This is hooey. The GPL may or may not be based on socialist concepts, but the Linux kernel is definitely not; Linus bases all of his administration of the kernel's development solely on technical, non-political concerns. It is, in fact, based on people writing software that they want to use, and the only real reason most of them also then give it away is because it costs them relatively little to do so. (In particular, they can give it away and still have it.)
> Not only Linux, but public schools and hospitals
This is a common misconception, but in fact most hospitals in the US are firmly in the private sector. It's less obvious than in other industries (especially retail), partly because of the heavy regulation to which hospitals are subject, but hospitals do in fact participate in most if not all all of the usual Dilbert-style business activities (creative downsizing, cutting employee benefits and claiming it's an improvement, raising prices "to serve you better", and so forth) -- and they turn a profit, or try to.
There *are* socialist hospitals, mostly in the big cities, but the other hospitals all look down their noses at them and want to tell you they're not "real" hospitals and should be known as "clinics" or somesuch instead. The big university medical centers could also be considered socialist, and they get a little more respect, but still the other ("real") hospitals want very much to paint them as inferior. ("You get some *student* doctor...")
If you want to see socialized medicine in action, look at Canada or Sweden.
This is all not enough, if bots can sign up for accounts. You're going to need eventually to set things up so that only a verified human can get an account, and only a logged-in user can edit anything.
> Problem with approaches like these ("User Levels") is that some users > learn along the way
This is why you always have a way to change the current user level (one level at a time; if a newbie goes directly to expert mode by mistake he'll get all confused and have to reinstall). Generally, you want to tuck this away just enough so that the user on average has to get at least moderately comfortable with the current level to find it, without making it impossible to find for those who know most of the interface at the current level. Some settings may take effect on next reboot, and if so, which user level options are available should be one of them (to ensure that the user has the opportunity to see all the features of the current level before moving up any further).
Incidentally, my dream involves having applications be aware of the current user level and show more options at the higher levels. For example, at the End User level there should be very few preferences, but by PowerUser level all the available preferences should be shown in the prefs dialogs. This can apply to more than prefs dialogs, as well, so that e.g. a package manager might show only one or two best-of-breed applications in each category at End User level but at Power User level show descriptions of each option and allow the user to select a much less popular option (say, fvwm) if desired. Also, at the poweruser level, certain things (e.g., a shell prompt) should be prominent and easy to get to; whereas, at End User level you want them tucked away out of the way, because the user won't know what they are and will be afraid of them most likely. So that means you want things like your panel to be aware of user levels...
> Also, if memory serves me right, user testing has indicated that people are > afraid to choose "Dummies" level because they don't want to miss something > important.
That's a large part of why the "Dummies" level exists -- so they won't feel stupid choosing the End User level, which is still pretty basic.
> Too bad these are preInternet days, almost like prehistory.
You can find discussions in some of the older parts of usenet that refer to such things. There is quite a lot of history going back into the early days before the web. Lots of information is out there about earlier operating systems before Unix, for example.
> IT jobs require significant interaction from a Software Engineering standpoint. > Having your architects/sales/management on one side of the world and ppl doing > the "grunt" work on the other side can be very frustrating and impede progress.
A lot of this can be overcome using non-aural communication mechanims (e.g., email). That takes accents almost totally out of the picture. I can have a totally coherent conversation via email with e.g. a Korean; in person, I would have a *really* hard time understanding him (much worse than someone from India; Hindustani language have *more* consonants than English, so the accent is really not all that bad; Hangul is tonal and has as near as I can determine *no* consonants (well, no stops -- or very few at any rate), so the accent is really hard to follow for someone who grew up in the middle of the US; all you hear is vowells and liquids).
Sure, there are still barriers to perfect understanding, but that's always going to be the case when you have more than one person involved. Using written forms of communication does help a lot.
But you can't do that for telephone tech support, so it's good to see phone tech support coming back to the US, for US customers.
> End users, of course, will use whatever came preinstalled for them.
Speaking of that... my dream for an installer system goes like this. The first screen that pops up has several clickable links (reminiscent of a web page), along these lines...
How much do you want to customize your system?
1. Don't ask me any more questions. Just get me on the internet.
2. I don't mind a few simple choices, like what kind of tasks I want
to use the computer for, but don't get technical on me.
3. Show me some screenshots and let me choose my desktop setup, and put
a link to the control panel on the desktop so I can configure stuff.
4. I have specific ideas about exactly which packages I want to install
and how I want them configured. Show me a detailed package list.
(Traditional Linux Distro Installer)
5. I'm an expert. I grok the shell. Give me a working base system with
commandline package managers and I'll set up the rest myself.
Developers can refer to these conveniently as "For Dummies", "End User", "Moderate", "PowerUser" and "Expert" installs, respectively. I suspect the End User and Moderate installs would see the most use, followed by Dummies, with PowerUser and Expert trailing (but still important, because you want IT people to like your system well enough to deploy, recommend, and develop for it).
> - You may not like the lack of preferences, but corporations (and > inexperienced users) do. MSN is the default home page on millions of > systems because users are too lazy to change it. You may like a control > center with 40 pages (and multiple tabs per page), but such a thing flies > in the face of usability. Users are so overwhelmed by options that they > don't find the ones that really matter.
You're confusing end users with powerusers. End users do not want to change settings. At all. Ever. For them, you have defaults. Powerusers *love* changing settings. For them, you have preferences dialogs, control panels, and the like. Big tree-views with multiple panes containing multiple options, plus "Advanced" buttons in some of the panes that lead to additional dialoges? Yep, powerusers love that sort of thing. End users never see it, because as you rightly point out they never open up the prefs dialog in the first place, not even so much as to change the default home page. That's why the defaults should be aimed at them -- which is why, e.g., automatically prefixing www. and suffixing.com onto text typed into the browser's address bar makes a good default, even though it's loathsome in the extreme to more clueful users. The clueful users know where to find the prefs dialog and turn it the heck off.
That said, I've used both KDE and Gnome, and I'm not sure this is a major difference between them. I'd like to see more things like theme managers that theme both of their toolkits similarly, panel applets that can run in either panel, and so forth. Also, I'd like to see more ability to mix and match. For example, I specifically need to use the Gnome panel, because the drawers are an important feature for me. But Metacity blows chunks, and so I'm currently using sawfish, but I'd like to have the option of using the KDE window manager if such were my desire and have everything (e.g., the window entries on the task list in the Gnome panel) work together correctly. Similarly, someone else ought to have the option of using mostly KDE but replacing the filemanager with Nautilus. (Not that I can imagine why anyone would voluntarily use Nautilus; I've removed it from my Gnome session and do all my file management using gnome-terminal... but it ought to be an option.) In other words, the components of the desktop environments ought to be standardised (in terms of how they interact) and commoditized so that people (and distributions) can choose whichever parts they want. End users, of course, will use whatever came preinstalled for them.
