I read TFA, and I can't believe that this guy is advocating software engineering. I reread an old, but not too old, paper by Edsgar Dijkstra called "On the Cruelty of Really Teaching Computer Science". In it, he refers to software engineering as "how to program if you cannot." That one had me rolling on the floor because, more often than not, it agrees with my own observations and experience.
Our university teaches software engineering. The professor who teaches most of the software engineering courses is an idiot. For my project, my group and I wrote a pretty kick-ass app with some pretty kick-ass code. How? We snuck in some agile methodology, which just seemed perfect given the size of the group and the size of the project. Even though our project was the best in the class, we got a C because the specifications weren't complete, and she wasn't convinced that our automated tests would really test the code. (The other groups who got A's for testing wrote some very non-specific paragraphs about how they might test their code. None of them actually did any testing)
What this proves to me is that software engineering is easily the most useless discipline in computer science. I have never had a good experience with it. I've never known anyone who said that software engineering really makes things run smoothly. It's a business-centered/management-centered unrealistic approach to software development. It may make your boss feel all warm and fuzzy, but it won't get the software out the door on time, nor will the developers have any degree of confidence in it.
I was chatting with my thesis advisor just the other day, and we were talking about how much it bothers both of us that students see a degree as simply a means to an end. Having a job is nice, but if you're not interested in learning the SCIENCE in computer science, then your university experience will be four plus years of hell.
You see, when people tell you that the skills you learn in a CS program will be out of date in five years, they're absolutely right. So what do you do? You don't go to school for skills. You go to school to learn how to learn. You go to learn the science that undergirds computer science, which is independent of any particular technology, programming language, computer, what-have-you.
Theory of Computation was my favorite course as an undergrad -- and it was again when I took it as a grad student. Now, a lot of what you learn in the beginning of that course -- regular languages, context-free languages -- is really practical in a real job, particularly if you ever implement an interpreter or compiler. (If you don't think that happens, I have an uncle who developed a special purpose language specifically for energy futures traders.)
But the stuff you learn later in that course -- computability, complexity, and all that good stuff -- it may not be as practical, but it's sure a hell of a lot of fun, and it becomes practical because it influences the way you see a computer or program. It enhances your theoretical understanding of what a computer can and can't do. You tend to waste a lot less time trying to develop an (efficient) algorithm to solve an unsolvable problem.
Here's the way I see a university education in CS: if you wanted to be an auto mechanic, you'd go to a trade school for a couple of years to learn the skills you'll need. If you want to become an automotive engineer, you go to a university for an engineering degree: it requires a different kind of understanding to be an engineer than a mechanic. The same is true of CS. If you want to be a "programmer", go to a community college and get your MCSE. If you want to be a professional computer scientist, you go to a university and get a CS degree. It's all a matter of what you want to do.
That said, a university is certainly not the right place for everyone. There's nothing wrong with going to a community college. But you have to understand the benefits and costs of doing one instead of the other.
And did you read what that dude's job title is? Do you think that the "head of technology in Asia Pacific for antivirus firm Sophos" might have a vested interest in saying that Macs need antivirus software?
Here's another handy tip: if you want to save yourself a few more MB in Windows, completely disable the Themes service. Plus, you can force anyone else who uses that computer to live without eye candy. It pisses off the design students in the lab I run. It makes me smile.
You're right about using the command line. I use the command line myself, even though I've got a nice (allegedly user friendly) OS X computer. But I would argue that there are more than two classes of computer users (beginner and advanced) and that most users lie somewhere in between. That is, most users are neither novices nor experts, so the user interface should be geared toward them more than anyone else. It's like a bell curve, and by hitting that "intermediate level" sweet spot, you can cover 80% of users.
I think that most UIs -- OS X's in particular, and even sometimes Windows (but not Office) -- do a good job of making those "intermediate" features available to the people who know how to use them. For example, I think that toolbars are great for beginner users, but you can cram a lot more functionality into a menu. That's why toolbars are big and graphic, and menus are small and textual.
I've never been a fan of the "three tiered" interface. Some apps do it -- I think Eudora has something like that. But as I recall, if you're using "advanced" mode and you're explaining something to someone who's using beginner or intermediate mode, they may not have the same options in front of them. I think a better solution, and one that accomplishes approximately the same thing, is to design the interface so that those advanced features are more or less invisible to, or out of the way of, the beginner users, and they can discover how to use them just by exploring. (That's also why it's important to have a big visible undo button and make it consistent across all apps so users don't have to worry about screwing things up. My mom, for example, never learned how to use cut and paste because she was afraid of ruining her documents. So instead, I showed her how to use the undo button, and knowing that she had a fallback, became much more adventurous in learning how to use her apps. Now she can cut and paste like a pro.)
I don't think having a nice interface to a CLI would necessarily make it easier to use for those intermediate users. The problem with the CLI, as you say, is that it is byzantine. Want to output a file to the screen? Then use cat of course. Isn't it obvious? Want directory contents, then it's ls for you.
