> Can you elaborate a little bit on what you mean by 'painfully > illogical backwards step'?
I don't know for sure what the other poster meant, but I can elaborate on what *I* think of when I think of painfully illogical backwards steps in the WinXP UI, as compared to Win9x. I think of opening up Windows Explorer windows only to find that, instead of showing you the contents of the folder, it shows you something else. I think of going to Explore view only to find that, when I click on a folder in the tree view at left, instead of just highlighting that folder and showing me the contents, it takes away the tree view at left and shows something entirely different, and after fifteen minutes of fiddling I eventually find that the way to get the Explore view back involves changing which sidebar is checked in some menu. I think of even *more* stupid checked-by-default options under View->Folder Options that I have to uncheck to keep Explorer from hiding various things from me. (Windows 95 had *none* of these; Windows 98 had the ones for hiding the file extension, and I think hiding the full path. I suppose Longhorn will have about three dozen of them for hiding everything up to and including the fact that you're using a computer at all, like some kind of demented two-dimensional Matrix clone trying to keep you from becomming self-aware.) I think of a My Network Places that, unlike Network Neighborhood in Win98, does not actually show you the computers in your workgroup on your local network unless you first go through Entire Network and then Microsoft Windows Network and then click on your workgroup. Three extra steps, and even then frequently it cannot be made, through any amount of rebooting or network-reconfiguring, be made to see one-half to one-third of the other Windows computers on the network, for no apparent reason, even though the ones it can see can see the others, and vice versa, and Samba on Linux has no trouble seeing all of them. (Reinstalling Windows XP is the only way I've discovered to fix this. I think running the Network Configuration Wizard is a common cause of this problem, but I haven't tracked down the precise details.) I think of a start menu that pushes all the programs another whole level deeper into the hierarchy, for no good reason. I think of a control panel that is completely reorganized, to be twice as unintuitive as ever and scare newbies even *more*, pretty icons notwithstanding. (They could have introduced prettier icons without dorking with the arrangmenent...) I think of menu items that mysteriously disappear, so that every other time you use a menu you have to hunt down a small double-arrow thingydoo just to get your menu entries back. I think of little pop-up tooltip-like things that jump out of the system tray every seven or eight minutes and bug you with questions that either A) don't require any action or B) are confusing to end users or C) both. (The worse offenders here are security-related things; as near as I can determine, the *only* way to avoid having WindowsUpdate and NAV pester the user incessently is to disable them completely; they both have a "just do it and don't bug me option", but if you turn said option on, it still bugs the user all the time. So desktop sysadmins are essentially forced to disable all security-related features, which is not really what they want to do, but you *can't* have every desktop user calling you up three times a day to come deal with these things.) I could go on, but I think this paragraph is long enough now.
WinXP has memory protection, which is really important, plus filesystem-level permissions, which are nice to have, so in that respect it's a much better option than Win9x. And it has better hardware/driver support than Win2000, and some UI enhancements as well. But it sure is not without its stupid idiotic changes nobody would ever conceivably want. And yeah, most of the really horrid ones can be turned off, but I can definitely see where the other poster is coming from.
There are some things I use fixed-width fonts for exclusively, notably email and all source code (both markup and programming languages). Mostly I use either Andale Mono or Bitstream Vera Sans Mono, although for some things I still use bitmapped screen fonts.
> I have also found that my head OCR's faster with certain fonts (Courier 10 > point, San Serif 10 or 12 point) than in all the fruity variable width fonts.
Courier is one of the ugliest, most bletcherous font families in the history of the universe. Especially Courier New. I can't *stand* Courier New. For fixed-width, I like Andale Mono or Bitstream Vera Sans Mono. Lucida Console will do in a pinch, if neither of the others is installed for some bizarre reason.
> I like to bash MS as well, but Georgia and (especially) Verdana are lovely
I like Georgia and Verdana as well. Also Andale Mono. At work, I have them on my checklist of stuff to install on every new Windows system. Occasionally a new Windows system comes with either Georgia or Verdana already installed, but so far in my experience never both, and never Andale Mono. I don't know if this is the OEMs screwing things up or what, but I always have to track them down -- formerly at MS Typography, these days at corefonts.sourceforge.net instead -- and install them myself. (And people wonder why the Core Fonts initiative never succeeded in getting all the world's webmasters to rely on those fonts being installed. The webmasters' own systems probably didn't have them half the time.)
Other things on my new-Windows-system install checklist include TweakUI, Pegasus Mail, a decent browser (these days usually Firefox), OpenOffice, ActivePerl, the lastest Acrobat Reader, a recent Java vm, and the CommandPromptHere powertoy. Microsoft *ought* to bundle at least half of that (specifically, TweakUI and the CommandPromptHere should just be standard, and the Core Fonts of course, and I'm sure MS could reach an agreement with Adobe on acroread, and there's no excuse for not including a decent browser in any modern OS these days, and runtimes for Perl and Java are standard in every other OS I think, so MS really needs to get on the stick in terms of these things; and if MS doesn't, the OEMs should at least make some attempt to fill in the most glaring gaps; it's pathetic how worthless a new computer is out of the box, until you download umpteen updates and enhancements, all of which are free downloads and most of which presumably could have been included by the manufacturer if they would just do so).
> Say what you want about the Windows UI. In my opinion it is far better than > the OS X UI, NOT because of its inherent functionalty, but because its far > more customizable than the Mac UI.
If you like a customizeable UI, you really should have a look at Gnome. Screenshot: http://www.bright.net/~jonada b/screenshot.png
My favourite feature is the ability to have panel drawers (visible as the icons with little white arrows in their lower-right corners, on the left panel), each of which can contain a full row of icons; this completely obviates the need for icons on the desktop, entirely removing the need to minimize stuff all the time. But in general, the Gnome UI is *way* more customizeable than the Windows one. You'll notice I've removed the Gnome foot (equivalent of the Start button/menu), because I never use it.
Yeah, it wouldn't be completely *worthless* then or anything, considering each entry would be the minimum size, so you wouldn't be able to tell one window from another, and in addition they wouldn't all fit.
When I used Windows as my main desktop, I settled on putting the taskbar on the left edge, sized to about a third of the screen, so there was room for three columns of task list buttons, and set to autohide so it didn't take up a third of the screen permanently. This was rather suboptimal, but it seemed to be the best I could do.