> Linux is much smaller than Windows XP when you take everything into account > properly.
Only if you spend an hour selecting and unselecting a bunch of individual packages. If you pick the "Just Install Everything" option and go make yourself some dinner while it copies all the files, most major distros are a tad bit larger. When was the last time you saw a version of Windows that came on three CDs, all nearly full?
Granted, most of that is redundant, and most of what's not redundant (e.g., Gimp, Perl,...) is stuff that Windows does not come with out of the box. Really we're comparing apples with orange groves. When you count up the size of a Windows install, do you count all the festering ooze that a typical big-name OEM installs on there for no reason? When you count up the size of a Linux install, do you count all the redundancies in a typical distribution (e.g., thirty different window managers)? I mean, it's not really fair *not* to count them, because part of the *value* of Linux is the customizeability, the abundance of choice. (I should know; I hated Metacity and so replaced it with sawfish.) On the other hand, it's not fair to count them either, because it's possible to download multiple everything for Windows too, but nobody does, generally; they just download or buy the one they want. e.g., (relatively) a lot of Windows users (here we're talking about powerusers, of course) go download Netscape or Mozilla, but relatively few download both plus Opera, K-Meleon, CrazyBrowser, and four or five others. So the choice is still there, but each and every thing that you want you have to go and separately obtain; in Linux you get most of what you want OOTB, but the price you pay is that the install is larger, because you also get a lot of stuff you don't care about. It's a tradeoff between convenience and disk space, IMO.
Of course, you could always get Gentoo and individually emerge each and every package you want to use... but that is not the path for most users.
G500? What's that? I use a Matrox Mystique, for several reasons: It came with the PC I bought in 1998, so I have it. It's the best graphics card I've ever used, in terms of never manifesting a display quirk, ever. It works OOTB with every operating system that supports PCI at all. Oh, and there's no compelling reason to switch, because it does everything I need a graphics card to do.
Actually, that last point isn't entirely accurate: I *want* my graphics card to do automatic alpha blending of layers, so that my GUI can just write each window or whatever as an RGBA layer and allow the hardware to composite them. If anybody knows of such a card with reasonably broad OS support, let me know.
3D doesn't interest me, however. The screen is, like, you know, _flat_, so a 3D interface comes out as an ugly and impractical kludge. If I had VR goggles, I might feel differently about that.
> > Cable select is a major pain > Only if you've got a frankenputer.. > Try working on some Tier-1 machines
By Tier-1, I guess you mean big-name, like Compaq and Dell? I do work with them, at work. I hate them. (Well, Dell's cases are good, but I'm not buying a whole computer just for the case, when I'm going to have to replace most of the components.) Whitebox systems have, generally, much more reliable components and are also more likely to work with any random operating system. Further, with the big-name systems you don't usually get proper driver disks, just an All-Or-Nothing "Restore CD", which is worthless if you don't want to delete all your data and a big pain even if you don't need your data. Ugh. (Dell is the exception here again; they actually send you a Drivers & Utilities CD. I'm starting to like them, for situations where I absolutely have to pick a big name.)
Additionally, big-name computers invariably come with one CD drive and one hard drive; that's nowhere near enough. Usually they only have two IDE channels, so you can only have four drives total, and that's absolutely not enough. (I'm a multibooter; even with only one CD drive, that would be only 3*4=12 primary partitions, which is inadequate.)
Cable select is a major pain, because it limits where in the case you can put each drive. If e.g. you want to put your ATAPI CD-ROM drive in the bottom bay for reasons that have to do with the physical positioning of your case, and your hard drive needs one of those full-bay cooler/mounting kit combos, Cable Select makes it virtually impossible to get the cable to reach everything, assuming you don't want your CD-ROM drive to be the primary master. It's much easier to jumper the drives and plug the cable in the way that it can reach most easily.
That said, Cable Select does have its usefulness. Sometimes you want to be able to switch the order of the drives without rejumpering them, especially when you're moving a drive back and forth between two systems. But until SATA takes over entirely you still need the option of jumpering your drives for Master and Slave.
The author of the memo is clearly an ignoramus unqualified to discuss the subject. To wit:
> One such recent example included the manufacturer's labeling of equipment > where the words "Master/Slave" appeared to identify the primary and > secondary sources.
Err, no. Master and Slave do *not* refer to Primary and Secondary. They refer to Master and Slave. The Primary and Secondary channels can each have both a Master and a Slave, as anyone who has even the smallest amount of knowledge of the subject would not only know, but take for granted. The author of the memo is entirely out of his depth and should refrain from sending memos of any kind to suppliers of technical equipment, lest he inadvertently order something he didn't intend that will cost the county a lot of money and get him fired. He probably thinks RAID terminology is insensitive to Scandinavians because of its resemblance to Viking activities.
> Is it possible to have a clean organized grown-up home, without throwing > everything away?"
Yes, mostly. Designate one area of your home as the "work area" or "mess zone". A basement room is a good choice. Outfit it with shelves on every wall and a table, some benches, a desk, that sort of thing (i.e., lots of horizontal surface area). Here you can have computer components sitting around, boxes and bins of different kinds of cables, stacks of manuals and papers and books, and so forth. A three-drawer cabinet or two (like a filing cabinet but with high-sided drawers made for holding things besides papers) is a good kind of furniture for this room, too. Also, you want a couple of easily-repositioned lamps, for shedding light in exactly the right part of a computer case, or whatever.
Keep all your junky stuff in this room, and out of the rest of the house. When (not if) the room starts to get awefully cluttered so that it becomes difficult to navigate to the opposite side of the room, box up some of the stuff you use least often. If there's no place to put the box, cart it to the attic for longer-term storage. (Striking a balance between keeping a high enough percentage of your stuff boxed up that you have room for most of it in the room and yet keeping enough stuff layed out that you can reach the things you need constantly does take some thought, but it's not an impossible balance to strike, generally.)
Now, the rest of the house can have some computer equipment, but only in the form of complete, working systems with covers on the cases that have some hope of being used in any given 24-hour period. When you need to work on a computer, or when you stop using it regularly, you bring it to the work room. If you have family members who don't like mess, you make a bargain with them: they leave your work room alone, and you do your part to keep the rest of the house picked up. (Dirty clothes in the basket, and all that sort of nonsense.)
Things that may happen in 2004...
on
Linux in 2004?
