> This is Slashdot, home of the penguin I know. But you have to give MS some credit. They're doing better.
Yeah, wake me up when their UI is customizeable to any meaninful extent.
Say, for instance, just by way of example, that I want to put the task list, and only the task list, across the bottom edge of the screen, so that I can fit lots of open Windows on that list, and then I want all the *other* stuff, such as the start menu, quick launch, and system tray, to be along the left edge of the screen, about 32px or maybe 48px wide at most (not that great big honking Vista sidebar that takes up a quarter of the screen). Can I do that in Windows?
(No, don't say LiteStep. No. Just, no.)
Every release since 2.0, the Gnome panel gets fewer and fewer features. What I *really* want is the panel from Gnome 1.x, which was awesome. But of course that's not compatible with modern libraries.
But even though it keeps *losing* useful features, the latest and worst version of Gnome is *still* ahead of the panel in Windows, any version of Windows you care to name.
> Not to mention the fact that the US with all their might couldn't defeat the Vietcong.
The US didn't use most of its real might in that conflict, because of the risk of Soviet retaliation if we did. Militarily, the USSR was comparable in power to the US at the time, so we could only do what we could do without getting them directly involved. (We eventually won the cold war hands down, yes, but we did that by economic means, not military.)
In other words we lost (or conceded) the Vietnam conflict for external political reasons.
> It just seems like it'd be easier now to find out the spam mail servers and block everything that comes from them.
They migrate from IP address to IP address too often for that. Apparently some ISPs will just hand you a different Class A block every couple of weeks, no questions asked, as long as you're paying your bill. This has been common in APNIC space at least since the late nineties.
> Headline should read, IE8 May Be End of the Line for Internet Explorer _Engine_ .
Indeed. Although, as much work as they put in to making IE8 render so many things so much better than IE7, I have a hard time believing they'll abandon it so soon.
> I don't see any reason why ActiveX apps couldn't be sandboxed like anything else
I'll go you one better: I don't see any reason why IE9 needs to support ActiveX apps at all. ActiveX* has been deprecated for so many years, the only things that still use it are line-of-business intranet apps, most of which were never updated to work in IE7 anyhow, because they're things that just don't get updated that much. There's no way in the world they're going to be updated for IE9. They're not developed that actively. They're *used* actively, but that's not the same thing.
So to support them, you just roll out an IE6-on-Win98 VirtualPC image and Bob is your uncle. IE9 doesn't need to do anything at all.
*Above I was speaking of ActiveX controls other than common well-known plugins like Java and Flash. These plugins will be needed for IE9, of course, but if IE9 uses a different plugin interface other than ActiveX, the plugins will be produced for that interface, and sites that use Java or Flash or whatever will work. There's no problem there.
> People seem to forget that we have sold way more computers than people in > the world... no reason to replace them all to run IE9.
Bear in mind, the person who suggested this was talking about compatibility for ActiveX-based content. This is not something that normal users need to view normal websites.
ActiveX has been deprecated ever since Microsoft found out that when you connect your PC to the internet security *is* an issue after all. Automatic ActiveX download-and-install support has been turned off by default, except in the intranet zone, for years. I don't remember *exactly* when they fixed this; it might have been XP SP1 that did it, or it could've just been an IE6 security update at some point. In any event, it was a while ago, well before they started working on IE7. Since it's been turned off for so many years, there are no public websites of any significance that still use it. There was some whining at first, but when it became clear that Microsoft was going to actually listen to the security people on this one, the web developers fell in line and found other ways to do things.
So for almost everyone, ActiveX is not needed. There are some things that will be needed that are currently *implemented* as ActiveX controls, but those things don't *need* to be implemented as ActiveX controls. They can be implemented another way. Microsoft already did this with XMLHttpRequest, which as of IE7 is built in, rather than being an ActiveX control.