For those intermediate users, there are a couple of solutions that can do some pretty advanced things. I've already named some of them, menus being the most obvious. But what about your particular task of renaming a batch of files? I don't know if you've had a chance to use Automator in OS X Tiger, but that's exactly what it's designed to do: give you a nice graphical interface to do scripting-type things without having to know the language. Granted, there are those of us who know that the scripting language is still faster, so there's Terminal.app and bash for us.
Anyway, the idea behind Automator is that you make "flows". Data (files, usually) come from one source, like the Finder -- which can be generalized to accept data from any folder, so a flow is reusable -- and it goes into other workflow objects that change it in some way, like renaming files, resizing images, emailing things. (There's a neat little sample script that takes your unread messages from Mail.app, concatenates them in an iPod note, then uploads itself the next time you synch your iPod. Poor/cheap man's Palm!) You can write apps for Tiger that expose actions to Automator so people can script your program with it. I imagine that things like that occupy the space that is both understandable to intermediate users, and powerful enough for advanced users. Plus, the way in which it interacts with one's graphical apps is obvious. It opens the apps to run the appropriate workflows. Nifty, no?
It would be nice to be able to pipe a list of files into a *nix command, though. I definitely agree with you on that. That could probably be done by simply adding a "copy file names" command to the menu or something like that.
Not only that, but this guy really doesn't understand the things that current HCI does well. For example, that options should be visible (well, the screen corners are visible... but they're really not distinct), feedback (moving the cursor to the screen corner doesn't immediately produce an obvious effect; there must be a delay to avoid slip-ups), and promoting recognition over recall.
I'm going to harp on the last one. We have menus with titles because we can recognize the type of functionality we want. We've got toolbars and file navigators with icons so we can -- at a glance -- recognize what we're looking for. A screen corner has none of these properties. It's just one of those things you have to "remember", which means that people won't remember. What's there to remind them? If you used a screen corner for some real functionality, within minutes, you'd have people posting "cheat sheets" on the web to help people remember.
Remember back in the (bad) old days when you needed a keyboard template just so you could remember what key does what in your apps? Modern HCI says that's bad, but this guy wants to bring a slightly more sophisticated version back.
By the way, Mac OS X has the same hotspot corner thing. So Apple must have lifted that from After Dark as well.
You could always bootstrap your own compiler. How fluent are you with hexadecimal opcodes on your architecture of choice? Make an assembler with plain hex, then assemble progressively better assemblers with that, then you'll be at a point where you can make your very own little C compiler. Build it up, little by little, and before you know it, you'll have your very own fancy-pants compiler to write your TI-99/4A OS (which I used to own, by the way. There were some fun games. Oh, the things you'll find at a flea market in northwest Iowa.).
The problem is, IIRC, that Creative applied for the patent in January 2001. The iPod wasn't even on the market until (I think) October 2001.
However, the idea of hierarchical navigation and using metadata to group things is obviously prior art -- evidently no one at Creative has ever used a directory tree. So, the test will be if it makes a difference that metadata from the files is being used to create that navigation tree (although that's not how the iPod stores files internally).
It just goes to show how broken the patent system is when such a simple and obvious idea, with so much prior art, can be granted a patent and used to sue a competitor in a "we can't beat you, so we're going to litigate you to death" sort of way. Can't Creative be more, uh, creative than that?
Part of my senior thesis when I was an undergrad was to research the Mac OS dynamic recompiler. It was truly remarkable that they built such a thing at an extremely low level -- like emulating device drivers, even.
VLIW is really nice. If you ever have a chance to play with an Itanium -- OK, Itanium 2 -- take it. Use the -S flag (I think that's what it is) in gcc, and prepare to be amazed by the assembly that comees out. Even though VLIW could make things easier, the underlying problem of turning machine code into an IR and then outputting a different machine language is still just as tough. But with DR, it's a sort of "one time investment" as you run a program.
AMD makes some of the best processors on the planet. I'm a fan of anyone who can build a better Intel processor than Intel. When he announced the Intel switch, Steve said that he was most impressed by the Intel roadmap. That means one of two things: either AMD can't maintain the level of output that Apple would require (likely) or Intel's got some kick-ass processor coming down the pike that will blow AMD away (unlikely). It may simply be that Intel has better brand name recognition (years upon years of that five note Intel thingy at the end of almost every computer commercial gets in peoples' heads, I imagine). In either event, Apple and AMD both occupy the same mindspace in most geeks -- better, but underappreciated. I'm thinking that Apple's trying to get out of that space by favoring Intel over AMD.
And though Tiger lets you create Smart Folders (saved searches), this feature is neither easily discoverable nor particularly integrated into the system. Specifically, Tiger doesn't ship with pre-made Smart Folders for commonly-accessed searches.
OK, so the argument here is that one of Vista's big advantages over Tiger is that it ships with pre-made Virtual Folders. I can think of lots of reasons why Apple didn't do that.
Apple's fervently pursuing switchers, users who are new to the Mac. Try explaining the difference between folders and smart folders to someone who's not, as people often say, "good with computers." Tell them something like, "well, OK, you see, the file's there, but it's not really there. It's actually in a real folder somewhere else." You're likely to get a glazed expression from that one, and possibly an existential argument about "is anything really where it is?"