·
· Score: 1
* Perl6 won't come out, but several Apocalypse and Exegesis articles will, and
we'll get all excited about waiting with baited breath for Perl6 to come out
some future year. * The 2.6 series Linux kernels will have a number of important patches. * Sun will continue to hype Java, and Microsoft will continue to hype.NET,
but few people will adopt either of them that haven't done so already. * gcc 2.x will begin to see serious disuse as 3.x pretty much takes over. * People will start talking about features that will be in Emacs 22. * Some of the features being developed at xserver.freedesktop.org will make
it into a major distro but will be optional and probably not the default. * A major ISP will threaten to adopt Mozilla technology, forcing Microsoft
to strongly reiterate their promise to update MSIE to block popups. * A major desktop vendor will ship low-end systems aimed at consumers with
an office suite built on OpenOffice.org technology. * A major desktop vendor will announce a total discontinuance of systems
with "legacy" ports, saying that they're going over entirely to USB, but
the "legacy-free" systems will continue to flop in the marketplace and the
vendor will retract their position and continue to ship systems with
RS232, parallel, and PS/2 ports, on the grounds that they still want to
turn a profit. * A "luggable" laptop will be released with an 18" viewable screen.
It will have one hardware component (probably a NIC or modem) that
doesn't support Linux initially. * Macromedia will release a Linux version of one of their programs that
was previously unavailable for Linux. Nobody much will care, but slashdot
will get at least three stories out of it. * Apple will announce Mac OS X 10.4, with redesigned box art, changes to
Finder, synced with a newer version of BSD, a newer Apache, and assorted
minor improvements to the GUI. People will complain about having to pay
for the upgrade, but it'll get great reviews. My best guess for the
codename is "Wildcat". People will switch to it, mostly from earlier
versions of Mac OS X, but some from other systems as well. * The SCO thing will not be over by the end of 2004, but we *will* be
tired of hearing about it.
> Apparently, the only way to avoid a lawsuit entirely is to cover songs > from the public domain.
Yeah, but this is really easy to do. Pull a short theme (about twelve notes) from anything composed prior to the advent of modern copyright law, repeat the first part of it three times before adding the last part, and you've got your melody for the chorus. Pick another theme and do the same thing to it for your verse melody. Grab any piece of love poetry written during the same period, retranslate the poetry into modern English with an abcb rhyming scheme, completely ignoring meter, resyncopate the melody to fit the lyrics, add a couple of lame harmony parts to complete the monody, throw in a musical bridge between the fifth and sixth repetition of the chorus, and voila, you've got crap. Get a pop singer to title an album after it, and it'll sell like hotcakes. Anybody tries to sue you for infringement, you've got ironclad defense: "Err, the chorus melody is from Canon in D, and the verse melody is from the Musical Offering. The lyrics are straight out of Voltaire, translated into English. What did we take from your music, exactly?"
> I know some people who really don't like Beethoven too much since the > orchestral part they play in his pieces is utterly boring (a commonly > seen thing from the classical period forward for supporting parts).
Quite. Probably the most innovative thing about Beethoven is his use of widely varying volume. Frankly, I'm not sure he would be as popular as he is (not that he wouldn't be popular at all, but that a dozen other composers might ecclipse him for the top most-famous-composer-of-all-time spot) if it weren't for his having gone deaf in the late part of his career.
And yes, I find the supporting parts very boring in most modern music, from about the Romantic period forward. It's all monody, and monody by 1750 or so had been done and done and done, done to death, then done some more. I much prefer non-monodic forms (true polyphony, canon, fugue,...). They're harder to write well, granted, but they're also much more interesting when they are written well. The interaction between the parts in a good fugue is absolutely *fascinating*. You can lose yourself in that kind of music for hours.
The only downside is, its complexity makes it hard to hum; you never know which part you should be humming, so it's not as *catchy* as monody, because you hum a few bars and get lost, switch parts, hum a few more bars, get lost again, lose track of where you are, and either quit or start over. But IMO this elusive quality also makes such music more intriguing.
I'd be interested in seeing what a high-quality contemporary group or composer would do if they set out to create a modern polyphonic form. I can imagine Stomp creating a percussive fugue, or Mannheim Steamroller putting together something polyphonic with nontraditional instruments, or John Williams weaving together two disparate but equal themes into a movie soundtrack (say, a theme for the good guys and one for the bad guys, both going at once whenever both are onscreen). Unfortunately, as near as I can tell, nobody seems to be composing anything these days that's not monodic, unless you count the fugue portion of Glad's "History of A Capella Music" bonus cut.
> Other sites that utilize moderation in a Slashdot style fashion? > If so, name an example. www.kuro5hin.org doesn't allow post editing.
www.perlmonks.org now allows you to edit your own root nodes. (You could always edit your own replies, and root nodes in certain sections (e.g., in Obfuscations), but now you can edit all of them.)
Granted the moderation is a bit different over there, as your karma (err, XP) depends directly on the voting done by lots of people (not just a handful of mods), but your post first has to be "approved" by one scribe (or higher), else it will languish in the realm of only being seen by people who have turned on unapproved nodes in their prefs. Also, particularly worthless nodes can be Considered, and with enough downvotes can be Reaped, which causes them to not be shown inline, but a "Reaped: Reason" with a link to the actual contents. Oh, and for those special cases where that's not enough, a small handful of people have the power to nuke a node altogether, though this is almost never done.
So, basically, whereas the slashdot mod system is oversimple, the perlmonks one is baroque -- but based on the same basic idea.
> learning a video game is a recreational process.
Learning anything computer-related is a recreational process for some and painful for others, mostly depending on your attitude toward computers. My mom would consider learning a video game interface just as painful as any other computer-related thing, and I would consider it no more recreational than learning a new programming language (assuming it's a fun language to learn, by which I mean a dynamically typed and multiparadigmatic language).
> I don't know why you are having issues with winXP.
In fairness, most of my problems with WinXP have been directly related to identifying the model of the video card so as to locate the correct driver on the manufacturer's website. After that, you install it and reboot a couple of times, and then it will generally let you change the video settings to something reasonable. Still, the whole process takes a few minutes, so the other poster's "it just works" charactarization is a bit on the optimistic side, IMO.
> But really, how hard is it to get drivers for 2 very well supported chipsets > and throw them into an OS?
Not too bad, once you identify the chipsets. (It's not always obvious which chip to read the numbers off of...) But letting HardDrake do this stuff for you is easier, and I've yet to see it fail to detect a video card. (Some other kinds of hardware are another matter, granted.)
> IE, not some freaky looking acid-trip OEM box from taiwan that you can't > find the manufacturer name on.
Actually, my *worst* problems have been with big-name systems, where I had to reinstall Windows for one reason or another, and the specific model name is nowhere on the system (it just says, e.g., "Deskpro", which is roughly as specific as identifying a nursing home resident as "the one with glasses"), so you don't know which softpaq to get. But then, too, my worst problems have been not with video cards but with NICs. In two cases I ultimately gave up and bought replacement NE2000 cards (and later put the old cards into Mandrake systems, where HardDrake configured them automatically, and all I had to do was tell it what IP settings to use).