The other major thing is plugins like Java and Flash, but, again, those don't *need* to be ActiveX controls. Another technology could be used. Firefox uses a different plugin technology, for instance. Believe me, if Microsoft still has anywhere near its current market share when IE9 comes out, the plugin developers will make their plugins for *whatever* plugin infrastructure IE9 uses. Sun talks big about not needing Microsoft, but when it comes down to brass tacks they will do the hundred yard belly crawl over splintered glass in their birthday suits if necessary to make sure the Java plugin works in Internet Explorer. Adobe will be even more motivated to make Flash work, because they won't want to leave a door open for Silverlight to get a foot into. (Right now there are approximately three websites that use Silverlight, and two of them were created by Microsoft to promote the technology; Adobe will want to keep it that way.)
Whatever solution Microsoft develops for ActiveX compatibility, almost all of the world's computers WILL NOT NEED IT, so there's no reason to suppose that all of the world's computers would need to be replaced because of whatever performance demands said solution might impose.
The only reason ActiveX would be needed at all, by anyone, is because there are some third-party line-of-business applications out there that rely on it. We're talking here about intranet stuff. This kind of software tends to develop at a glacial pace and then be upgraded even more slowly. Where I work, we were using DOS-based software until August of 2005, no fooling. So stuff that only works in IE6 will, as you can imagine, probably be around for a while. That's why the article says it will be a sticking point.
Honestly, though, such software would probably run just fine in emulation on today's hardware. Just roll out an IE6 VirtualPC image and Bob is very likely your uncle. I really don't see why IE9 needs ActiveX support at all.
Just make the sign-up form ask a question that all the locals can easily answer.
If I were doing one for Galion, and non-locals or bots were making a nuisance of themselves, I'd probably ask something like "The haunted house that the Jaycees run every October is on the corner of which two streets?" or "What's the name of the drug store on the square?" or perhaps "Who teaches US Government at Galion High School?" That would let in people who used to live here or have family here even if they aren't located here currently, but on the whole that's probably a good thing.
Of course, it's not worth messing with this at all unless you are having an actual problem. In most cases I would not expect any significant number of non-locals to find or bother molesting a site of purely local interest.
> The conclusions: Windows 7 is, overall, faster than both Vista and XP
Thanks, I'm going to have to spend the next hour and a half winding the needle on my bogometer back around to zero.
Every version of Windows is always said to be overall faster than the previous. One of the selling points for Windows 95 was that it would make your computer faster (as compared, presumably, to DOS 6 and Windows 3.11). Windows 98 was faster than 95. Windows 2000 was faster again, and XP was faster than that.
Except, if you do a side-by-side comparison on identical hardware, it's extremely obvious that in fact exactly the opposite is true. If you run Windows 95 on a 233 MHz system with 64MB of RAM, it performs well. Try that with Seven!
Vista was *theoretically* supposed to be faster than XP, except nobody believes that because it's system requirements are SO much higher, mostly because of the large number of years that passed while it was being developed. Seven will be more similar to Vista than Vista was to XP, because not as many years have passed and not as many changes were made. Nonetheless, it's officially going to be faster than Vista and faster than XP, but I'm pretty confident that if you run it side-by-side with XP on identical hardware, XP will come out faster for most tasks.
Although, it wouldn't be at all hard to beat XP at extracting large.zip files. I don't know how the Windows Explorer team managed to make that particular task so ridiculously slow in XP. Info-zip can actually do -9 compression in less time than it'll take Windows just to extract it. So I suppose Seven could actually beat XP at that particular task, and they'll probably find a couple of other corner cases to bookmark, probably involving new kinds of hardware acceleration that normal applications don't use.
> > Why should synchronous writes be the default? > Because it's what the user wants and expects
I suppose I should elaborate on this a bit...
Old Macs used to allow the user to drag a diskette icon to the trash and eject it even if there were files on that diskette that were currently open. Naturally, when you then tried to do anything at all with the windows in question, including close them, the Mac would ask for the disk back, because it wanted to cleanly close the file, write any pending changes, and so on. Well, guess what? The disk is gone. Why do you think the user ejected it in the first place? They were leaving the computer lab and they wanted to take their disk with them. The disk has left the building. You may as well ask for a billion dollars as ask for the disk back.