The moral: smart folders are an advanced feature. People who want them will know how to find them. People who don't understand them won't have to worry about them.
Again, from TFA:
In Tiger, there is no easy or obvious way to edit meta data for the documents and other data files you create, and you typically have to rely on document processing applications (such as Microsoft Word) to add and edit this information.
Spotlight relies on Spotlight Importers, little bundles of code that know how to read files and return metadata about them. More often than not, the importers are written by the original application designer, who should know better than anyone what bits of data are most important in a document. Apple's implicit position is that metadata should be either derived from the document on its own, or that metadata should be provided in some manner by the creating application (which the importer can then retrieve).
Again, should people have to care what "metadata" is? There are lots of ways the programs themselves can gather all the metadata you'd care about. Standard info, such as the file's author and what-not, can easily be provided automatically by the program. That's the way it should be, because programs can automatically add relevant metadata that improves searches without the user ever having to do a thing. Plus, there's a matter of confidence. If Vista's got a great big box for me to enter metadata, should I take that to mean that there's a good chance Vista doesn't really know how to index my files? If that's the case, then forget about it. I'm not going to add metadata to every document I've ever written just so I can find it.
The moral of the story is this: having a wide arsenal of tools is great. But many users don't know how to use them, don't need them, and don't much care to learn. Vista seems to favor forcing users to learn how to use these new features. A forcing function is a good idea sometimes, but forcing users to use features that just complicate their experience is foolishness. The crux of Thurott's complaints against Tiger is that it's not complicated enough. There aren't enough exposed features. I've learned that in UI design, the more buttons you give someone to push, the better the chance is that they'll pick the wrong one, and the better the chance they'll blame you for it. And they'll be right.
What is impressive, however, is that when the first PPCs came out, they actually ran a 680x0 version of the Mac OS -- mainly because no PPC version existed. The whole OS ran in a dynamic recompiler, which is like an emulator, but it actually translates the non-native code into native code on the fly so it can reuse the portions already translated.
Plus, remember that when the PPC Macs first came out, there was literally no software for them, and Apple had an installed base of millions of 680x0 Macs to support. Fat binaries didn't please everyone -- hard drive space being expensive in those days -- but it made it easy for average consumers to just buy a program for their Mac, PPC or 680x0, and know that it would work.
The difference this time is that Apple's been planning ahead. NeXT made an Intel version of NeXTSTEP/OpenSTEP, so it cost them very little to develop OS X for Intel in parallel with OS X for PPC. Again, they're using a dynamic recompiler (Rosetta) so that PPC apps will run on Intel-based Macs. This time around, Apple is much better prepared to make this transition. The goal, as it was the last time, is to make it easy for Mac users to pick up a piece of off-the-shelf software for the Mac and know that it will run, regardless of the processor inside. On the developer boxes, as I understand it, many of Apple's own apps are still PPC-only and are running in Rosetta; iTunes is a prime example of this. Plus, the particular reason Apple's going with Intel right now is because they know that they can build more impressive computers -- right now.
Ask any sociologist -- if you dare -- about standardized testing. They'll tell you that it's utter crap because all standardized tests are biased toward the writer. That is, if men write IQ tests, then men will do better on it.
What would be interesting, although a different study altogether, would be to assemble a group of women similar to the group of (presumably) men that write IQ tests now. Find yourself a couple of groups of men and women to take each test, and SUPRISE! the women will do better on the test written by women.
There have been numerous studies just like this that show that other kinds of standardized testing (of the "No Child Left Behind" variety) is biased against women and minorities, and that the scores women and minorities earn on these tests are not representative of their actual abilities.
What I've learned from all of this is that intelligence, or even knowledge in a particular area, is impossible to quantify. The problem is that individuals use different mental models to understand their world. If they don't use the same mental model as the test-writer, then they won't do as well.
HAHA. I think I might still have that text around somewhere. Fortunately, the professor who taught the intro AI class was well-known for being an easy A, so I never felt the need to read more than about two or three pages.
Although, come to think of it, his PowerPoint slides were usually pretty sleep-inducing.
It's been a really long time since I've smoked. Although I think I did sleep better in those days. Hmm... wonder if I still know any dealers...
A while back I was prescribed a tri-cyclic anti-depressant which was to help me sleep at night (doc thought the insomnia might be related to some form of depression), I forget the name of it but it works as a sleeping pill, so I know the groggy feeling in the morning, like there's ambisol in your head.
It wasn't trazodone, was it? I took that, and as I described in an earlier post, made me into a groggy basketcase the next day.
Plus, I invested in an expensive espresso machine. There's no way I can give up coffee altogether.
You don't want drugs unless you like to feel like ass the next morning. Believe me, I've been there. Even the so-called "next-generation" sleep drugs -- your Ambiens, Sonatas, and what-have-you's -- can still make you feel lousy.
I used to take one -- very briefly -- that actually made me into an emotional basketcase the next day. Literally can't stop crying, can't explain why you're crying in the first place, plus usually a nasty headache. I used to work with a couple of my good friends, and they said it's the most bizarre they've ever seen me act -- and they've seen me drunk.