Mind you, I'm not saying that hardware support is universally better in Linux. (It's not. Some hardware isn't supported at all.) What I'm saying is that it's not universally "it just works" in Windows, as the other post seemed to be claiming. That's a significant exaggeration.
> The console commands are confusing to those without experience with them
Well, duh. Dude, it's the eighties. Err, nineties. No, wait... well, anyway, it's not the seventies anymore, and all the end users are using this new-fangled underpowered type of interface with graphics, where they point at things with a mouse device and, you know, click.
The commandline is for powerusers. And, for powerusers, the commandline in Windows (current generation; bring this up again in 2006 and we'll talk) is pretty vastly inferior to the 1975 state of the art in Unix command shells. When I'm on Windows, I find myself having to use a GUI file manager (Explorer or some equivalent) for certain things, even though it's clunky, because doing them from the command line, as I would normally do on any POSIX system, is difficult or impossible. (Yeah, yeah, cygwin, but if it was a system I was going to use on an ongoing basis I'd just install Mandrake. We're talking here about doing maintenance on systems that are mostly used by other people.)
> From someone without computer experience, what the hell is "mounting" a > hard drive?
Something you don't have to worry about. The installer for any of the major newbie-friendly distros (RedHat, Mandrake, and so forth -- as opposed to, say, Gentoo, which is intended for people with some *nix experience) takes care of setting up fstab automatically. If your distro includes Supermount (as e.g. Mandrake does) you don't have to manually mount your floppies or CDs either. (I always turn this off, but I'm a poweruser.)
It sounds to me like you've either been using a quite old version, or a distro that wasn't intended for newbies.
> Where do you put your pictures? Not in "C:\pictures", but in > "/mnt/users/username/home/pictuers" or somesuch.
You click on the little picture of a house on your panel, and a window opens up (that's your "home folder" inside your computer, where you keep all your stuff; the equivalent on WinXP is "My Documents") and then click on the "Pictures" folder. That's where you keep your pictures.
> Hey, how about "How does someone with no Linux experience install things?"
Under the foot menu, click on "Configuration" and then "Install Software". The Software Manager (technically known in geek circles as rpmdrake, but you don't need to know that) will come up. Checkmark the packages you want to install, and click the "Install" button.
What, you mean *download and install*? End users don't do that, under any OS, generally. They get someone like me to do it for them, or (you're going to think I'm kidding here, but I've seen 'em do this for office software) they go buy a computer that has the software in question installed on it. (*That* can get expensive...) They're either afraid to install software (because it might mess something up or possibly give them a virus) or else they just plain don't know how. If you're comfortable installing software, I'd classify you as at least somewhat a poweruser (though, of course, there are a range of abilities among powerusers; I'm more toward the high end of it myself, pushing the boundary between poweruser and developer; I don't actually write software that anyone uses besides myself, unless you count CGI-based database frontends, but I know and use Emacs lisp, understand algorithm analysis, and can think in Perl... so I'm *almost* a developer.)
Just having the *desire* to install software that didn't come preinstalled on your computer betrays a poweruser tendency in your mindset: you want to make the computer do more, add capabilities to its feature set. End users don't think that way; they don't even want to learn to use the features they've already got, except for the smallest subset they need to get their stuff done. (Learn ways to get it done better, faster, and easier? No thanks, they haven't got time for that. They're busy, after all, and they have other things to do than sit in front of a computer all the time.)
> It is really a shame that someone like Canon does not create some linux > software and include it with their cameras.
This is silly. Doing that, you have to deal with different configurations on the user's PC, and then when you go to write Windows and Mac interfaces you have to start over from scratch.
Much better to just embed a web server in the camera/scanner/whatever. This is no harder than creating interface software for any one platform, and it gets you all platforms, and as an added bonus, it gets you a user interface that nearly everybody already understands. Given the price of NE2000 network cards these days, you can bundle one and a short crossover cable for roughly the same cost as the packaging, just in case the user doesn't have a LAN setup already. Embed a DHCP server on the camera and Bob is everybody's uncle.
> I can throw 2 video cards into windows 2000 and they just.... work.
Wow. That's awesome. Maybe I should recommend a Win2k site license at work, so we can migrate all our Win98 and WinXP systems to it. Because in Win98 and WinXP, you're lucky if *one* video card works when you plug it in; usually, after each Windows reinstall you have to spend fifteen minutes or so (if you've done it a lot of times before -- it can take a newbie an hour) getting the graphics card to work in anything but 16-color mode at the minimum res and 60Hz refresh rate. That's if you know the manufacturer and model number of your graphics card (or motherboard for onboard video); if you don't, you're pretty much totally screwed.
> As much as I hate spam, it shouldn't be a criminal offense, and especially > should not have a prison sentence.
This bill does not make spamming illegal (it specifically makes it legal, and regulates it), and the only things it attaches prison sentences to are the online equivalents of crimes that in the real world can already get you jail time -- things like fraud.
Almost everyone on/. is concentrating on what this bill *doesn't* do (it doesn't outright make spam illegal), but what it *does* do (within the US) is make some of the more heinous abuses (e.g., forging headers) illegal. That's a good thing. I said months ago, if we can just make forging headers illegal worldwide, just that alone will make spam twenty times easier to filter. This bill does that in the US, which is the first step. Does it stop spam? No. Can it help make it more manageable? Yeah, it can. Just the provision against forging headers is worth passing it for. The spammers will of course move overseas, but the same was true of copyright-spurning printing houses[1], at one time -- until most of the world adopted international copyright law. But some nation has to lead the way.
[1] I'm talking about the businesses based on printing many copies of a
copyrighted work without permission and selling them in bulk at a lower
price than the legitimate publisher, cutting the author out entirely.
Yes, it's easier to stop such books from being imported than it is to
stop spam at the borders, but it remains that most of the world now has
adopted international copyright law -- pretty much every country with
phone lines. If we get that level of cooperation for stopping the
forging of email headers, we can put away the bayesian classification
and markov chain analysis and go back to simple domain-based blacklists.
> I would expect, all factors being equal, to receive some spam in Hindi...
Now that I think about it, I wonder if all the major ISPs in India use the local two-digit TLD, which would have the side-effect of making it easy for spammers using those languages to target their spam toward people who can maybe read it.
Slashdot Mirror
> Linux is based on socialist concepts.
This is hooey. The GPL may or may not be based on socialist concepts, but
the Linux kernel is definitely not; Linus bases all of his administration of
the kernel's development solely on technical, non-political concerns. It is,
in fact, based on people writing software that they want to use, and the only
real reason most of them also then give it away is because it costs them
relatively little to do so. (In particular, they can give it away and still
have it.)