Fast forward to 2002. Windows XP has built-in support for creating data CDs, if you've got a CD-R drive. You just open My Computer, and you drag the files you want to put on the data CD, and you drop them on the little CD icon in the My Computer window. And then, if you're the sort of person who understands the relevance of the number 9660 to this workflow and knows that the phrase "random access" does *not* apply, you double-click the CD icon to get a window with the "write these files to disk" option in the sidebar, which you click. But normal users don't do this. Normal users drag the files onto the CD icon and drop them there, and then when the operation appears to have completed they eject the disk, take it, and leave. This is why Windows XP computers always have "files waiting to be written to the CD". Later the users are absolutely mystified that the files they put on the disk are not there. Did they bring the wrong disk? They *thought* this was the right one...
And while we're at it, we may as well talk about USB mass storage devices, because, again, as soon as the operation appears to complete, they yank the thing out of the port immediately. Always. "Remove safely"? I've NEVER seen a normal user bother with that. Ever. (*I* do it, but I have also been known to compile Emacs from source because binaries were not yet available of the version I wanted, so obviously I don't count as a normal user.)
Users expect the data to be on the disk when the save or copy or whatever appears to complete. This is ALWAYS what they expect. It's always what they WANT. Under no circumstances would they ever prefer for the actual write to happen later in the background for improved performance.
Because it's what the user wants and expects 99.9987849782% of the time, and almost no applications bother with it every time they should.
Of course, Unix systems generally have a sync mount option, which forces *all* writes to be done synchronously. But then that includes stuff like the cron job that updates locatedb, which SHOULD be asynchronous.
Personally, I still consider write cacheing to be a fundamentally bad idea. The process doing the writing should block until the data is ON THE DISK, physically (or until an error comes back saying it can't be written).
Other applications, on the other hand, should be able to continue about their merry way as if nothing were happening. This is where most current operating systems completely fall down, IMO. If a background process is flogging the disk, the whole system becomes unresponsive. That shouldn't happen. Only the application that's doing all that heavy disk work should be slow. (Frankly, I don't care how slow that background process gets; let the thing that updates locatedb run for a month and a half, so what?) Other processes should still get their normal time slices AND should have access to the disk during that time.
The other thing is that the window manager should have a way (a privileged way not available to other programs) to inform the kernel which process has keyboard focus at the moment, because that process should get a significantly larger portion of the available CPU time (if it needs it) than anything in the background.
> And if you happen to be running an Alpha, you can still run FreeBSD 6.3
What if I have a Vax? What version of BSD can I get that will run on that?
(This is largely theoretical, although I *do* have a MicroVax 3100-40 sitting around doing nothing in particular...)
Also, you forgot to mention ARM processors, which are pretty common, and 8-bit and 16-bit x86 systems, which used to be pretty common, pre-PowerPC Macs (6800 processor), and the Z80, among other things.
> To the degree that Norton does not exist except as a legal construct by state and federal law
That's absurd. Norton exists in the physical world. They have tangible assets (buildings, equipment,...).
Additionally, state and federal law also allow plenty of other things that are, nonetheless, not part of the government. 501(c)3 non-profit organizations, including churches, have a legal existence granted by the law, even though of course churches have existed, de facto, for centuries before said laws were written. Come to think of it, commercial private-sector entities have also existed, de facto, for centuries before the current laws under which they exist today were ever written. The law allows labor unions to exist, but they are not part of the government. The law (indeed, the constitution, ostensibly the supreme law of the land) allows citizens to assemble peaceably in an organized fashion to protest government policies, and yet, we do not consider such protests to be actions of the government. The "first sale" doctrine -- a legal concept bound up in federal law in the US -- allows me, as a private individual, to resell an item that I have purchased, and yet, by doing so, I am not therefore part of the government.
> Examples abound-- they cannot refuse to sell you products or hire you because of race or gender.
That's because there are specific laws that say they can't.
> Similarly, they don't have to create a forum, but they have* to be fair in letting people post there.
No. Freedom of the press has always belonged to whoever owns the press. The newspaper does not have to publish your letter to the editor, or your press release. They can pick and choose and publish whichever ones they like, because it's their paper, and their printing press.