For some reason, I seem to have a lot fewer windows open on my Gnome system. (No, it's not because of task grouping -- I have that turned off, as I dislike it.) One reason is tabbed browsing, of course -- it didn't exist yet (in any browser) back when I used Windows. But I think the big reason is that with a decent command shell, I no longer need to use a graphical file manager all the time, so that's about twelve or so Windows Explorer windows I don't have open, because I do all my file management from either one of the terminal windows (which correspond to command windows I had open in Windows) or else in eshell in Emacs (which, again, is something I also had open in Windows). Gimp2 helps also, because with the docking I get by with just one more windows than the number of images I have open (so, usually two windows instead of six or so). Hmmm... I guess maybe the difference is not so much Windows versus Gnome as it is three years ago versus now. I guess the Windows Explorer windows are the only difference I've discussed that is platform-related really -- and that one will theoretically go away in Longhorn, when the spiffy new command shell is introduced.
There is also the minor difference of not having a start menu and quicklaunch on the left (but that only makes room for about one more window) and system tray on the right, so the task list takes up the entire bottom panel. ISTR having twelve or fifteen items in my system tray on Windows 95, just to get up to a basic level of functionality. Come to think of it, I don't know that that's really necessary in recent versions of Windows, at least not to that extent.
Eh, what do I know? Trying to compare Mandrake 9.2 to Windows 95 isn't fair. Forget I said anything.
I'm still not sure about the universality of the practicality of having a one-row taskbar in all cases, though.
I think the ultimate alarm clock program would play Buddy's Carpet commercials, or possibly car dealership commercials, but in the absense of that, a fairly effective and very easy one to write does this:
* Uses a fractal to generate the sound. A basic plasma fractal in one dimension (wavelength) creates a rather irritating sound that is hard to ignore. * Increases the volume with each passing second. * Is easy to snooze (press any key), but requires significant brain function to turn off (type complicated command sequence involving the current date and time, which must be accurate within one minute, so you can't just type it with muscle memory). * Reduces the length of the snooze period by 1/3 to 1/2 each iteration. * Stores the wake time on the filesystem, and checks it at system startup time. This way a power outage during the night does not foil the alarm clock.
When I was in college, I wrote this in GW-BASIC on my 8086 system and used it for several years.
> Debian stable eschews the bleeding edge in favor of reliability.
The bleeding edge? What an understatement! Debian stable eschews the whole modern era in favor of metaphorical stone knives and bearskins. It is not so much venerable as decrepit, senile. Installing it is like travelling back in time the better part of a decade, to the era when 2.4 was the new kernel and not everyone used it, when X servers had to have their settings tweaked and modelines frequently had to be written by hand, when dinosaurs roamed the earth, the era when most laptops didn't work, and a lot of desktops didn't have all their hardware supported either. Remember that? Wasn't that fun? Now you can experience the late nineties _again_. Just install Debian Stable, and re-live the spent days of your lost youth...
There were also some applications that required UMBs. At one time I had a set of BAT files that changed AUTOEXEC.BAT to different versions for the different configurations required.
> USB 2.0 Full-speed (12 Mbps) > USB 2.0 Hi-speed (480 Mbps)
Screen resolutions were like that... 300 x 200 Medium Resolution (CGA four-color mode) 600 x 200 High Resolution (CGA monochrome graphics mode) And then EGA and VGA came along and more superlatives had to be invented to describe the resolutions that were higher than high. Eventually everyone decided to just stick with the numbers, because it's actually easier to remember which resolution is which that way; when we say 800x600 and 1024x768 and 1280x1024, everyone knows which is the highest of those three resolutions.
Oh, and do you remember Full Color, High Color, and True Color? Fortunately these days we can just say 24bpp or 32bpp or whatever it is, which is much easier to keep straight.
Will USB/1394/whatever head in that direction, so that we can just talk about the port's (or the device's) speed, rather than quoting version numbers? Wouldn't that be nice?
Also, how about floppies: 360K Double-sided Double-density (5.25") 1.2M Double-sided High-density (5.25") 720K Double-sided Double-density (3.5") 1.4M Double-sided High-density (3.5") There were also at one time single-sided and/or single-density 5.25" diskettes and drives. (And if you had double-sided diskettes and a single-sided drive, you could flip the diskette over and format the other side and use it separately, without interfering with the first side; people used to put a program on one side and data on the other, sometimes.) I don't know whether there were ever single-sided 3.5" diskettes.
> if you continue this "open up a terminal lamer" attitude you won't have > too much future in OS X or Win32
The other poster was perhaps not clear enough. I will try to explain this clearly so that you can understand it: some programs can never be useful to people who don't want to open up a terminal, just because of the inherently very technical nature of what the programs do. panpipes is exactly the sort of program that such a person would never be interested in using. All it does is make the computer crash. That's all it does. It doesn't do anything else. Furthermore, the whole point of writing it was so that people could analyze it and try to figure out how it does that, in terms of the technical under-the-hood things, the nuts and bolts of what it does to make the computer crash.
Suggesting that a program like that needs a GUI, just because end users don't like to open up terminals, is completely missing the point. This program does not do anything an end user would want to do. Ever. Under any circumstances. Making a GUI for it would be utterly and completely worthless, a total waste of time. People who analyze code that causes kernel panics are *NOT* the sort of people who need a GUI for everything and don't want to open a terminal. For crying out loud, they had to use gdb to analyze the thing -- gdb is really hardcore technical stuff, *way* beyond not having a GUI. Even most people who use the terminal on an hourly bases are not experienced using gdb (or any other lowlevel debugger like that). There is absolutely no way on earth that anyone who doesn't want to open up a terminal window and thinks all applications should have a GUI would have *any* interest *whatsoever* in participating in this sort of contest. If they thought they were interested, it is obviously because they saw the word "contest" and didn't read the explanation of what kind of contest it was, what it was actually about, or what they had to do to participate. If they had, they would NOT have been interested.
As for the other part of your post, suggesting that the other poster will somehow not have a future, just because he writes programs that are technical in nature... that does not even deserve a response. I suppose the people who design microchips are doomed to extinction by next week, because regular people are not interested in examining their work.