> Not only Linux, but public schools and hospitals
This is a common misconception, but in fact most hospitals in the US are firmly
in the private sector. It's less obvious than in other industries (especially
retail), partly because of the heavy regulation to which hospitals are subject,
but hospitals do in fact participate in most if not all all of the usual
Dilbert-style business activities (creative downsizing, cutting employee
benefits and claiming it's an improvement, raising prices "to serve you
better", and so forth) -- and they turn a profit, or try to.
There *are* socialist hospitals, mostly in the big cities, but the other
hospitals all look down their noses at them and want to tell you they're not
"real" hospitals and should be known as "clinics" or somesuch instead. The
big university medical centers could also be considered socialist, and they
get a little more respect, but still the other ("real") hospitals want very
much to paint them as inferior. ("You get some *student* doctor...")
If you want to see socialized medicine in action, look at Canada or Sweden.
This is all not enough, if bots can sign up for accounts. You're going to need
eventually to set things up so that only a verified human can get an account,
and only a logged-in user can edit anything.
> Problem with approaches like these ("User Levels") is that some users
> learn along the way
This is why you always have a way to change the current user level (one level
at a time; if a newbie goes directly to expert mode by mistake he'll get
all confused and have to reinstall). Generally, you want to tuck this away
just enough so that the user on average has to get at least moderately
comfortable with the current level to find it, without making it impossible
to find for those who know most of the interface at the current level.
Some settings may take effect on next reboot, and if so, which user level
options are available should be one of them (to ensure that the user has
the opportunity to see all the features of the current level before moving
up any further).
Incidentally, my dream involves having applications be aware of the current
user level and show more options at the higher levels. For example, at the
End User level there should be very few preferences, but by PowerUser level
all the available preferences should be shown in the prefs dialogs. This
can apply to more than prefs dialogs, as well, so that e.g. a package
manager might show only one or two best-of-breed applications in each
category at End User level but at Power User level show descriptions of
each option and allow the user to select a much less popular option (say,
fvwm) if desired. Also, at the poweruser level, certain things (e.g., a
shell prompt) should be prominent and easy to get to; whereas, at End User
level you want them tucked away out of the way, because the user won't
know what they are and will be afraid of them most likely. So that means
you want things like your panel to be aware of user levels...
> Also, if memory serves me right, user testing has indicated that people are
> afraid to choose "Dummies" level because they don't want to miss something
> important.
That's a large part of why the "Dummies" level exists -- so they won't feel
stupid choosing the End User level, which is still pretty basic.
> Too bad these are preInternet days, almost like prehistory.
You can find discussions in some of the older parts of usenet that refer
to such things. There is quite a lot of history going back into the early
days before the web. Lots of information is out there about earlier operating
systems before Unix, for example.
> IT jobs require significant interaction from a Software Engineering standpoint.
> Having your architects/sales/management on one side of the world and ppl doing
> the "grunt" work on the other side can be very frustrating and impede progress.
A lot of this can be overcome using non-aural communication mechanims (e.g.,
email). That takes accents almost totally out of the picture. I can have
a totally coherent conversation via email with e.g. a Korean; in person, I
would have a *really* hard time understanding him (much worse than someone
from India; Hindustani language have *more* consonants than English, so the
accent is really not all that bad; Hangul is tonal and has as near as I can
determine *no* consonants (well, no stops -- or very few at any rate), so the
accent is really hard to follow for someone who grew up in the middle of the
US; all you hear is vowells and liquids).
Sure, there are still barriers to perfect understanding, but that's always
going to be the case when you have more than one person involved. Using
written forms of communication does help a lot.
But you can't do that for telephone tech support, so it's good to see phone
tech support coming back to the US, for US customers.
> End users, of course, will use whatever came preinstalled for them.
Speaking of that... my dream for an installer system goes like this. The
first screen that pops up has several clickable links (reminiscent of a web
page), along these lines...
How much do you want to customize your system?
1. Don't ask me any more questions. Just get me on the internet.
2. I don't mind a few simple choices, like what kind of tasks I want
to use the computer for, but don't get technical on me.
3. Show me some screenshots and let me choose my desktop setup, and put
a link to the control panel on the desktop so I can configure stuff.
4. I have specific ideas about exactly which packages I want to install
and how I want them configured. Show me a detailed package list.
(Traditional Linux Distro Installer)
5. I'm an expert. I grok the shell. Give me a working base system with
commandline package managers and I'll set up the rest myself.
Developers can refer to these conveniently as "For Dummies", "End User",
"Moderate", "PowerUser" and "Expert" installs, respectively. I suspect the
End User and Moderate installs would see the most use, followed by Dummies,
with PowerUser and Expert trailing (but still important, because you want
IT people to like your system well enough to deploy, recommend, and develop
for it).
> - You may not like the lack of preferences, but corporations (and
.com onto text typed into the browser's address
> inexperienced users) do. MSN is the default home page on millions of
> systems because users are too lazy to change it. You may like a control
> center with 40 pages (and multiple tabs per page), but such a thing flies
> in the face of usability. Users are so overwhelmed by options that they
> don't find the ones that really matter.
You're confusing end users with powerusers. End users do not want to change
settings. At all. Ever. For them, you have defaults. Powerusers *love*
changing settings. For them, you have preferences dialogs, control panels,
and the like. Big tree-views with multiple panes containing multiple options,
plus "Advanced" buttons in some of the panes that lead to additional dialoges?
Yep, powerusers love that sort of thing. End users never see it, because
as you rightly point out they never open up the prefs dialog in the first
place, not even so much as to change the default home page. That's why
the defaults should be aimed at them -- which is why, e.g., automatically
prefixing www. and suffixing
bar makes a good default, even though it's loathsome in the extreme to more
clueful users. The clueful users know where to find the prefs dialog and
turn it the heck off.
That said, I've used both KDE and Gnome, and I'm not sure this is a major
difference between them. I'd like to see more things like theme managers
that theme both of their toolkits similarly, panel applets that can run in
either panel, and so forth. Also, I'd like to see more ability to mix and
match. For example, I specifically need to use the Gnome panel, because the
drawers are an important feature for me. But Metacity blows chunks, and so
I'm currently using sawfish, but I'd like to have the option of using the
KDE window manager if such were my desire and have everything (e.g., the
window entries on the task list in the Gnome panel) work together correctly.
Similarly, someone else ought to have the option of using mostly KDE but
replacing the filemanager with Nautilus. (Not that I can imagine why anyone
would voluntarily use Nautilus; I've removed it from my Gnome session and do
all my file management using gnome-terminal... but it ought to be an option.)
In other words, the components of the desktop environments ought to be
standardised (in terms of how they interact) and commoditized so that people
(and distributions) can choose whichever parts they want. End users, of
course, will use whatever came preinstalled for them.
> Linux is much smaller than Windows XP when you take everything into account
...) is stuff that Windows does not come with out of the box.
> properly.