Book publishers do not have to publish every book sent to them. If they did, they'd all go out of business, because they all get approximately five hundred bajillion manuscripts sent to them day, most of which are completely unpublishable drivel. They pick and choose and publish the ones they think will actually sell, and the decision of whether to publish or not publish a specific book is entirely the purview of the publisher (and, actually, specific people working for the publisher, who do screen manuscripts for a living, a job I can assure you I do NOT want, no matter how much it pays). If Random House turns down your manuscript, you are free, as the author, to take it to Penguin and try to get *them* interested in publishing it, and if they turn you down flat you can go to another publisher, and another, and another -- and authors do exactly this. And when you get fed up with all that, you can go to a print-on-demand publisher and pay them up front for five hundred copies, which you can cart around to every bookstore and library you can find and try to interest them in it -- good luck with that, because they don't have shelf space for stuff nobody will want to read, which as a general rule tends to include most self-published material.
But, you know, that would be your problem. The book stores and libraries don't have to put your book on their shelves. Because, you know the shelves are THEIR shelves, not yours. And the publishers don't have to print your book on their press, because it's THEIR press, not yours. You can go out and get your OWN press, if you can afford it, or you can try to *build* your own, out of corrugated cardboard and popsicle sticks if you want -- and if you can get your handmade press to work, Slashdot will probably run a story on it, because that's the kind of story they like to feature; but, of course, that's up to the Slashdot editors.
Similarly, Norton can publish on their website whatever they want (as long as it doesn't break any laws, e.g., it's not outright fraudlent and it's not child porn and they have permission from the copyright holder and so on). And they can *NOT* publish anything they *don't* want to publish. Because, you know,
Just give some astronauts a whole lot of iced tea to drink, in those little bubble pouches, then fly them in a slightly higher orbit, and let them have contests to knock the debris out of orbit!
> By their argument, `cd/; rm -rf.' still ought to work. Sigh.
Actually, I think their argument is that / and . and.. should all be treated (equally) specially. I'm not sure I buy that, but the whole thing is kind of silly anyhow. I mean, really, who entrusts root access, on any system that matters, to somebody who thinks combining -f with -r is a good idea? Seriously, why is this even an issue? It's not like the combination of recursion, coersion, the root directory, *and* doing the whole operation as root is something you can put together accidentally when you're trying to do something totally innocuous that wouldn't generally require caution otherwise. There are at least three different things going on there that *individually* should be treated with respect and seriousness. If you manage to combine all of them without extreme caution, it's because you're fundamentally a careless person.
The horror stories you hear about people "accidentally" messing up a production system this way always involve combining several inherently unsafe behaviors, and then inadvertently (or, a cynic might suggest, "inadvertently" on purpose) getting just one more unsafe thing into the mix. For instance, the user was *trying* to do rm -rf/foo as root on a production system, and he hit the enter key prematurely. Or the user was trying to do rm -rf / as root, for kicks, on a system that was about to be decommissioned, but he got the wrong terminal window that was shelled into a more important system. Meh. If the sysadmin is that much of a clown, he *WILL* find a way to hose the system sooner or later, no matter how much rm is neutered.
I mean, burning magnesium reacts violently with nitrogen. Does this mean nitrogen gas -- which makes up more than half of our atmosphere -- is an unsafe and reactive chemical? Personally, I think it means burning magnesium should be handled with caution and by qualified people. In terms of rm, I think the -r and -f arguments are both fairly serious things that should be handled with caution by qualified people. And then there's root access to a production system, which *definitely* should not be handed out like candy to careless idiots. Putting all three together... you're playing with fire in a magnesium warehouse. If you're careless and burn the place down, it's not fair to blame the atmosphere.
I've seen some of the HTML these tools (Frontweaver, Dreampage, HotMetal, etc) produce, and I Do Not Want It.
I use Emacs and w3schools, and my HTML is clean, scalable, efficient, reasonably accessible, and very maintainable, and honestly I don't spend that much time on it. HTML is, fundamentally, very easy, once you know what you're doing.
In terms of keeping all the pages on a site updated with side-wide changes, I mostly use a combination of keyboard macros, custom elisp, Perl, regular expressions, chewing gum, and bailing wire. But it works, and it works the way I *want* it to work.
As far as Drupal, though, I thought that was a CMS. Do people really try to use it as an HTML editor? Ugh.