I wouldn't think so. The moon is only, what, 280 thousand miles out from the Earth's surface? Give or take? Assuming your networking equipment is optical, you should be able to achieve reasonable ping times. Good enough that a sysadmin in Denver could ssh in and perform system maintenance in an emergency, I should think. Granted, the networking equipment would cost a little more than residential-consumer-grade WiFi stuff...
> Bush said he wants to use it as a stepping stone for Mars.
I don't know the context of this quote, but is it possible he meant a stepping stone in terms of experience, knowledge gained, et cetera? That seems more likely to me than a physical stepping stone in terms of launching ships. We can launch ships from here, but a lunar base would be within a few days' reach in an emergency; whereas, a Mars base would not be, so an argument could be made for using a lunar base first to work out the kinks in the base planning methodology.
We theoretically have the technology to go to Mars, yes. That doesn't make it an easy thing at all, nor cheap. We do *not* have the technology to establish a self-sustaining base there (not on the Moon either, though), so you're looking at a short-term purely scientific endeavor, one that's really *really* expensive and requires a whole new *kind* of astronaut, one who's willing to live in a confined space for *years* at a time with effectively *zero* possibility of returning to Earth on a shorter timeframe in an emergency. Possible candidates for the trip should be screened first for the ability to stay a couple of years at someplace like Vostok (or at least Amundsen-Scott), come back sane, and still want to go to Mars.
> Yes, just a big rock, chock full of raw materials we need for your > trip to Mars
What specific material is on the Moon that we need in order to go to Mars?
> Once established, the Moon Base will py for itself countless times over.
Now you're completely out of your chair. The cost of maintaining it would absolutely dwarf any possible financial benefit. You know that the Moon has no atmosphere, right? No atmosphere? Have you got that? A base there cannot sustain itself; it needs continual supplies from Earth, not least of all food. That means regular supply trips, which have to deal with full Earth gravity, every little bit, for the whole time we maintain the base. The only reasons to go to the Moon or establish a base there are strictly scientific in nature. Economically, it's a big fat negative. Heck, even inland Antarctica, which is *way* less hostile an environment than the moon, is economically-speaking a completely worthless destination. The Moon is even worse. (So, for that matter, is Mars.)
That doesn't (necessarily) mean we shouldn't do it, but expecting it to pay for itself, much less anything else, is industrial-strength naive. It will not pay for itself. It will cost a fortune to build and a fortune every year to maintain.
> Ultimately it doesn't matter who owns it until there's some feasable way > to get there.
Getting there isn't too hard. Okay, so it's not an afternoon jaunt, but it's been done a handful of times, and that was without modern technology. The hard part is figuring out how to derive substantial benefit from staying there long term. Nobody's solved that one yet. Visiting the moon is an interesting endeavor, but after a short stay, everybody seems to want to come back to Earth. I guess it's not a tangible thing, just some kind of nebulous psychological thing or something. People keep saying Earth has "better atmosphere", whatever that means.
I just use grip with the "rip only" option (no encoding). Metadata are stored at the filesystem level: the author's name as a directory, and within that the album title as a directory, and within that the track name as a filename, with.wav postpended. Works for me. One nice thing about WAV is that it's lossless. Another is that it's supported by, you know, absolutely everything.
> This problem is that Linux is still largely a text interface if you want > to tap into any of its power.
This is also true of Win32. And the shell is pathetically bad -- though the new one in Longhorn is supposed to be an earthshattering improvement, and I'm inclined to believe it.
I have a feeling you're confusing "easier to use" with "easier to *learn* to use", which is a different thing. The learning curve for Windows starts off less steeply than the one for *nix. (Note that I'm talking about the first part of the learning curve, the part that a newbie encounters; once you reach the power user level, the curves on both platforms change, but in different ways.) However, once you've learned how to use it, the command line (on both platforms) makes a lot of things WAY easier to do than the equivalent GUI way, and the *nix command line is (at this time) indisputably better than the NT command line, which in turn is universally regarded as superior to the Win9x comand line, which in turn puts the (classic) MacOS command-line completely to shame:-)
> Go a week without using a shell. Easier said than done.
*shrug* Go a week without using a web browser. Go a week without using a text editor. What's your point, exactly? Prohibiting the use of whole classes of important standard application categories is *always* going to be rather limiting.
Sure, new users don't use some of the more advanced types of applications -- command line stuff, text editors, raytracers, compilers, image editors (except maybe those ones that come with digital cameras and have auto-remove-redeye buttons and no real features; whether those are really image editors is an open question), remote login facilities (not just the command-line ones; new users don't use RDC or VNC or cetera either; generally even mentioning the possibility of controlling a second computer remotely will overwhelm them) -- anything they don't have an immediate felt need for, because, being new, they still have plenty of basic stuff on their plate to learn, without adding more esoteric things to the cirruculum. Indeed, some people dabble with computers for several years before they even learn to use such essential things as the clipboard.
And yes, some users learn things (everything) more rapidly than others, mostly due to motivational factors, and to a lesser extent due to aptitude; and yes, some users reach a certain point and decide that's pretty much enough for them, at least for a good while (though often they change their mind several years later and want to learn how to do X, Y, or Z). Which is fine.
But none of that is specific to any particular platform.
> I have NEVER had that good of luck with any version of windows...
Luck is what you make of it. Windows does require more reboots, and of course it helps to be at least a Windows poweruser (just the same as it helps to be a Linux poweruser if you're installing that -- and don't get me started on installing BSD), and sometimes you will have to download drivers from the manufacturer, but despite all that the success rate for an experienced user can be very high, as almost all hardware is capable of running it (or vice versa, depending how you look at that) -- much higher than 78%. (Although, that number also seems low to me for modern Linux distros -- what are you installing, Debian stable?)
There are, of course, some things to know when installing a Microsoft OS. Number one is, get the data you want to keep off, fdisk the thing, and create new partitions from scratch. Always. Saves HOURS of headaches. (In a pinch, if you really HAVE to keep the partition in tact for data's sake, installing somewhere other than C:\WINDOWS can work as an interim solution, if you have enough extra drive space, but I recommend against it if it can possibly be avoided, since the cruft left around by this approach can really wreak havoc with application installations.) Number two, as soon as you have a decent working state, create a system restore point *immediately*, *before* you tweak anything, *especially* the network settings. But, there are not-so-dissimilar equivalents to these with the Linux distributions I have used also. Getting networking to work, and keeping it working while you set everything else up, is generally easier with Linux, but OTOH getting printers to work can be more of a pain. Really, they're quite comparable, albeit certainly not the same.