Only if you spend an hour selecting and unselecting a bunch of individual
packages. If you pick the "Just Install Everything" option and go make
yourself some dinner while it copies all the files, most major distros are
a tad bit larger. When was the last time you saw a version of Windows that
came on three CDs, all nearly full?
Granted, most of that is redundant, and most of what's not redundant (e.g.,
Gimp, Perl,
Really we're comparing apples with orange groves. When you count up the
size of a Windows install, do you count all the festering ooze that a
typical big-name OEM installs on there for no reason? When you count up the
size of a Linux install, do you count all the redundancies in a typical
distribution (e.g., thirty different window managers)? I mean, it's not
really fair *not* to count them, because part of the *value* of Linux is
the customizeability, the abundance of choice. (I should know; I hated
Metacity and so replaced it with sawfish.) On the other hand, it's not
fair to count them either, because it's possible to download multiple
everything for Windows too, but nobody does, generally; they just download
or buy the one they want. e.g., (relatively) a lot of Windows users (here
we're talking about powerusers, of course) go download Netscape or Mozilla,
but relatively few download both plus Opera, K-Meleon, CrazyBrowser, and
four or five others. So the choice is still there, but each and every
thing that you want you have to go and separately obtain; in Linux you
get most of what you want OOTB, but the price you pay is that the install
is larger, because you also get a lot of stuff you don't care about. It's
a tradeoff between convenience and disk space, IMO.
Of course, you could always get Gentoo and individually emerge each and
every package you want to use... but that is not the path for most users.
> Ami Pro (which invented the icon toolbar, as far as I know!)
I very much doubt it. The concept of a toolbar with icons is certifiably
ancient, certainly much older than Macintosh.
G500? What's that? I use a Matrox Mystique, for several reasons: It came
with the PC I bought in 1998, so I have it. It's the best graphics card
I've ever used, in terms of never manifesting a display quirk, ever. It
works OOTB with every operating system that supports PCI at all. Oh, and
there's no compelling reason to switch, because it does everything I need
a graphics card to do.
Actually, that last point isn't entirely accurate: I *want* my graphics card
to do automatic alpha blending of layers, so that my GUI can just write each
window or whatever as an RGBA layer and allow the hardware to composite them.
If anybody knows of such a card with reasonably broad OS support, let me know.
3D doesn't interest me, however. The screen is, like, you know, _flat_, so a
3D interface comes out as an ugly and impractical kludge. If I had VR goggles,
I might feel differently about that.
> > Cable select is a major pain
> Only if you've got a frankenputer..
> Try working on some Tier-1 machines
By Tier-1, I guess you mean big-name, like Compaq and Dell? I do work with
them, at work. I hate them. (Well, Dell's cases are good, but I'm not buying
a whole computer just for the case, when I'm going to have to replace most of
the components.) Whitebox systems have, generally, much more reliable
components and are also more likely to work with any random operating system.
Further, with the big-name systems you don't usually get proper driver disks,
just an All-Or-Nothing "Restore CD", which is worthless if you don't want to
delete all your data and a big pain even if you don't need your data. Ugh.
(Dell is the exception here again; they actually send you a Drivers & Utilities
CD. I'm starting to like them, for situations where I absolutely have to pick
a big name.)
Additionally, big-name computers invariably come with one CD drive and one
hard drive; that's nowhere near enough. Usually they only have two IDE
channels, so you can only have four drives total, and that's absolutely not
enough. (I'm a multibooter; even with only one CD drive, that would be only
3*4=12 primary partitions, which is inadequate.)
> What about cable select?
Cable select is a major pain, because it limits where in the case you can put
each drive. If e.g. you want to put your ATAPI CD-ROM drive in the bottom bay
for reasons that have to do with the physical positioning of your case, and
your hard drive needs one of those full-bay cooler/mounting kit combos, Cable
Select makes it virtually impossible to get the cable to reach everything,
assuming you don't want your CD-ROM drive to be the primary master. It's much
easier to jumper the drives and plug the cable in the way that it can reach
most easily.
That said, Cable Select does have its usefulness. Sometimes you want to be
able to switch the order of the drives without rejumpering them, especially
when you're moving a drive back and forth between two systems. But until
SATA takes over entirely you still need the option of jumpering your drives
for Master and Slave.
The author of the memo is clearly an ignoramus unqualified to discuss
the subject. To wit:
> One such recent example included the manufacturer's labeling of equipment
> where the words "Master/Slave" appeared to identify the primary and
> secondary sources.
Err, no. Master and Slave do *not* refer to Primary and Secondary. They
refer to Master and Slave. The Primary and Secondary channels can each
have both a Master and a Slave, as anyone who has even the smallest amount
of knowledge of the subject would not only know, but take for granted.
The author of the memo is entirely out of his depth and should refrain
from sending memos of any kind to suppliers of technical equipment, lest
he inadvertently order something he didn't intend that will cost the county
a lot of money and get him fired. He probably thinks RAID terminology is
insensitive to Scandinavians because of its resemblance to Viking activities.
> Is it possible to have a clean organized grown-up home, without throwing
> everything away?"
Yes, mostly. Designate one area of your home as the "work area" or "mess
zone". A basement room is a good choice. Outfit it with shelves on every
wall and a table, some benches, a desk, that sort of thing (i.e., lots of
horizontal surface area). Here you can have computer components sitting
around, boxes and bins of different kinds of cables, stacks of manuals and
papers and books, and so forth. A three-drawer cabinet or two (like a filing
cabinet but with high-sided drawers made for holding things besides papers)
is a good kind of furniture for this room, too. Also, you want a couple of
easily-repositioned lamps, for shedding light in exactly the right part of
a computer case, or whatever.
Keep all your junky stuff in this room, and out of the rest of the house.
When (not if) the room starts to get awefully cluttered so that it becomes
difficult to navigate to the opposite side of the room, box up some of the
stuff you use least often. If there's no place to put the box, cart it to
the attic for longer-term storage. (Striking a balance between keeping
a high enough percentage of your stuff boxed up that you have room for
most of it in the room and yet keeping enough stuff layed out that you can
reach the things you need constantly does take some thought, but it's not
an impossible balance to strike, generally.)
Now, the rest of the house can have some computer equipment, but only in the
form of complete, working systems with covers on the cases that have some
hope of being used in any given 24-hour period. When you need to work on
a computer, or when you stop using it regularly, you bring it to the work
room. If you have family members who don't like mess, you make a bargain
with them: they leave your work room alone, and you do your part to keep
the rest of the house picked up. (Dirty clothes in the basket, and all
that sort of nonsense.)
* Perl6 won't come out, but several Apocalypse and Exegesis articles will, and .NET,
we'll get all excited about waiting with baited breath for Perl6 to come out
some future year.
* The 2.6 series Linux kernels will have a number of important patches.
* Sun will continue to hype Java, and Microsoft will continue to hype
but few people will adopt either of them that haven't done so already.