I mean, fundamentally, if the fed funds rate is 1%, and somebody offers you a AAA-rated safe investment that pays 4% per annum, shouldn't your brain be screaming "TOO GOOD TO BE TRUE" so hard it blows steam out your ears? You don't have to understand what's wrong with it. You don't have to analyze it in depth. All you need to know is, it's paying out quadruple the fed funds rate in interest, and there's GOTTA be a reason.
Oh, dear. Your patented method for trademarking a copyright violates my intellectual property rights under the PRO-IP act, the DMCA, and also the Berne Convention. I'm afraid I'm going to have to ask you to cease and desist all use of the method covered by your copyright trademark patent.
Obviously, if you've got two things orbiting together, the larger one isn't going to be classed as a satellite of the smaller (although, under standard Newtonian physics, the relationship is actually mutual, even with a quite large disparity in size).
There's some question about where to draw the line between twin-planets that aren't quite exactly the same size, versus a planet with a fairly large moon, but there are various ways to write a definition for that that doesn't rely on arbitrary numbers. (For instance, you could say that if the center of mass of the combined system lays outside of either body they are companions, and otherwise the smaller object is a satellite of the larger. This has the consequence of making Pluto and Charon companion minor planets, but we've already taken the big step of demoting Pluto to minor-planet status in the first place.)
The harder question is where to draw the line between a small moon and a mere particle/rock/meteor/asteroid that happens to be in orbit around a much larger body. Personally, I'd vote for gravitationally-self-rounded, because I never felt that the asteroids around Mars were large enough to be proper moons anyway. This would also greatly simplify and reduce the moon counts for the gas giant planets. Smaller orbiting objects would still be natural satellites, of course; the definition would only apply to the word "moon". But the astronomical community has never reached any significant degree of consensus on this issue.
> Firms who discriminate against people who aren't ashamed of their > life and like to talk openly about it will wind up full of drones
I'm thinking they'll wind up full of people who use pseudonyms and/or whose names are along the lines of Jennifer Miller or Robert Anderson.
But yeah, if you really believe that this practice does not help the firm get better employees, and may even hinder them from doing so, then it logically follows that there's no point in worrying much about it, because employers who don't engage in the practice will do well and compete well against the ones who do, and so there will always be jobs available for people who don't worry about the issue.
I really am not sure I see what the big deal is. If I search for myself on Google (by my name, not the pseudonym), the first result is some dude's facebook profile who clearly isn't me, but after that you get my contributions to the Open Clip Art Library, my CPAN profile, the reviews I've written on Amazon,...
As long as they only changed the *default* number of blinks from three to seven, so that you can still configure it down to zero, I don't see a problem there. Because anybody who would be annoyed by seven was probably annoyed by three (the default since WinXP) anyway.
> They don't give you the choice of when to reboot after an update
Yeah, you' think they'd give the system administrator the option to configure a delay into that, so that it wouldn't bug the users for the first N hours, on the assumption that by then the system will be shut down anyway. A lot of office PCs routinely get shut down every night, so if the sysadmin set the upgrade-reboot-delay (or whatever they want to call it) to twenty hours, the logged-in user would never be bothered, and upgrades would still all get installed in reasonably a timely fashion. And if for some reason a user *doesn't* shut down the PC at night, it'll then force the reboot when the delay does expire.
> Now that I think about it, why does an OS need to reboot after > every installation / removal of an application or update?
Because NTFS doesn't have inodes to abstract away the stored information from the directory entry. As a result, it's not safe to update a file that is currently in use. For some kinds of files, the only way for it to not be in use is if the system isn't fully booted up.
Microsoft has needed for several years now to introduce a successor to NTFS, for a variety of reasons. (Windows should continue to support both, just like it still supports FAT, although after a version or two the main filesystem on which the OS is installed might be required to be the new fs, or certainly it would at least be strongly encouraged.) However, their original concept for what that would mean was a database-like filesystem built on top of NTFS, and when it eventually became clear that that wasn't actually the way to go, they pulled that feature from Longhorn/Vista and sort of set the issue aside for a while, because Vista was already overdue and they needed to get it out the door, with or without a new filesystem. Seven is a minor upgrade built mostly on Vista technology, so a new filesystem isn't in the cards there either. Perhaps for blackcomb (or whatever they're going to call the release after Seven)? Who knows. But sooner or later they will need to do it.