> Next week Jacob will be telling us how washing machines are great for > keeping your clothes clean but not very good for making cups of tea.
The thing is, a lot of people over the years have tried to use PDF as a format for online distribution of content. It sucks for that. In many cases, HTML or even plain ASCII would have been a MUCH better choice.
If you're using PDF for distributing content that is no good unless printed (on a certain size and shape of paper), such as IRS tax forms, then that's a good use of PDF. If you're using PDF to deliver documents to a commercial printer, that's another good use of PDF. If you're using it to protect the precious layout of your white paper, the world would be better served if you stuck your precious layout in the nearest available unsavory bodily orifice and distributed the content as HTML or plain text.
The release of xpdf I am comparing to is not the latest and greatest (it's not an app that I use all that often), so maybe this has been fixed, but last I knew xpdf does not handle transparent images very well, often giving them a black background when in fact it should be white; if there is text in front of the images, this can result in documents' being illegible. acroread also used to have this problem in about version 4, but it has been fixed for some time now.
> Now if only Adobe would bring Photoshop over as well...
What, and give the Gimp developers the ability to easily run the two side-by-side on their platform of choice, directly compare them, and so forth? I'm not sure the Photoshop team really wants that level of competition. Gimp is already fairly impressive in terms of functionality, but the Gimp developers as a rule don't have Photoshop to compare to, so the interfaces are quite different. (This can be construed as a good thing... certainly it's good for allowing the respective applications to retain their existing user bases, which is much to Adobe's advantage.)
It may be that the Gimp developers would not spend the cash on Photoshop even if they *could* run it on *nix, and so it could be that little or nothing would change. But I can quite readily imagine Adobe's not wanting to chance it.
> To most end users, a consistent look and feel, that works right out of the > box, is really important.
Yeah, but Win32 has never had this, because it allows application developers to make their own widgets out of whole cloth, resulting in fruitcakey nonsense UIs like AIM, Winamp, et cetera, ad infinitum, ad nauseam, ad bedlam. Yes, X11 allows this also; fortunately, *most* of the OSS community these days is beginning to see the folly of this approach and use one of the leading GUI widget sets, mainly GTK and/or Qt, which can be themed to look highly similar or at least coordinate nicely, and some of the major apps with their own toolkits (OOo, Mozilla,...) are at _least_ starting to pick up basic color schemes from the system now, although really they ought to use the actual native widgets. But it only takes one $#@! Athena application, or something like xmms, to spoil the consistency.
The thing is, it is utter hypocracy for Microsoft advocates to criticize "Linux" (by which they mean X11 or the distros) for this, because exactly the same thing is true, and to pretty much exactly the same extent, on Win32. Okay, so Win32 has one core widget set instead of two (GTK and Qt), but the end user does not need to know about the GTK/Qt distinction, if they are themed similarly.
For Microsoft, there's an obvious solution to this: deprecate the parts of the API that allow custom widgets, and in five years remove support for them. ISVs won't like this one bit, and the question is, which is more important: retaining ISV mindshare, or giving users a consistent look in every app? I tend to think the ISV mindshare is more important, but Microsoft does have this option to consider. In the OSS community, of course, there's no such option, because of the nature of the freedoms developers inherently have -- although distributions could refuse to *bundle* non-compliant apps. (But, no sane distribution would refuse to bundle, say, Emacs, which last I checked is still built against Xaw3D or some such ugliness, presumably because for some strange reason pretty scrollbars and tear-off menus don't seem to be a major show-stopping priority for Emacs developers. XMMS is also pretty major, but I think that one is sufficiently modular that they could adjust its GUI to comply, rather than exclude it. Command-line programs, of course, are exempt from GUI-oriented visual consistency guidelines; nobody expects cd to use a GTK2 file selection dialog, for instance; that way lies madness.)
Mandrake does put some of their own work into their distro. The *drake tools (printerdrake and so forth) really do speed up initial configuration -- especially for relative *nix newbies, but even for experienced users they are convenient.
Not that Ubuntu is bad, but that's another thread.
That's because ping is braindead. You won't have this problem if you construct the echo-request packets yourself. (Okay, so the packets do have to have the IP addresses in them -- but you can send them directly to a certain MAC address.)
> Why is it ridiculous to run stable applications at work?
Well, let's all just run PC-DOS 3.3 while we're at it.
Seriously. PC-DOS 3.3 was very stable; I never managed to find a bug in it, period, not even a little one. Then MS-DOS 5.0 came out, and it had all kinds of bugs. 6.0 was only a little better. 6.22 had most of the bugs ironed out, but it was still not as stable as 3.3. Let's all use 3.3, since it's stable.
Do you see the problem?
Installing Woody is like travelling back in time the better part of a decade. Just for starters, the installer is not something that could ever be recommended in a work setting: you have to use dselect. Manually. It makes NT4 look positively modern. And the installer is just the beginning. Once you get it installed, you find that any application released in the last couple of years won't run on it. You want to run an application that requires GTK2? Haha, you're joking, right? Heavens, you wouldn't want anything so bleeding edge as *that* on your system; after all, it's only three or four years old, so just get it out of your head.
There's stable, and then there's really stable, and then there's Just Plain Old. Woody, at this point, is too far into the last category.
There's nothing stopping people who don't need new software from continuing to use old software, such as Woody, or Potato, or TOPS20. But when your *latest* stable release has such thick cobwebs on it that people with seven years of Linux experience try to install it and immediately go "Yikes!", it's time to get a slightly more modern one out the door soon.
The marketing fluff is normal -- in fact, I barely noticed it, as out of long habbit I've trained myself to read through hype and pick out the relevant bits (at least, usually); it's the grammatical anomoly that I didn't expect to see coming out of Apple. They're usually relatively careful about that.
Slashdot Mirror
> Can you elaborate a little bit on what you mean by 'painfully
> illogical backwards step'?