Actually, yes. I've actually never been healthier. I've totally given up soda and sugary crap, and I exercise about four times a week, usually in the afternoon. My one vice is still coffee, but only in the morning.
Oddly enough, my boss suggested having a beer before bed, too. But, my doctor says that's a bad idea. It makes your sleep more restless and you're more likely to wake up in the middle of the night.
As someone who suffers from chronic insomnia -- and yeah, I've gone through all the medical nonsense for them to tell me there's nothing wrong with me physically or emotionally -- having a drug to counteract the effects of sleep deprivation sounds like a godsend. For me, all the sleep deprivation effects in the world can't help me fall asleep. For example, I finally fell asleep at about 5:30 this morning and had to get up about an hour and a half later for work.
Some of us are jealous of the relative ease with which the rest of you fall asleep. (The absolute worst is sharing a hotel room after a long trip, where your traveling companion falls asleep right away, but you don't fall asleep for hours) I'd be happy to at least feel as awake as most people seem. The only time I feel that way is when I can sleep in on the weekends. It's mostly just depressing that I can't be that alert the rest of the week -- you know, when it matters most.
You laugh at that, but I know people who *really think* that using microorganisms to do our bidding is a form of slavery. There are people who are honestly opposed to using recombinant organisms to produce oh, say, insulin.
And you can't point out that those organisms exist in an environment where they reproduce in numbers that would never be possible in nature. According to them -- well, one person in particular -- "it doesn't matter how loosely we hold the whip; it's still a form of slavery." That's more or less a direct quote.
And when you point out the symbiotic relationships that already exist in most living organisms -- animals in particular -- they refuse to believe that it's true. Something about the military-industrial complex of medicine trying to make us believe what they want us to believe or some crap like that.
It's hard not to laugh when someone tells me that humans are evil evil creatures for creating antibiotics and vaccines. Because someone think of the poor bacteria and viruses! I only wish I were joking right now.
Even in conservative northwest Iowa, the policy was very plain: you can look at anything you want, but you will be asked to look at something different, or asked to leave the computer if anyone complains.
And, they had what I consider to be one of the most sensible policies for monitoring children I've ever seen: for children under 13, a parent or other guardian *must* be present for the kid to be allowed to use the computers.
The rationale was, well, rational: as a parent, you know what you consider to be inappropriate for your child, so it's up to the parent to make the judgment call.
I haven't been back to Sioux City in quite some time, and I certainly haven't been in the library, so I'm not sure if they've changed their policies. What I really liked about their policies is that they accounted for the unspoken question of appropriateness: appropriate for whom? The courts are never going to be able to decide what little Johnny's parents think is appropriate for him, so let/force them to decide. That's why I think it's crazy that librarians are expected to parent other people's children for them while they're at the library.
Not only that, but Dillard's (a department store I think you mostly find in the midwest/east coast) puts an individual proof-of-purchase barcode on everything you buy at the cashwrap. It links the product to the particular sale, and to the salesperson. It's also probably a lot harder to substitute an inferior product, as the stickers tend to "self-destruct" when you remove them (they're the kind that tear into hundreds of little pieces if you try to remove them)
It's the same principle, and they won't accept anything for a return unless it has a proof of purchase sticker they can verify with their system. Nice thing about that is that if you don't have the receipt, they can still verify it with the sticker and an ID.
Idiot Retailer: Sorry, we can't accept this for a return.
Me: But the receipt doesn't say I can't return it.
IR: Sorry. It's on our super-secret list of non-returnable items.
Me: Well, I don't like to make a big deal out this, but I'm an attorney, and I think your business practices border on fraudulent.
IR: So, would you like cash or store credit?
It also used to work with telemarketers before the "do-not-call" list went into effect. They don't know you're really not a lawyer, but if you dress nicely and look (and/or sound) old enough, they'll believe anything.
Actually, you can ban your enemies, people you've fired, or those who complain about unfair practices unless there's specific legislation to stop you. Never seen protesters escorted out of a store, have you? Hell, until the US passed a friggin Constitutional amendment, it was legal to deny even the the most basic of public services to people on the basis of their race. (By the way, there are a lot of groups that aren't covered by that amendment. And even for the ones that are, it's not exactly a cakewalk.)
And your mythical bridge would be neither a toll bridge, nor privately owned if it's the only way across town. In fact, when have you ever seen privately owned roadways? Governments exist for the public good, not businesses. Businesses exist to make a profit. But everyone's got a profit motive, haven't they? Even (especially?) the Telus union. After all, they're holding out for more money. I doubt they give a shit about the public good if they're willing to undermine the company by threatening its employees.
BTW: Telus, you know, used to be a public entity. Just like BT. If you want to blame someone for their monopoly status, blame the Canadian government. Competition is good. So are bananas. Bananas are good.
Your analogy doesn't hold, because the Telus employees are using their employer's equipment to attack them. I'm having a hard time imagining how you're going to use your green hair and chlorinated water to attack the water department.