There are several things they really ought to build into their next filesystem. The aforementioned inodes are one. Versioning is another. Optional per-user automatic encryption, tied to login, is another.
> #7: Programs COULD install themselves to your gnome-panel. Wouldn't be that hard.
It would be highly undesirable for a lot of Gnome users, though. We tend to want things they way we want them, set them up that way, and expect them to STAY THAT WAY unless and until we change them. We don't want software making changes to our setup without our permission. If we wanted that, we'd probably use Windows.
> #15: Have you seen the new GNOME network manager?
No, I'm still on etch here, so if Gnome has a *new* feature, I haven't seen it yet.
Come to think of it, though, aren't new features against Gnome's religion these days?
> I finally read the manpage for interfaces and just use ifupdown without the network manager now.
I set up the network settings the way I want them when I installed, and haven't needed to change them since.
> #23 is EVIL!
I never liked Windows Media Player. It always seemed to be too focused on disrupting the user's experience with undesirable features nobody would ever want, things that actively got in the way of playing music (or video, or whatever you were trying to play). Instead of potentially-useful features, like looking up the CD you just inserted in the cddb and filling in the track titles, it instead chose to do stupid annoying stuff, like hide the tracklist and instead show stupid screen-saver-esque effects, advertisements, and other malarkey. Also, the interface appeared to have been designed by a first-year art student with no sense of aesthetics and even less concern for function. When I had a version of Windows that came with WMP, I initially replaced WMP with the old equivalent applications that had come with Windows 95, because they were significantly better. Later I downloaded some freeware media player or another. These days I mostly use xmms. It does what I want and mostly stays out of my way otherwise. If I were using Windows, I'd probably be using a different media player than xmms, but I can tell you this for free: it wouldn't be Windows Media Player.
> #24 is WEIRD! You couldn't resume playback after coming back from sleep mode? > That seems like a bug fix, not a feature enhancement.
> I sure hope that one day we get some decent open video drivers > for ATI, because that's all that's keeping me from switching.
I realize this won't help you right *now* (unless you have money to spend on the switch, which is not usually how it works for most of us), but you can always keep this in mind next time you upgrade your hardware, and consider getting a different brand of video card.
Personally, I won't buy ATI or nVidia cards. I don't have money to spend on overpriced wonky hardware that's only good for playing games. Depending on the needs for any given system, I either go with cheap onboard video (typically, Intel chipset), which is plenty good enough for systems with only very basic video needs (servers, and systems that are only used for the web and email and office documents), or a Matrox video card if it actually matters (e.g., for a workstation).
> It is not human nature, to repeat a failed action in the hopes that the results change...
That depends on the human. Intelligent people try to figure out why it didn't work and do something about that. But a double-digit percentage of the population just plain isn't that bright.
You know what would be really cool, and make network administration a lot more fun? If the print spool would do a checksum or something and, if the print job is exactly the same as the last one, sound an audible alarm and refuse ("Error: Duplicate Print Job. This document has already just been sent to the print spool."), or at least require administrator approval.
Slashdot Mirror
> This is Slashdot, home of the penguin I know. But you have to give MS some credit. They're doing better.
Yeah, wake me up when their UI is customizeable to any meaninful extent.
Say, for instance, just by way of example, that I want to put the task list, and only the task list, across the bottom edge of the screen, so that I can fit lots of open Windows on that list, and then I want all the *other* stuff, such as the start menu, quick launch, and system tray, to be along the left edge of the screen, about 32px or maybe 48px wide at most (not that great big honking Vista sidebar that takes up a quarter of the screen). Can I do that in Windows?
(No, don't say LiteStep. No. Just, no.)
Every release since 2.0, the Gnome panel gets fewer and fewer features. What I *really* want is the panel from Gnome 1.x, which was awesome. But of course that's not compatible with modern libraries.
But even though it keeps *losing* useful features, the latest and worst version of Gnome is *still* ahead of the panel in Windows, any version of Windows you care to name.
