a checkbox. A check box is suppose to toggle a single option. Only when an obvious dependancy to another item exists should it also change the state of another option.


In this scenario, there is obviously no dependency between choosing to install Office 2004 Test Drive and any other option. If there were, it would be indicated by a hierarchy as it is done with iWeb, GarageBand and iDVD. Bypassing this expected behavior and thus tricking the user is unethical.

Forcing By Deception

After installing iWork ’08 I was, not surprisingly, presented with an option to register* my use of the product with Apple. Unlike in most registration dialogs, I was not given an option to Never Register or Cancel. Rather, I was only allowed to register or postpone my decision.

I thought: “Man, am I really gonna have to click on Register Later for all of eternity?”

Well, it turns out that I am stubborn enough so that is exactly what I decided to do even if it meant dealing with the dialog every time I wanted to use the software that I purchased.


But once I launched the product a third time, the Register Later button turned into a Never Register button, which is exactly what I had wanted in the first place. Why couldn’t that option have been offered initially? Why did they try to deceive me to register before giving me the option to not do so? They could at least have informed me from the get-go that the option would change down the line.

Close-up view of the progression of the available options.

To change the options available to a user without any evident reason to do so is deceptive. It makes the user think that those initial options are the only ones available. I for one, could have easily been convinced to just go ahead and register if I wasn’t as stubborn as I am. Deceiving people with an artificial mandate is unethical because you influence their ultimate decision by omitting valid options.

Forcing By…Forcing

I encountered a similar registration dialog upon installing Aperture and this time the fields contained even more personal information — all of which was pre-filled with data from my address book:

As in my previous example, I clicked on the Register Later button hoping that eventually I would be given the option to Cancel or Never Register. Well, I’ve been clicking on the same button for more than six months now so I think it’s safe to assume that it will never come up. I guess persistence isn’t always a solution.

Even though I’m not forced to register, eventually I might click the Register Now button accidentally. In reality, I have very little control over the situation since all my information is automatically filled in and my options are limited. Always having an option to postpone my registration is a clear indication that it isn’t a necessary step and I should therefore be given an option to forgo it permanently.

If one considers how some viruses work, this conduct doesn’t sound so innocent anymore. While a virus would obtain and send my information without my knowledge or approval, the behavior in use here waits for me to make a mistake and click on the Register Now button by accident. It should be my choice whether to have all my contact information sent to God knows where and in this scenario, I’m not really being given the control that I deserve.

The Rant

So who’s to decide what is ethical or not? Well…we are. Corporations only get away with this crap because users of their products put up with it. The majority of the population would rather “just move on” even if it means complying with an imposed action. I feel differently. I find it annoying and I refuse to conform simply because “it’s more convenient.”

Who knows, maybe the manipulative intentions in the examples above weren’t premeditated or will change in future releases. But it’s important for developers and product managers to realize that little things such as these can damage a company’s image and the trust that users have in their products. If you’re manipulating me in this way, how am I to trust that you’re not doing it in other, dare I say worse ways without my knowledge?

Users need to be aware of when they’re being pushed around unnecessarily and should definitely be mindful of what they click on. After all, we can’t expect corporations to have the user’s best interest in mind since their ultimate concerns are to their stock holders and their profit margins. However, we can expect them to behave ethically and we should hold them accountable for just that.

* Registering is different from activating a product with a serial number and should be optional. The information normally requested is not at all necessary for the operation of the software.

I never register because 99% of the time there is no need or benefit to release the kind of personal information requested.

Apple Security Threat

Thursday, November 6th, 2008

A recent occurrence has made me think twice about Apple’s Target Disk Mode boot option. Indeed it can be a very convenient feature, but like most conveniences this one is riddled with security threats. What is most bothersome, though, is how few people realize the problems it poses — not to mention the simplicity of a solution that Apple does not provide…at least not by default.


For those of you not up to speed, most of Apple’s computers allow themselves to be temporarily turned into an external hard drive simply by pressing the corresponding hot key (‘T’) during boot up. If the computer supports this option (most do) it will enter what is called Target Disk Mode (TDM) and allow itself to become a mass storage device and be connected to another computer via an IEEE 1394 interface (aka FireWire, i.LINK, Lynx…whatever).

Yes, this feature is convenient for transferring large amounts of data or if you need a quick makeshift external hard drive (assuming you have a male-male Firewire cable). Unfortunately, the feature also inherently bypasses the OS from ever being started on your computer allowing others access to all sorts of files that you assumed were secure by the OS’s login.

How It Works

When you press the power button on your computer the first thing to come to life is the firmware (a very low level program that lives in the hardware) and it decides what happens next — whether to boot into the installed OS, boot from a CD, boot from a network drive, etc. The decision is based on multiple factors, one of which is to check for certain hot keys on the keyboard.

The Problem

The problem with this convenience is that anyone with a finger has the ability to transform your computer into a large external drive. Yeah, including that person that just walked away with your laptop while you were getting another soy latte at Star Bucks.

Some would argue that if I’m this concerned with the security of my files, that I should enable FileVault in order to encrypt every file on my hard drive. Yeah? Well, I don’t think I should have to enable something that will have incredible amounts of overhead just because a back door exists that can completely circumvent the OS’s login prompt.

Solution (but not really)

Firmware Password Utility ApplicationThe solution is simple: eliminate the hot keys from influencing the firmware’s decision. Welding a steel plate on top of your keyboard would work I guess, but that’s not very convenient. A better idea would be to tell the firmware to not check the hot keys.