Here's a (perhaps) better analogy: if I own a hardware store, and I rent a sledge hammer to you (yeah, it's a weird hardware store... I know) and you use it to smash up the place and threaten my employees, don't you think, just maybe, it wouldn't be entirely unreasonable for me to ask for it back. I'll give your deposit back, even.
I guess what I want to know is this: why should a company invest its resources and money into hosting a site that calls for its demise? As an executive in that company, it would be wrong of them to leave it up there, for the safety and wellbeing of their other employees, as well as themselves.
I'll bet there are other people in that union who are embarrassed by this little happening, as it gives all of them a bad name, not just the ones who support such an action. In my experience, threatening the people who have what you want makes them all the more determined not to give it to you.
Hell, in the long run, it may be better for the union employees who posted it in the first place. If there are any inaccurate accusations intended to damage Telus's reputation, they could easily be sued for libel (if Canada's laws are anything like Britain's). I really don't understand why this is some big "censorship" argument. I see this as a reasonable action by a company protecting its own interests against those who would harm it.
On the face of it, if that large a percentage were using Macs, Apple would have shown tremendous market share gains in its past several quarterly announcements
Not necessarily. There have been plenty of studies that show that Apple owners buy computers less frequently than PC owners. Market share is a measure of relative volume of sales, not install base.
If, like my University, you buy new Macs every 3-4 years and new PCs every 2-3 years, in 10 years, you would have only bought 2-3 Macs, and somewhere between 3-5 PCs. And, we easily have three times as many PCs as Macs, so at the most extreme, in ten years, you would buy 2 Macs and 3 * 5 = 15 PC's. (And an education-oriented college campus is probably a lot more Mac-heavy than other institutions) 2/17 implies that 11.7% of our computers are Macs, but the actual ratio is 1/4, or 25%.
That's one of the main reasons I always caution people about using market share statistics to judge install base. It's simply not accurate, but it is the most readily-accessible tool.
I agree. If I were to imagine any video-based product coming down the pipeline, it would most likely take the form of a video analogue to the AirPort Express. I love my AP-E, and I think it would be fantastic if I could watch movies on it as well.
I don't think people really will watch movies on their iPods, if for no other reason than the situations in which you would listen to an iPod are generally times when you can't/shouldn't be watching a screen.
Slashdot Mirror
Our university teaches software engineering. The professor who teaches most of the software engineering courses is an idiot. For my project, my group and I wrote a pretty kick-ass app with some pretty kick-ass code. How? We snuck in some agile methodology, which just seemed perfect given the size of the group and the size of the project. Even though our project was the best in the class, we got a C because the specifications weren't complete, and she wasn't convinced that our automated tests would really test the code. (The other groups who got A's for testing wrote some very non-specific paragraphs about how they might test their code. None of them actually did any testing)
What this proves to me is that software engineering is easily the most useless discipline in computer science. I have never had a good experience with it. I've never known anyone who said that software engineering really makes things run smoothly. It's a business-centered/management-centered unrealistic approach to software development. It may make your boss feel all warm and fuzzy, but it won't get the software out the door on time, nor will the developers have any degree of confidence in it.
You see, when people tell you that the skills you learn in a CS program will be out of date in five years, they're absolutely right. So what do you do? You don't go to school for skills. You go to school to learn how to learn. You go to learn the science that undergirds computer science, which is independent of any particular technology, programming language, computer, what-have-you.
Theory of Computation was my favorite course as an undergrad -- and it was again when I took it as a grad student. Now, a lot of what you learn in the beginning of that course -- regular languages, context-free languages -- is really practical in a real job, particularly if you ever implement an interpreter or compiler. (If you don't think that happens, I have an uncle who developed a special purpose language specifically for energy futures traders.)
But the stuff you learn later in that course -- computability, complexity, and all that good stuff -- it may not be as practical, but it's sure a hell of a lot of fun, and it becomes practical because it influences the way you see a computer or program. It enhances your theoretical understanding of what a computer can and can't do. You tend to waste a lot less time trying to develop an (efficient) algorithm to solve an unsolvable problem.
Here's the way I see a university education in CS: if you wanted to be an auto mechanic, you'd go to a trade school for a couple of years to learn the skills you'll need. If you want to become an automotive engineer, you go to a university for an engineering degree: it requires a different kind of understanding to be an engineer than a mechanic. The same is true of CS. If you want to be a "programmer", go to a community college and get your MCSE. If you want to be a professional computer scientist, you go to a university and get a CS degree. It's all a matter of what you want to do.
That said, a university is certainly not the right place for everyone. There's nothing wrong with going to a community college. But you have to understand the benefits and costs of doing one instead of the other.
And did you read what that dude's job title is? Do you think that the "head of technology in Asia Pacific for antivirus firm Sophos" might have a vested interest in saying that Macs need antivirus software?
Here's another handy tip: if you want to save yourself a few more MB in Windows, completely disable the Themes service. Plus, you can force anyone else who uses that computer to live without eye candy. It pisses off the design students in the lab I run. It makes me smile.
I think that most UIs -- OS X's in particular, and even sometimes Windows (but not Office) -- do a good job of making those "intermediate" features available to the people who know how to use them. For example, I think that toolbars are great for beginner users, but you can cram a lot more functionality into a menu. That's why toolbars are big and graphic, and menus are small and textual.