I don't know for sure what the other poster meant, but I can elaborate on what *I* think of when I think of painfully illogical backwards steps in the WinXP UI, as compared to Win9x. I think of opening up Windows Explorer windows only to find that, instead of showing you the contents of the folder, it shows you something else. I think of going to Explore view only to find that, when I click on a folder in the tree view at left, instead of just highlighting that folder and showing me the contents, it takes away the tree view at left and shows something entirely different, and after fifteen minutes of fiddling I eventually find that the way to get the Explore view back involves changing which sidebar is checked in some menu. I think of even *more* stupid checked-by-default options under View->Folder Options that I have to uncheck to keep Explorer from hiding various things from me. (Windows 95 had *none* of these; Windows 98 had the ones for hiding the file extension, and I think hiding the full path. I suppose Longhorn will have about three dozen of them for hiding everything up to and including the fact that you're using a computer at all, like some kind of demented two-dimensional Matrix clone trying to keep you from becomming self-aware.) I think of a My Network Places that, unlike Network Neighborhood in Win98, does not actually show you the computers in your workgroup on your local network unless you first go through Entire Network and then Microsoft Windows Network and then click on your workgroup. Three extra steps, and even then frequently it cannot be made, through any amount of rebooting or network-reconfiguring, be made to see one-half to one-third of the other Windows computers on the network, for no apparent reason, even though the ones it can see can see the others, and vice versa, and Samba on Linux has no trouble seeing all of them. (Reinstalling Windows XP is the only way I've discovered to fix this. I think running the Network Configuration Wizard is a common cause of this problem, but I haven't tracked down the precise details.) I think of a start menu that pushes all the programs another whole level deeper into the hierarchy, for no good reason. I think of a control panel that is completely reorganized, to be twice as unintuitive as ever and scare newbies even *more*, pretty icons notwithstanding. (They could have introduced prettier icons without dorking with the arrangmenent...) I think of menu items that mysteriously disappear, so that every other time you use a menu you have to hunt down a small double-arrow thingydoo just to get your menu entries back. I think of little pop-up tooltip-like things that jump out of the system tray every seven or eight minutes and bug you with questions that either A) don't require any action or B) are confusing to end users or C) both. (The worse offenders here are security-related things; as near as I can determine, the *only* way to avoid having WindowsUpdate and NAV pester the user incessently is to disable them completely; they both have a "just do it and don't bug me option", but if you turn said option on, it still bugs the user all the time. So desktop sysadmins are essentially forced to disable all security-related features, which is not really what they want to do, but you *can't* have every desktop user calling you up three times a day to come deal with these things.) I could go on, but I think this paragraph is long enough now.
WinXP has memory protection, which is really important, plus filesystem-level permissions, which are nice to have, so in that respect it's a much better option than Win9x. And it has better hardware/driver support than Win2000, and some UI enhancements as well. But it sure is not without its stupid idiotic changes nobody would ever conceivably want. And yeah, most of the really horrid ones can be turned off, but I can definitely see where the other poster is coming from.
There are some things I use fixed-width fonts for exclusively, notably email and all source code (both markup and programming languages). Mostly I use either Andale Mono or Bitstream Vera Sans Mono, although for some things I still use bitmapped screen fonts.
> I have also found that my head OCR's faster with certain fonts (Courier 10
> point, San Serif 10 or 12 point) than in all the fruity variable width fonts.
Courier is one of the ugliest, most bletcherous font families in the history of the universe. Especially Courier New. I can't *stand* Courier New. For fixed-width, I like Andale Mono or Bitstream Vera Sans Mono. Lucida Console will do in a pinch, if neither of the others is installed for some bizarre reason.
> I like to bash MS as well, but Georgia and (especially) Verdana are lovely
I like Georgia and Verdana as well. Also Andale Mono. At work, I have them on my checklist of stuff to install on every new Windows system. Occasionally a new Windows system comes with either Georgia or Verdana already installed, but so far in my experience never both, and never Andale Mono. I don't know if this is the OEMs screwing things up or what, but I always have to track them down -- formerly at MS Typography, these days at corefonts.sourceforge.net instead -- and install them myself. (And people wonder why the Core Fonts initiative never succeeded in getting all the world's webmasters to rely on those fonts being installed. The webmasters' own systems probably didn't have them half the time.)
Other things on my new-Windows-system install checklist include TweakUI, Pegasus Mail, a decent browser (these days usually Firefox), OpenOffice, ActivePerl, the lastest Acrobat Reader, a recent Java vm, and the CommandPromptHere powertoy. Microsoft *ought* to bundle at least half of that (specifically, TweakUI and the CommandPromptHere should just be standard, and the Core Fonts of course, and I'm sure MS could reach an agreement with Adobe on acroread, and there's no excuse for not including a decent browser in any modern OS these days, and runtimes for Perl and Java are standard in every other OS I think, so MS really needs to get on the stick in terms of these things; and if MS doesn't, the OEMs should at least make some attempt to fill in the most glaring gaps; it's pathetic how worthless a new computer is out of the box, until you download umpteen updates and enhancements, all of which are free downloads and most of which presumably could have been included by the manufacturer if they would just do so).
> Say what you want about the Windows UI. In my opinion it is far better than
a b/screenshot.png
> the OS X UI, NOT because of its inherent functionalty, but because its far
> more customizable than the Mac UI.
If you like a customizeable UI, you really should have a look at Gnome.
Screenshot:
http://www.bright.net/~jonad
My favourite feature is the ability to have panel drawers (visible as the icons with little white arrows in their lower-right corners, on the left panel), each of which can contain a full row of icons; this completely obviates the need for icons on the desktop, entirely removing the need to minimize stuff all the time. But in general, the Gnome UI is *way* more customizeable than the Windows one. You'll notice I've removed the Gnome foot (equivalent of the Start button/menu), because I never use it.
> Taskbar should be one row.
Yeah, it wouldn't be completely *worthless* then or anything, considering each entry would be the minimum size, so you wouldn't be able to tell one window from another, and in addition they wouldn't all fit.
When I used Windows as my main desktop, I settled on putting the taskbar on the left edge, sized to about a third of the screen, so there was room for three columns of task list buttons, and set to autohide so it didn't take up a third of the screen permanently. This was rather suboptimal, but it seemed to be the best I could do.