Currently, there is no way to disable these hot keys, but it turns out there is a way to password protect the firmware with some extra software. But after reading Apple documentation that states that the firmware password can be circumvented (quite easily), and that it could in fact be hazardous to your system, and that it is temperamental, I disabled it on my machine and don’t recommend it. Way to fuck us over, Apple:

“WARNING: Open Firmware settings are critical. Take great care when modifying these settings and when creating a secure Open Firmware password.”

“An Open Firmware password provides some protection, but it can be reset if a user has physical access to the machine and changes the physical memory configuration of the machine.”

“Open Firmware password protection can be bypassed if the user changes the physical memory configuration of the machine and then resets the PRAM three times (by holding down Command, Option, P, and R keys during system startup).”

The Rant

First of all, I think that the extra Firmware Password Utility (not included in a default installation…but available from the software installation disc (/Applications/Utilities/) and online) should not be necessary. I think there should be a simple check box in the System Preferences that enables/disables whether or not the keyboard is “heard” by the firmware.

I also think that the hot keys should be disabled by default. Apple is all about an ‘out of the box, ready to go’ mentality so I suspect they leave the feature enabled by default because that makes it more convenient for their users to make use of the TDM functionality. We’ve seen this same behavior before, but I think the security threat outweighs the convenience factor. Tisk, tisk Apple.

A More Intuitive iPod Shuffle Switch

Wednesday, October 1st, 2008

The iPod Shuffle is a wonderful little product and in my opinion is by far the best mp3 player that doesn’t display information about what is playing. But there is an element of its design that I consider to be flawed and which I attribute to Apple’s consistent choice to sacrifice options for the sake of simplicity.

iPod Shuffle Close Up


One of the two switches on the iPod Shuffle chooses the play mode: either Continuous Playback or Shuffle. The former will repeatedly loop through all the songs in the order that they were added to the iPod from iTunes. The Shuffle mode will obviously play through them randomly, but will it stop playing after all the songs have been played through once? Are Shuffle and Continuous Playback mutually exclusive?

Different Shuffle Symbol Suggestions for the iPod ShuffleThis conflict seems minor, but there might be a major design flaw here. A switch is a user interface element that chooses between two possibilities. Ideally, neither choice implies the other (or else a different user interface element would be used…more about this later). But in the case of the iPod Shuffle, the Shuffle mode implies continuous playback as well. And I agree that it should—it’s intuitive to me that Shuffle would also continuously loop through all the songs. But with the current setup (Default Shuffle Symbol), however, moving the switch from Continuous Playback to Shuffle is contradictory. I would have at least made the symbol something like: Alternate Shuffle Symbol Suggestion 1 or Alternate Shuffle Symbol Suggestion 2.

You might now be saying to yourself: “That’s all fine and dandy because the current symbols on either side of the switch are simply graphical representations of two preset modes: one that continuously loops through the songs in order and another that continuously loops through the songs but in a random order.” Well, let’s not be so quick to make that assumption because as much as I’d agree with you, that’s not the way iTunes does it. iTunes uses buttons to select the mode allowing the user to turn both Shuffle and Continuous Playback on at the same time.

Playback choices in iTunesAnd let’s not kid ourselves that it is ok to have iTunes do something different than all the iPods out there. Apple’s success with their line of mp3 players (as well as their other products) is almost entirely attributable the almost necessary connection between their hardware and software components.

Polarization and Hair Color

Tuesday, May 27th, 2008

I’d like to share something that I’ve observed quite often in my experiences with a polarizing filter, but that is not a common topic of articles that I’ve read. I hear a lot about polarization’s abilities to reduce glare, darken the blue in the sky, and enhance the greens in plants. However, rarely do I hear about it’s ability to make hair color (particularly red) significantly ‘pop.’


A circular polarizing filter is like a pair of polarized sunglasses for your camera. Polarization relies on the fact that certain forms of light (such as most reflected light) travel in a uniform manner. Due to this consistency, a polarized lens is able to filter out a specific group of light waves while allowing the remaining light waves (usually what we care about) to pass through with less disturbance. To illustrate this behavior, here’s a quick example of what polarization is most known for: it’s ability to significantly reduce, and sometimes completely remove, reflections so that you can see “through” glass and water.

Glare and reflections obscure subject matter without a polarizing filter.

A polarizing filter can reduce glare and reflections.

The two pictures above are from a recent trip to Rome where I caught three police officers reading about a soccer match in their car. You can’t see the third officer in the passenger seat in the first picture because of some intense reflections on the windshield. But with the polarized filter adjusted correctly in the second picture, the reflections are almost completely removed and the light waves that I care about (those bouncing off of the person in the car) are able to hit my camera’s sensor without the interference of the reflections.

Now let me show you an interesting example of how this same process of reducing reflections in shiny hair can significantly saturate it’s color. The following pictures have not been graphically altered and are straight out of my camera. The difference you see in the colors are a result of the polarizing lens blocking out the reflections in the woman’s hair!

A polarizing filter can saturate hair color.

The only down side that I can think of is that objects with reduced reflections tend to flatten a bit. Reflections naturally help our eyes perceive the 3D, spacial dimensions of an object. Removing these reflections sometimes has an equal result of flattening the image. I’ve zoomed into an area of the pictures above to illustrate this point. Notice below how the reflections in the woman’s hair on the left help define the curvature of her head and how it is somewhat lost in the picture to the right.

A polarizing filter can also flatten shapes.

Though, I’ll take the saturated colors over the curvature 98% of the time!