I've never been a fan of the "three tiered" interface. Some apps do it -- I think Eudora has something like that. But as I recall, if you're using "advanced" mode and you're explaining something to someone who's using beginner or intermediate mode, they may not have the same options in front of them. I think a better solution, and one that accomplishes approximately the same thing, is to design the interface so that those advanced features are more or less invisible to, or out of the way of, the beginner users, and they can discover how to use them just by exploring. (That's also why it's important to have a big visible undo button and make it consistent across all apps so users don't have to worry about screwing things up. My mom, for example, never learned how to use cut and paste because she was afraid of ruining her documents. So instead, I showed her how to use the undo button, and knowing that she had a fallback, became much more adventurous in learning how to use her apps. Now she can cut and paste like a pro.)
I don't think having a nice interface to a CLI would necessarily make it easier to use for those intermediate users. The problem with the CLI, as you say, is that it is byzantine. Want to output a file to the screen? Then use cat of course. Isn't it obvious? Want directory contents, then it's ls for you.
For those intermediate users, there are a couple of solutions that can do some pretty advanced things. I've already named some of them, menus being the most obvious. But what about your particular task of renaming a batch of files? I don't know if you've had a chance to use Automator in OS X Tiger, but that's exactly what it's designed to do: give you a nice graphical interface to do scripting-type things without having to know the language. Granted, there are those of us who know that the scripting language is still faster, so there's Terminal.app and bash for us.
Anyway, the idea behind Automator is that you make "flows". Data (files, usually) come from one source, like the Finder -- which can be generalized to accept data from any folder, so a flow is reusable -- and it goes into other workflow objects that change it in some way, like renaming files, resizing images, emailing things. (There's a neat little sample script that takes your unread messages from Mail.app, concatenates them in an iPod note, then uploads itself the next time you synch your iPod. Poor/cheap man's Palm!) You can write apps for Tiger that expose actions to Automator so people can script your program with it. I imagine that things like that occupy the space that is both understandable to intermediate users, and powerful enough for advanced users. Plus, the way in which it interacts with one's graphical apps is obvious. It opens the apps to run the appropriate workflows. Nifty, no?
It would be nice to be able to pipe a list of files into a *nix command, though. I definitely agree with you on that. That could probably be done by simply adding a "copy file names" command to the menu or something like that.
I'm going to harp on the last one. We have menus with titles because we can recognize the type of functionality we want. We've got toolbars and file navigators with icons so we can -- at a glance -- recognize what we're looking for. A screen corner has none of these properties. It's just one of those things you have to "remember", which means that people won't remember. What's there to remind them? If you used a screen corner for some real functionality, within minutes, you'd have people posting "cheat sheets" on the web to help people remember.
Remember back in the (bad) old days when you needed a keyboard template just so you could remember what key does what in your apps? Modern HCI says that's bad, but this guy wants to bring a slightly more sophisticated version back.
By the way, Mac OS X has the same hotspot corner thing. So Apple must have lifted that from After Dark as well.
You could always bootstrap your own compiler. How fluent are you with hexadecimal opcodes on your architecture of choice? Make an assembler with plain hex, then assemble progressively better assemblers with that, then you'll be at a point where you can make your very own little C compiler. Build it up, little by little, and before you know it, you'll have your very own fancy-pants compiler to write your TI-99/4A OS (which I used to own, by the way. There were some fun games. Oh, the things you'll find at a flea market in northwest Iowa.).
Want to see a kind of cool example of bootstrapping? Go here: http://www.rano.org/bcompiler.html
P.S. I know you were being facetious, but the idea of bootstrapping really is a lot of fun.
However, the idea of hierarchical navigation and using metadata to group things is obviously prior art -- evidently no one at Creative has ever used a directory tree. So, the test will be if it makes a difference that metadata from the files is being used to create that navigation tree (although that's not how the iPod stores files internally).
It just goes to show how broken the patent system is when such a simple and obvious idea, with so much prior art, can be granted a patent and used to sue a competitor in a "we can't beat you, so we're going to litigate you to death" sort of way. Can't Creative be more, uh, creative than that?
Part of my senior thesis when I was an undergrad was to research the Mac OS dynamic recompiler. It was truly remarkable that they built such a thing at an extremely low level -- like emulating device drivers, even.
VLIW is really nice. If you ever have a chance to play with an Itanium -- OK, Itanium 2 -- take it. Use the -S flag (I think that's what it is) in gcc, and prepare to be amazed by the assembly that comees out. Even though VLIW could make things easier, the underlying problem of turning machine code into an IR and then outputting a different machine language is still just as tough. But with DR, it's a sort of "one time investment" as you run a program.