For some reason, I seem to have a lot fewer windows open on my Gnome system. (No, it's not because of task grouping -- I have that turned off, as I dislike it.) One reason is tabbed browsing, of course -- it didn't exist yet (in any browser) back when I used Windows. But I think the big reason is that with a decent command shell, I no longer need to use a graphical file manager all the time, so that's about twelve or so Windows Explorer windows I don't have open, because I do all my file management from either one of the terminal windows (which correspond to command windows I had open in Windows) or else in eshell in Emacs (which, again, is something I also had open in Windows). Gimp2 helps also, because with the docking I get by with just one more windows than the number of images I have open (so, usually two windows instead of six or so).
Hmmm... I guess maybe the difference is not so much Windows versus Gnome as it is three years ago versus now. I guess the Windows Explorer windows are the only difference I've discussed that is platform-related really -- and that one will theoretically go away in Longhorn, when the spiffy new command shell is introduced.
There is also the minor difference of not having a start menu and quicklaunch on the left (but that only makes room for about one more window) and system tray on the right, so the task list takes up the entire bottom panel. ISTR having twelve or fifteen items in my system tray on Windows 95, just to get up to a basic level of functionality. Come to think of it, I don't know that that's really necessary in recent versions of Windows, at least not to that extent.
Eh, what do I know? Trying to compare Mandrake 9.2 to Windows 95 isn't fair. Forget I said anything.
I'm still not sure about the universality of the practicality of having a one-row taskbar in all cases, though.
I think the ultimate alarm clock program would play Buddy's Carpet commercials, or possibly car dealership commercials, but in the absense of that, a fairly effective and very easy one to write does this:
* Uses a fractal to generate the sound. A basic plasma fractal in one dimension (wavelength) creates a rather irritating sound that is hard to ignore.
* Increases the volume with each passing second.
* Is easy to snooze (press any key), but requires significant brain function to turn off (type complicated command sequence involving the current date and time, which must be accurate within one minute, so you can't just type it with muscle memory).
* Reduces the length of the snooze period by 1/3 to 1/2 each iteration.
* Stores the wake time on the filesystem, and checks it at system startup time. This way a power outage during the night does not foil the alarm clock.
When I was in college, I wrote this in GW-BASIC on my 8086 system and used it for several years.
> Debian stable eschews the bleeding edge in favor of reliability.
The bleeding edge? What an understatement! Debian stable eschews the whole modern era in favor of metaphorical stone knives and bearskins. It is not so much venerable as decrepit, senile. Installing it is like travelling back in time the better part of a decade, to the era when 2.4 was the new kernel and not everyone used it, when X servers had to have their settings tweaked and modelines frequently had to be written by hand, when dinosaurs roamed the earth, the era when most laptops didn't work, and a lot of desktops didn't have all their hardware supported either. Remember that? Wasn't that fun? Now you can experience the late nineties _again_. Just install Debian Stable, and re-live the spent days of your lost youth...
> XMS - eXtended Memory Specification
> EMS - Expanded Memory Specification
There were also some applications that required UMBs. At one time I had a
set of BAT files that changed AUTOEXEC.BAT to different versions for the
different configurations required.
> USB 2.0 Full-speed (12 Mbps)
> USB 2.0 Hi-speed (480 Mbps)
Screen resolutions were like that...
300 x 200 Medium Resolution (CGA four-color mode)
600 x 200 High Resolution (CGA monochrome graphics mode)
And then EGA and VGA came along and more superlatives had to be invented to describe the resolutions that were higher than high. Eventually everyone decided to just stick with the numbers, because it's actually easier to remember which resolution is which that way; when we say 800x600 and 1024x768 and 1280x1024, everyone knows which is the highest of those three resolutions.
Oh, and do you remember Full Color, High Color, and True Color? Fortunately these days we can just say 24bpp or 32bpp or whatever it is, which is much easier to keep straight.
Will USB/1394/whatever head in that direction, so that we can just talk about the port's (or the device's) speed, rather than quoting version numbers? Wouldn't that be nice?
Also, how about floppies:
360K Double-sided Double-density (5.25")
1.2M Double-sided High-density (5.25")
720K Double-sided Double-density (3.5")
1.4M Double-sided High-density (3.5")
There were also at one time single-sided and/or single-density 5.25" diskettes and drives. (And if you had double-sided diskettes and a single-sided drive, you could flip the diskette over and format the other side and use it separately, without interfering with the first side; people used to put a program on one side and data on the other, sometimes.) I don't know whether there were ever single-sided 3.5" diskettes.
> if you continue this "open up a terminal lamer" attitude you won't have
... that does not even deserve a response. I suppose the people who design microchips are doomed to extinction by next week, because regular people are not interested in examining their work.
> too much future in OS X or Win32
The other poster was perhaps not clear enough. I will try to explain this clearly so that you can understand it: some programs can never be useful to people who don't want to open up a terminal, just because of the inherently very technical nature of what the programs do. panpipes is exactly the sort of program that such a person would never be interested in using. All it does is make the computer crash. That's all it does. It doesn't do anything else. Furthermore, the whole point of writing it was so that people could analyze it and try to figure out how it does that, in terms of the technical under-the-hood things, the nuts and bolts of what it does to make the computer crash.
Suggesting that a program like that needs a GUI, just because end users don't like to open up terminals, is completely missing the point. This program does not do anything an end user would want to do. Ever. Under any circumstances. Making a GUI for it would be utterly and completely worthless, a total waste of time. People who analyze code that causes kernel panics are *NOT* the sort of people who need a GUI for everything and don't want to open a terminal. For crying out loud, they had to use gdb to analyze the thing -- gdb is really hardcore technical stuff, *way* beyond not having a GUI. Even most people who use the terminal on an hourly bases are not experienced using gdb (or any other lowlevel debugger like that). There is absolutely no way on earth that anyone who doesn't want to open up a terminal window and thinks all applications should have a GUI would have *any* interest *whatsoever* in participating in this sort of contest. If they thought they were interested, it is obviously because they saw the word "contest" and didn't read the explanation of what kind of contest it was, what it was actually about, or what they had to do to participate. If they had, they would NOT have been interested.
As for the other part of your post, suggesting that the other poster will somehow not have a future, just because he writes programs that are technical in nature
> the ping delays would be a serious drag.