* gcc 2.x will begin to see serious disuse as 3.x pretty much takes over.
* People will start talking about features that will be in Emacs 22.
* Some of the features being developed at xserver.freedesktop.org will make
it into a major distro but will be optional and probably not the default.
* A major ISP will threaten to adopt Mozilla technology, forcing Microsoft
to strongly reiterate their promise to update MSIE to block popups.
* A major desktop vendor will ship low-end systems aimed at consumers with
an office suite built on OpenOffice.org technology.
* A major desktop vendor will announce a total discontinuance of systems
with "legacy" ports, saying that they're going over entirely to USB, but
the "legacy-free" systems will continue to flop in the marketplace and the
vendor will retract their position and continue to ship systems with
RS232, parallel, and PS/2 ports, on the grounds that they still want to
turn a profit.
* A "luggable" laptop will be released with an 18" viewable screen.
It will have one hardware component (probably a NIC or modem) that
doesn't support Linux initially.
* Macromedia will release a Linux version of one of their programs that
was previously unavailable for Linux. Nobody much will care, but slashdot
will get at least three stories out of it.
* Apple will announce Mac OS X 10.4, with redesigned box art, changes to
Finder, synced with a newer version of BSD, a newer Apache, and assorted
minor improvements to the GUI. People will complain about having to pay
for the upgrade, but it'll get great reviews. My best guess for the
codename is "Wildcat". People will switch to it, mostly from earlier
versions of Mac OS X, but some from other systems as well.
* The SCO thing will not be over by the end of 2004, but we *will* be
tired of hearing about it.
> Apparently, the only way to avoid a lawsuit entirely is to cover songs
> from the public domain.
Yeah, but this is really easy to do. Pull a short theme (about twelve notes)
from anything composed prior to the advent of modern copyright law, repeat
the first part of it three times before adding the last part, and you've got
your melody for the chorus. Pick another theme and do the same thing to it
for your verse melody. Grab any piece of love poetry written during the same
period, retranslate the poetry into modern English with an abcb rhyming scheme,
completely ignoring meter, resyncopate the melody to fit the lyrics, add a
couple of lame harmony parts to complete the monody, throw in a musical bridge
between the fifth and sixth repetition of the chorus, and voila, you've got
crap. Get a pop singer to title an album after it, and it'll sell like
hotcakes. Anybody tries to sue you for infringement, you've got ironclad
defense: "Err, the chorus melody is from Canon in D, and the verse melody
is from the Musical Offering. The lyrics are straight out of Voltaire,
translated into English. What did we take from your music, exactly?"
> I know some people who really don't like Beethoven too much since the
...). They're
> orchestral part they play in his pieces is utterly boring (a commonly
> seen thing from the classical period forward for supporting parts).
Quite. Probably the most innovative thing about Beethoven is his use of
widely varying volume. Frankly, I'm not sure he would be as popular as
he is (not that he wouldn't be popular at all, but that a dozen other
composers might ecclipse him for the top most-famous-composer-of-all-time
spot) if it weren't for his having gone deaf in the late part of his career.
And yes, I find the supporting parts very boring in most modern music, from
about the Romantic period forward. It's all monody, and monody by 1750 or
so had been done and done and done, done to death, then done some more. I
much prefer non-monodic forms (true polyphony, canon, fugue,
harder to write well, granted, but they're also much more interesting when
they are written well. The interaction between the parts in a good fugue
is absolutely *fascinating*. You can lose yourself in that kind of music
for hours.
The only downside is, its complexity makes it hard to hum; you never know
which part you should be humming, so it's not as *catchy* as monody, because
you hum a few bars and get lost, switch parts, hum a few more bars, get lost
again, lose track of where you are, and either quit or start over. But IMO
this elusive quality also makes such music more intriguing.
I'd be interested in seeing what a high-quality contemporary group or
composer would do if they set out to create a modern polyphonic form. I can
imagine Stomp creating a percussive fugue, or Mannheim Steamroller putting
together something polyphonic with nontraditional instruments, or John
Williams weaving together two disparate but equal themes into a movie
soundtrack (say, a theme for the good guys and one for the bad guys, both
going at once whenever both are onscreen). Unfortunately, as near as I can
tell, nobody seems to be composing anything these days that's not monodic,
unless you count the fugue portion of Glad's "History of A Capella Music"
bonus cut.
> Other sites that utilize moderation in a Slashdot style fashion?
> If so, name an example. www.kuro5hin.org doesn't allow post editing.
www.perlmonks.org now allows you to edit your own root nodes. (You could
always edit your own replies, and root nodes in certain sections (e.g., in
Obfuscations), but now you can edit all of them.)
Granted the moderation is a bit different over there, as your karma (err, XP)
depends directly on the voting done by lots of people (not just a handful of
mods), but your post first has to be "approved" by one scribe (or higher),
else it will languish in the realm of only being seen by people who have
turned on unapproved nodes in their prefs. Also, particularly worthless
nodes can be Considered, and with enough downvotes can be Reaped, which
causes them to not be shown inline, but a "Reaped: Reason" with a link to
the actual contents. Oh, and for those special cases where that's not
enough, a small handful of people have the power to nuke a node altogether,
though this is almost never done.
So, basically, whereas the slashdot mod system is oversimple, the perlmonks
one is baroque -- but based on the same basic idea.
> Am I the only one wondering who the hell Norah Jones is?
Yes, you're the only one wondering. Nobody else *cares* who she is.
> learning a video game is a recreational process.
Learning anything computer-related is a recreational process for some and
painful for others, mostly depending on your attitude toward computers. My
mom would consider learning a video game interface just as painful as any
other computer-related thing, and I would consider it no more recreational
than learning a new programming language (assuming it's a fun language to
learn, by which I mean a dynamically typed and multiparadigmatic language).
> I don't know why you are having issues with winXP.
In fairness, most of my problems with WinXP have been directly related to
identifying the model of the video card so as to locate the correct driver
on the manufacturer's website. After that, you install it and reboot a
couple of times, and then it will generally let you change the video
settings to something reasonable. Still, the whole process takes a few
minutes, so the other poster's "it just works" charactarization is a bit
on the optimistic side, IMO.
> But really, how hard is it to get drivers for 2 very well supported chipsets
> and throw them into an OS?
Not too bad, once you identify the chipsets. (It's not always obvious which
chip to read the numbers off of...) But letting HardDrake do this stuff for
you is easier, and I've yet to see it fail to detect a video card. (Some
other kinds of hardware are another matter, granted.)
> IE, not some freaky looking acid-trip OEM box from taiwan that you can't
> find the manufacturer name on.