AMD makes some of the best processors on the planet. I'm a fan of anyone who can build a better Intel processor than Intel. When he announced the Intel switch, Steve said that he was most impressed by the Intel roadmap. That means one of two things: either AMD can't maintain the level of output that Apple would require (likely) or Intel's got some kick-ass processor coming down the pike that will blow AMD away (unlikely). It may simply be that Intel has better brand name recognition (years upon years of that five note Intel thingy at the end of almost every computer commercial gets in peoples' heads, I imagine). In either event, Apple and AMD both occupy the same mindspace in most geeks -- better, but underappreciated. I'm thinking that Apple's trying to get out of that space by favoring Intel over AMD.
OK, so the argument here is that one of Vista's big advantages over Tiger is that it ships with pre-made Virtual Folders. I can think of lots of reasons why Apple didn't do that.
Apple's fervently pursuing switchers, users who are new to the Mac. Try explaining the difference between folders and smart folders to someone who's not, as people often say, "good with computers." Tell them something like, "well, OK, you see, the file's there, but it's not really there. It's actually in a real folder somewhere else." You're likely to get a glazed expression from that one, and possibly an existential argument about "is anything really where it is?"
The moral: smart folders are an advanced feature. People who want them will know how to find them. People who don't understand them won't have to worry about them.
Again, from TFA:
Spotlight relies on Spotlight Importers, little bundles of code that know how to read files and return metadata about them. More often than not, the importers are written by the original application designer, who should know better than anyone what bits of data are most important in a document. Apple's implicit position is that metadata should be either derived from the document on its own, or that metadata should be provided in some manner by the creating application (which the importer can then retrieve).
Again, should people have to care what "metadata" is? There are lots of ways the programs themselves can gather all the metadata you'd care about. Standard info, such as the file's author and what-not, can easily be provided automatically by the program. That's the way it should be, because programs can automatically add relevant metadata that improves searches without the user ever having to do a thing. Plus, there's a matter of confidence. If Vista's got a great big box for me to enter metadata, should I take that to mean that there's a good chance Vista doesn't really know how to index my files? If that's the case, then forget about it. I'm not going to add metadata to every document I've ever written just so I can find it.
The moral of the story is this: having a wide arsenal of tools is great. But many users don't know how to use them, don't need them, and don't much care to learn. Vista seems to favor forcing users to learn how to use these new features. A forcing function is a good idea sometimes, but forcing users to use features that just complicate their experience is foolishness. The crux of Thurott's complaints against Tiger is that it's not complicated enough. There aren't enough exposed features. I've learned that in UI design, the more buttons you give someone to push, the better the chance is that they'll pick the wrong one, and the better the chance they'll blame you for it. And they'll be right.
Plus, remember that when the PPC Macs first came out, there was literally no software for them, and Apple had an installed base of millions of 680x0 Macs to support. Fat binaries didn't please everyone -- hard drive space being expensive in those days -- but it made it easy for average consumers to just buy a program for their Mac, PPC or 680x0, and know that it would work.
The difference this time is that Apple's been planning ahead. NeXT made an Intel version of NeXTSTEP/OpenSTEP, so it cost them very little to develop OS X for Intel in parallel with OS X for PPC. Again, they're using a dynamic recompiler (Rosetta) so that PPC apps will run on Intel-based Macs. This time around, Apple is much better prepared to make this transition. The goal, as it was the last time, is to make it easy for Mac users to pick up a piece of off-the-shelf software for the Mac and know that it will run, regardless of the processor inside. On the developer boxes, as I understand it, many of Apple's own apps are still PPC-only and are running in Rosetta; iTunes is a prime example of this. Plus, the particular reason Apple's going with Intel right now is because they know that they can build more impressive computers -- right now.
What would be interesting, although a different study altogether, would be to assemble a group of women similar to the group of (presumably) men that write IQ tests now. Find yourself a couple of groups of men and women to take each test, and SUPRISE! the women will do better on the test written by women.
There have been numerous studies just like this that show that other kinds of standardized testing (of the "No Child Left Behind" variety) is biased against women and minorities, and that the scores women and minorities earn on these tests are not representative of their actual abilities.
What I've learned from all of this is that intelligence, or even knowledge in a particular area, is impossible to quantify. The problem is that individuals use different mental models to understand their world. If they don't use the same mental model as the test-writer, then they won't do as well.
Although, come to think of it, his PowerPoint slides were usually pretty sleep-inducing.
It's been a really long time since I've smoked. Although I think I did sleep better in those days. Hmm... wonder if I still know any dealers...
It wasn't trazodone, was it? I took that, and as I described in an earlier post, made me into a groggy basketcase the next day.
Plus, I invested in an expensive espresso machine. There's no way I can give up coffee altogether.
I used to take one -- very briefly -- that actually made me into an emotional basketcase the next day. Literally can't stop crying, can't explain why you're crying in the first place, plus usually a nasty headache. I used to work with a couple of my good friends, and they said it's the most bizarre they've ever seen me act -- and they've seen me drunk.
Oddly enough, my boss suggested having a beer before bed, too. But, my doctor says that's a bad idea. It makes your sleep more restless and you're more likely to wake up in the middle of the night.
Some of us are jealous of the relative ease with which the rest of you fall asleep. (The absolute worst is sharing a hotel room after a long trip, where your traveling companion falls asleep right away, but you don't fall asleep for hours) I'd be happy to at least feel as awake as most people seem. The only time I feel that way is when I can sleep in on the weekends. It's mostly just depressing that I can't be that alert the rest of the week -- you know, when it matters most.