I wouldn't think so. The moon is only, what, 280 thousand miles out from the Earth's surface? Give or take? Assuming your networking equipment is optical, you should be able to achieve reasonable ping times. Good enough that a sysadmin in Denver could ssh in and perform system maintenance in an emergency, I should think. Granted, the networking equipment would cost a little more than residential-consumer-grade WiFi stuff...
> Bush said he wants to use it as a stepping stone for Mars.
I don't know the context of this quote, but is it possible he meant a stepping stone in terms of experience, knowledge gained, et cetera? That seems more likely to me than a physical stepping stone in terms of launching ships. We can launch ships from here, but a lunar base would be within a few days' reach in an emergency; whereas, a Mars base would not be, so an argument could be made for using a lunar base first to work out the kinks in the base planning methodology.
We theoretically have the technology to go to Mars, yes. That doesn't make it an easy thing at all, nor cheap. We do *not* have the technology to establish a self-sustaining base there (not on the Moon either, though), so you're looking at a short-term purely scientific endeavor, one that's really *really* expensive and requires a whole new *kind* of astronaut, one who's willing to live in a confined space for *years* at a time with effectively *zero* possibility of returning to Earth on a shorter timeframe in an emergency. Possible candidates for the trip should be screened first for the ability to stay a couple of years at someplace like Vostok (or at least Amundsen-Scott), come back sane, and still want to go to Mars.
> Yes, just a big rock, chock full of raw materials we need for your
> trip to Mars
What specific material is on the Moon that we need in order to go to Mars?
> Once established, the Moon Base will py for itself countless times over.
Now you're completely out of your chair. The cost of maintaining it would absolutely dwarf any possible financial benefit. You know that the Moon has no atmosphere, right? No atmosphere? Have you got that? A base there cannot sustain itself; it needs continual supplies from Earth, not least of all food. That means regular supply trips, which have to deal with full Earth gravity, every little bit, for the whole time we maintain the base. The only reasons to go to the Moon or establish a base there are strictly scientific in nature. Economically, it's a big fat negative. Heck, even inland Antarctica, which is *way* less hostile an environment than the moon, is economically-speaking a completely worthless destination. The Moon is even worse. (So, for that matter, is Mars.)
That doesn't (necessarily) mean we shouldn't do it, but expecting it to pay for itself, much less anything else, is industrial-strength naive. It will not pay for itself. It will cost a fortune to build and a fortune every year to maintain.
> Ultimately it doesn't matter who owns it until there's some feasable way
> to get there.
Getting there isn't too hard. Okay, so it's not an afternoon jaunt, but it's
been done a handful of times, and that was without modern technology. The
hard part is figuring out how to derive substantial benefit from staying
there long term. Nobody's solved that one yet. Visiting the moon is an
interesting endeavor, but after a short stay, everybody seems to want to
come back to Earth. I guess it's not a tangible thing, just some kind of
nebulous psychological thing or something. People keep saying Earth has
"better atmosphere", whatever that means.
I just use grip with the "rip only" option (no encoding). Metadata are stored .wav postpended. Works for me. One nice thing about WAV is that it's
at the filesystem level: the author's name as a directory, and within that
the album title as a directory, and within that the track name as a filename,
with
lossless. Another is that it's supported by, you know, absolutely everything.
> This problem is that Linux is still largely a text interface if you want
:-)
> to tap into any of its power.
This is also true of Win32. And the shell is pathetically bad -- though the new one in Longhorn is supposed to be an earthshattering improvement, and I'm inclined to believe it.
I have a feeling you're confusing "easier to use" with "easier to *learn* to use", which is a different thing. The learning curve for Windows starts off less steeply than the one for *nix. (Note that I'm talking about the first part of the learning curve, the part that a newbie encounters; once you reach the power user level, the curves on both platforms change, but in different ways.) However, once you've learned how to use it, the command line (on both platforms) makes a lot of things WAY easier to do than the equivalent GUI way, and the *nix command line is (at this time) indisputably better than the NT command line, which in turn is universally regarded as superior to the Win9x comand line, which in turn puts the (classic) MacOS command-line completely to shame
> Go a week without using a shell. Easier said than done.
*shrug* Go a week without using a web browser. Go a week without using a text editor. What's your point, exactly? Prohibiting the use of whole classes of important standard application categories is *always* going to be rather limiting.
Sure, new users don't use some of the more advanced types of applications -- command line stuff, text editors, raytracers, compilers, image editors (except maybe those ones that come with digital cameras and have auto-remove-redeye buttons and no real features; whether those are really image editors is an open question), remote login facilities (not just the command-line ones; new users don't use RDC or VNC or cetera either; generally even mentioning the possibility of controlling a second computer remotely will overwhelm them) -- anything they don't have an immediate felt need for, because, being new, they still have plenty of basic stuff on their plate to learn, without adding more esoteric things to the cirruculum. Indeed, some people dabble with computers for several years before they even learn to use such essential things as the clipboard.
And yes, some users learn things (everything) more rapidly than others, mostly due to motivational factors, and to a lesser extent due to aptitude; and yes, some users reach a certain point and decide that's pretty much enough for them, at least for a good while (though often they change their mind several years later and want to learn how to do X, Y, or Z). Which is fine.
But none of that is specific to any particular platform.
> I have NEVER had that good of luck with any version of windows...
Luck is what you make of it. Windows does require more reboots, and of course it helps to be at least a Windows poweruser (just the same as it helps to be a Linux poweruser if you're installing that -- and don't get me started on installing BSD), and sometimes you will have to download drivers from the manufacturer, but despite all that the success rate for an experienced user can be very high, as almost all hardware is capable of running it (or vice versa, depending how you look at that) -- much higher than 78%. (Although, that number also seems low to me for modern Linux distros -- what are you installing, Debian stable?)