Actually, my *worst* problems have been with big-name systems, where I had to
reinstall Windows for one reason or another, and the specific model name is
nowhere on the system (it just says, e.g., "Deskpro", which is roughly as
specific as identifying a nursing home resident as "the one with glasses"),
so you don't know which softpaq to get. But then, too, my worst problems
have been not with video cards but with NICs. In two cases I ultimately
gave up and bought replacement NE2000 cards (and later put the old cards into
Mandrake systems, where HardDrake configured them automatically, and all I
had to do was tell it what IP settings to use).
Mind you, I'm not saying that hardware support is universally better in
Linux. (It's not. Some hardware isn't supported at all.) What I'm saying
is that it's not universally "it just works" in Windows, as the other post
seemed to be claiming. That's a significant exaggeration.
> The console commands are confusing to those without experience with them
Well, duh. Dude, it's the eighties. Err, nineties. No, wait... well,
anyway, it's not the seventies anymore, and all the end users are using this
new-fangled underpowered type of interface with graphics, where they point
at things with a mouse device and, you know, click.
The commandline is for powerusers. And, for powerusers, the commandline in
Windows (current generation; bring this up again in 2006 and we'll talk) is
pretty vastly inferior to the 1975 state of the art in Unix command shells.
When I'm on Windows, I find myself having to use a GUI file manager (Explorer
or some equivalent) for certain things, even though it's clunky, because
doing them from the command line, as I would normally do on any POSIX system,
is difficult or impossible. (Yeah, yeah, cygwin, but if it was a system I
was going to use on an ongoing basis I'd just install Mandrake. We're talking
here about doing maintenance on systems that are mostly used by other people.)
> From someone without computer experience, what the hell is "mounting" a
> hard drive?
Something you don't have to worry about. The installer for any of the
major newbie-friendly distros (RedHat, Mandrake, and so forth -- as opposed
to, say, Gentoo, which is intended for people with some *nix experience)
takes care of setting up fstab automatically. If your distro includes
Supermount (as e.g. Mandrake does) you don't have to manually mount your
floppies or CDs either. (I always turn this off, but I'm a poweruser.)
It sounds to me like you've either been using a quite old version, or a
distro that wasn't intended for newbies.
> Where do you put your pictures? Not in "C:\pictures", but in
> "/mnt/users/username/home/pictuers" or somesuch.
You click on the little picture of a house on your panel, and a window opens
up (that's your "home folder" inside your computer, where you keep all your
stuff; the equivalent on WinXP is "My Documents") and then click on the
"Pictures" folder. That's where you keep your pictures.
> Hey, how about "How does someone with no Linux experience install things?"
Under the foot menu, click on "Configuration" and then "Install Software".
The Software Manager (technically known in geek circles as rpmdrake, but
you don't need to know that) will come up. Checkmark the packages you want
to install, and click the "Install" button.
What, you mean *download and install*? End users don't do that, under any
OS, generally. They get someone like me to do it for them, or (you're going
to think I'm kidding here, but I've seen 'em do this for office software)
they go buy a computer that has the software in question installed on it.
(*That* can get expensive...) They're either afraid to install software
(because it might mess something up or possibly give them a virus) or else
they just plain don't know how. If you're comfortable installing software,
I'd classify you as at least somewhat a poweruser (though, of course, there
are a range of abilities among powerusers; I'm more toward the high end of
it myself, pushing the boundary between poweruser and developer; I don't
actually write software that anyone uses besides myself, unless you count
CGI-based database frontends, but I know and use Emacs lisp, understand
algorithm analysis, and can think in Perl... so I'm *almost* a developer.)
Just having the *desire* to install software that didn't come preinstalled
on your computer betrays a poweruser tendency in your mindset: you want to
make the computer do more, add capabilities to its feature set. End users
don't think that way; they don't even want to learn to use the features
they've already got, except for the smallest subset they need to get their
stuff done. (Learn ways to get it done better, faster, and easier? No
thanks, they haven't got time for that. They're busy, after all, and they
have other things to do than sit in front of a computer all the time.)
> It is really a shame that someone like Canon does not create some linux
> software and include it with their cameras.
This is silly. Doing that, you have to deal with different configurations
on the user's PC, and then when you go to write Windows and Mac interfaces
you have to start over from scratch.
Much better to just embed a web server in the camera/scanner/whatever. This is
no harder than creating interface software for any one platform, and it gets
you all platforms, and as an added bonus, it gets you a user interface that
nearly everybody already understands. Given the price of NE2000 network cards
these days, you can bundle one and a short crossover cable for roughly the
same cost as the packaging, just in case the user doesn't have a LAN setup
already. Embed a DHCP server on the camera and Bob is everybody's uncle.
> I can throw 2 video cards into windows 2000 and they just.... work.
Wow. That's awesome. Maybe I should recommend a Win2k site license at work,
so we can migrate all our Win98 and WinXP systems to it. Because in Win98 and
WinXP, you're lucky if *one* video card works when you plug it in; usually,
after each Windows reinstall you have to spend fifteen minutes or so (if you've
done it a lot of times before -- it can take a newbie an hour) getting the
graphics card to work in anything but 16-color mode at the minimum res and
60Hz refresh rate. That's if you know the manufacturer and model number of
your graphics card (or motherboard for onboard video); if you don't, you're
pretty much totally screwed.
> As much as I hate spam, it shouldn't be a criminal offense, and especially
/. is concentrating on what this bill *doesn't* do (it
> should not have a prison sentence.
This bill does not make spamming illegal (it specifically makes it legal, and
regulates it), and the only things it attaches prison sentences to are the
online equivalents of crimes that in the real world can already get you jail
time -- things like fraud.
Almost everyone on
doesn't outright make spam illegal), but what it *does* do (within the US) is
make some of the more heinous abuses (e.g., forging headers) illegal. That's
a good thing. I said months ago, if we can just make forging headers illegal
worldwide, just that alone will make spam twenty times easier to filter. This
bill does that in the US, which is the first step. Does it stop spam? No.
Can it help make it more manageable? Yeah, it can. Just the provision against
forging headers is worth passing it for. The spammers will of course move
overseas, but the same was true of copyright-spurning printing houses[1], at
one time -- until most of the world adopted international copyright law.
But some nation has to lead the way.
[1] I'm talking about the businesses based on printing many copies of a
copyrighted work without permission and selling them in bulk at a lower
price than the legitimate publisher, cutting the author out entirely.
Yes, it's easier to stop such books from being imported than it is to
stop spam at the borders, but it remains that most of the world now has
adopted international copyright law -- pretty much every country with
phone lines. If we get that level of cooperation for stopping the
forging of email headers, we can put away the bayesian classification
and markov chain analysis and go back to simple domain-based blacklists.
> I would expect, all factors being equal, to receive some spam in Hindi...
Now that I think about it, I wonder if all the major ISPs in India use the local
two-digit TLD, which would have the side-effect of making it easy for spammers
using those languages to target their spam toward people who can maybe read it.