And you can't point out that those organisms exist in an environment where they reproduce in numbers that would never be possible in nature. According to them -- well, one person in particular -- "it doesn't matter how loosely we hold the whip; it's still a form of slavery." That's more or less a direct quote.
And when you point out the symbiotic relationships that already exist in most living organisms -- animals in particular -- they refuse to believe that it's true. Something about the military-industrial complex of medicine trying to make us believe what they want us to believe or some crap like that.
It's hard not to laugh when someone tells me that humans are evil evil creatures for creating antibiotics and vaccines. Because someone think of the poor bacteria and viruses! I only wish I were joking right now.
And, they had what I consider to be one of the most sensible policies for monitoring children I've ever seen: for children under 13, a parent or other guardian *must* be present for the kid to be allowed to use the computers.
The rationale was, well, rational: as a parent, you know what you consider to be inappropriate for your child, so it's up to the parent to make the judgment call.
I haven't been back to Sioux City in quite some time, and I certainly haven't been in the library, so I'm not sure if they've changed their policies. What I really liked about their policies is that they accounted for the unspoken question of appropriateness: appropriate for whom? The courts are never going to be able to decide what little Johnny's parents think is appropriate for him, so let/force them to decide. That's why I think it's crazy that librarians are expected to parent other people's children for them while they're at the library.
It's the same principle, and they won't accept anything for a return unless it has a proof of purchase sticker they can verify with their system. Nice thing about that is that if you don't have the receipt, they can still verify it with the sticker and an ID.
It also used to work with telemarketers before the "do-not-call" list went into effect. They don't know you're really not a lawyer, but if you dress nicely and look (and/or sound) old enough, they'll believe anything.
Actually, you can ban your enemies, people you've fired, or those who complain about unfair practices unless there's specific legislation to stop you. Never seen protesters escorted out of a store, have you? Hell, until the US passed a friggin Constitutional amendment, it was legal to deny even the the most basic of public services to people on the basis of their race. (By the way, there are a lot of groups that aren't covered by that amendment. And even for the ones that are, it's not exactly a cakewalk.)
And your mythical bridge would be neither a toll bridge, nor privately owned if it's the only way across town. In fact, when have you ever seen privately owned roadways? Governments exist for the public good, not businesses. Businesses exist to make a profit. But everyone's got a profit motive, haven't they? Even (especially?) the Telus union. After all, they're holding out for more money. I doubt they give a shit about the public good if they're willing to undermine the company by threatening its employees.
BTW: Telus, you know, used to be a public entity. Just like BT. If you want to blame someone for their monopoly status, blame the Canadian government. Competition is good. So are bananas. Bananas are good.
Your analogy doesn't hold, because the Telus employees are using their employer's equipment to attack them. I'm having a hard time imagining how you're going to use your green hair and chlorinated water to attack the water department.
Here's a (perhaps) better analogy: if I own a hardware store, and I rent a sledge hammer to you (yeah, it's a weird hardware store... I know) and you use it to smash up the place and threaten my employees, don't you think, just maybe, it wouldn't be entirely unreasonable for me to ask for it back. I'll give your deposit back, even.
I guess what I want to know is this: why should a company invest its resources and money into hosting a site that calls for its demise? As an executive in that company, it would be wrong of them to leave it up there, for the safety and wellbeing of their other employees, as well as themselves.
I'll bet there are other people in that union who are embarrassed by this little happening, as it gives all of them a bad name, not just the ones who support such an action. In my experience, threatening the people who have what you want makes them all the more determined not to give it to you.
Hell, in the long run, it may be better for the union employees who posted it in the first place. If there are any inaccurate accusations intended to damage Telus's reputation, they could easily be sued for libel (if Canada's laws are anything like Britain's). I really don't understand why this is some big "censorship" argument. I see this as a reasonable action by a company protecting its own interests against those who would harm it.
Not necessarily. There have been plenty of studies that show that Apple owners buy computers less frequently than PC owners. Market share is a measure of relative volume of sales, not install base.
If, like my University, you buy new Macs every 3-4 years and new PCs every 2-3 years, in 10 years, you would have only bought 2-3 Macs, and somewhere between 3-5 PCs. And, we easily have three times as many PCs as Macs, so at the most extreme, in ten years, you would buy 2 Macs and 3 * 5 = 15 PC's. (And an education-oriented college campus is probably a lot more Mac-heavy than other institutions) 2/17 implies that 11.7% of our computers are Macs, but the actual ratio is 1/4, or 25%.
That's one of the main reasons I always caution people about using market share statistics to judge install base. It's simply not accurate, but it is the most readily-accessible tool.
I agree. If I were to imagine any video-based product coming down the pipeline, it would most likely take the form of a video analogue to the AirPort Express. I love my AP-E, and I think it would be fantastic if I could watch movies on it as well.
I don't think people really will watch movies on their iPods, if for no other reason than the situations in which you would listen to an iPod are generally times when you can't/shouldn't be watching a screen.