There are, of course, some things to know when installing a Microsoft OS. Number one is, get the data you want to keep off, fdisk the thing, and create new partitions from scratch. Always. Saves HOURS of headaches. (In a pinch, if you really HAVE to keep the partition in tact for data's sake, installing somewhere other than C:\WINDOWS can work as an interim solution, if you have enough extra drive space, but I recommend against it if it can possibly be avoided, since the cruft left around by this approach can really wreak havoc with application installations.) Number two, as soon as you have a decent working state, create a system restore point *immediately*, *before* you tweak anything, *especially* the network settings. But, there are not-so-dissimilar equivalents to these with the Linux distributions I have used also. Getting networking to work, and keeping it working while you set everything else up, is generally easier with Linux, but OTOH getting printers to work can be more of a pain. Really, they're quite comparable, albeit certainly not the same.
> Next week Jacob will be telling us how washing machines are great for
> keeping your clothes clean but not very good for making cups of tea.
The thing is, a lot of people over the years have tried to use PDF as a format for online distribution of content. It sucks for that. In many cases, HTML or even plain ASCII would have been a MUCH better choice.
If you're using PDF for distributing content that is no good unless printed (on a certain size and shape of paper), such as IRS tax forms, then that's a good use of PDF. If you're using PDF to deliver documents to a commercial printer, that's another good use of PDF. If you're using it to protect the precious layout of your white paper, the world would be better served if you stuck your precious layout in the nearest available unsavory bodily orifice and distributed the content as HTML or plain text.
> In what way is Adobe reader superior to Xpdf?
The release of xpdf I am comparing to is not the latest and greatest (it's not an app that I use all that often), so maybe this has been fixed, but last I knew xpdf does not handle transparent images very well, often giving them a black background when in fact it should be white; if there is text in front of the images, this can result in documents' being illegible. acroread also used to have this problem in about version 4, but it has been fixed for some time now.
> Now if only Adobe would bring Photoshop over as well...
What, and give the Gimp developers the ability to easily run the two side-by-side on their platform of choice, directly compare them, and so forth? I'm not sure the Photoshop team really wants that level of competition. Gimp is already fairly impressive in terms of functionality, but the Gimp developers as a rule don't have Photoshop to compare to, so the interfaces are quite different. (This can be construed as a good thing... certainly it's good for allowing the respective applications to retain their existing user bases, which is much to Adobe's advantage.)
It may be that the Gimp developers would not spend the cash on Photoshop even if they *could* run it on *nix, and so it could be that little or nothing would change. But I can quite readily imagine Adobe's not wanting to chance it.
> To most end users, a consistent look and feel, that works right out of the
...) are at _least_ starting to pick up basic color schemes from the system now, although really they ought to use the actual native widgets. But it only takes one $#@! Athena application, or something like xmms, to spoil the consistency.
> box, is really important.
Yeah, but Win32 has never had this, because it allows application developers to make their own widgets out of whole cloth, resulting in fruitcakey nonsense UIs like AIM, Winamp, et cetera, ad infinitum, ad nauseam, ad bedlam. Yes, X11 allows this also; fortunately, *most* of the OSS community these days is beginning to see the folly of this approach and use one of the leading GUI widget sets, mainly GTK and/or Qt, which can be themed to look highly similar or at least coordinate nicely, and some of the major apps with their own toolkits (OOo, Mozilla,
The thing is, it is utter hypocracy for Microsoft advocates to criticize "Linux" (by which they mean X11 or the distros) for this, because exactly the same thing is true, and to pretty much exactly the same extent, on Win32. Okay, so Win32 has one core widget set instead of two (GTK and Qt), but the end user does not need to know about the GTK/Qt distinction, if they are themed similarly.
For Microsoft, there's an obvious solution to this: deprecate the parts of the API that allow custom widgets, and in five years remove support for them. ISVs won't like this one bit, and the question is, which is more important: retaining ISV mindshare, or giving users a consistent look in every app? I tend to think the ISV mindshare is more important, but Microsoft does have this option to consider. In the OSS community, of course, there's no such option, because of the nature of the freedoms developers inherently have -- although distributions could refuse to *bundle* non-compliant apps. (But, no sane distribution would refuse to bundle, say, Emacs, which last I checked is still built against Xaw3D or some such ugliness, presumably because for some strange reason pretty scrollbars and tear-off menus don't seem to be a major show-stopping priority for Emacs developers. XMMS is also pretty major, but I think that one is sufficiently modular that they could adjust its GUI to comply, rather than exclude it. Command-line programs, of course, are exempt from GUI-oriented visual consistency guidelines; nobody expects cd to use a GTK2 file selection dialog, for instance; that way lies madness.)
Mandrake does put some of their own work into their distro. The *drake tools (printerdrake and so forth) really do speed up initial configuration -- especially for relative *nix newbies, but even for experienced users they are convenient.
Not that Ubuntu is bad, but that's another thread.
That's because ping is braindead. You won't have this problem if you construct the echo-request packets yourself. (Okay, so the packets do have to have the IP addresses in them -- but you can send them directly to a certain MAC address.)
> Why is it ridiculous to run stable applications at work?
Well, let's all just run PC-DOS 3.3 while we're at it.
Seriously. PC-DOS 3.3 was very stable; I never managed to find a bug in it, period, not even a little one. Then MS-DOS 5.0 came out, and it had all kinds of bugs. 6.0 was only a little better. 6.22 had most of the bugs ironed out, but it was still not as stable as 3.3. Let's all use 3.3, since it's stable.
Do you see the problem?
Installing Woody is like travelling back in time the better part of a decade. Just for starters, the installer is not something that could ever be recommended in a work setting: you have to use dselect. Manually. It makes NT4 look positively modern. And the installer is just the beginning. Once you get it installed, you find that any application released in the last couple of years won't run on it. You want to run an application that requires GTK2? Haha, you're joking, right? Heavens, you wouldn't want anything so bleeding edge as *that* on your system; after all, it's only three or four years old, so just get it out of your head.
There's stable, and then there's really stable, and then there's Just Plain Old. Woody, at this point, is too far into the last category.
There's nothing stopping people who don't need new software from continuing to use old software, such as Woody, or Potato, or TOPS20. But when your *latest* stable release has such thick cobwebs on it that people with seven years of Linux experience try to install it and immediately go "Yikes!", it's time to get a slightly more modern one out the door soon.
The marketing fluff is normal -- in fact, I barely noticed it, as out of long habbit I've trained myself to read through hype and pick out the relevant bits (at least, usually); it's the grammatical anomoly that I didn't expect to see coming out of Apple. They're usually relatively careful about that.