> Not to mention the fact that the US with all their might couldn't defeat the Vietcong.
The US didn't use most of its real might in that conflict, because of the risk of Soviet retaliation if we did. Militarily, the USSR was comparable in power to the US at the time, so we could only do what we could do without getting them directly involved. (We eventually won the cold war hands down, yes, but we did that by economic means, not military.)
In other words we lost (or conceded) the Vietnam conflict for external political reasons.
> It just seems like it'd be easier now to find out the spam mail servers and block everything that comes from them.
They migrate from IP address to IP address too often for that. Apparently some ISPs will just hand you a different Class A block every couple of weeks, no questions asked, as long as you're paying your bill. This has been common in APNIC space at least since the late nineties.
And then there are the botnets.
> Headline should read, IE8 May Be End of the Line for Internet Explorer _Engine_ .
Indeed. Although, as much work as they put in to making IE8 render so many things so much better than IE7, I have a hard time believing they'll abandon it so soon.
> I don't see any reason why ActiveX apps couldn't be sandboxed like anything else
I'll go you one better: I don't see any reason why IE9 needs to support ActiveX apps at all. ActiveX* has been deprecated for so many years, the only things that still use it are line-of-business intranet apps, most of which were never updated to work in IE7 anyhow, because they're things that just don't get updated that much. There's no way in the world they're going to be updated for IE9. They're not developed that actively. They're *used* actively, but that's not the same thing.
So to support them, you just roll out an IE6-on-Win98 VirtualPC image and Bob is your uncle. IE9 doesn't need to do anything at all.
*Above I was speaking of ActiveX controls other than common well-known plugins like Java and Flash. These plugins will be needed for IE9, of course, but if IE9 uses a different plugin interface other than ActiveX, the plugins will be produced for that interface, and sites that use Java or Flash or whatever will work. There's no problem there.
> People seem to forget that we have sold way more computers than people in
> the world... no reason to replace them all to run IE9.
Bear in mind, the person who suggested this was talking about compatibility for ActiveX-based content. This is not something that normal users need to view normal websites.
ActiveX has been deprecated ever since Microsoft found out that when you connect your PC to the internet security *is* an issue after all. Automatic ActiveX download-and-install support has been turned off by default, except in the intranet zone, for years. I don't remember *exactly* when they fixed this; it might have been XP SP1 that did it, or it could've just been an IE6 security update at some point. In any event, it was a while ago, well before they started working on IE7. Since it's been turned off for so many years, there are no public websites of any significance that still use it. There was some whining at first, but when it became clear that Microsoft was going to actually listen to the security people on this one, the web developers fell in line and found other ways to do things.
So for almost everyone, ActiveX is not needed. There are some things that will be needed that are currently *implemented* as ActiveX controls, but those things don't *need* to be implemented as ActiveX controls. They can be implemented another way. Microsoft already did this with XMLHttpRequest, which as of IE7 is built in, rather than being an ActiveX control.
The other major thing is plugins like Java and Flash, but, again, those don't *need* to be ActiveX controls. Another technology could be used. Firefox uses a different plugin technology, for instance. Believe me, if Microsoft still has anywhere near its current market share when IE9 comes out, the plugin developers will make their plugins for *whatever* plugin infrastructure IE9 uses. Sun talks big about not needing Microsoft, but when it comes down to brass tacks they will do the hundred yard belly crawl over splintered glass in their birthday suits if necessary to make sure the Java plugin works in Internet Explorer. Adobe will be even more motivated to make Flash work, because they won't want to leave a door open for Silverlight to get a foot into. (Right now there are approximately three websites that use Silverlight, and two of them were created by Microsoft to promote the technology; Adobe will want to keep it that way.)
Whatever solution Microsoft develops for ActiveX compatibility, almost all of the world's computers WILL NOT NEED IT, so there's no reason to suppose that all of the world's computers would need to be replaced because of whatever performance demands said solution might impose.
The only reason ActiveX would be needed at all, by anyone, is because there are some third-party line-of-business applications out there that rely on it. We're talking here about intranet stuff. This kind of software tends to develop at a glacial pace and then be upgraded even more slowly. Where I work, we were using DOS-based software until August of 2005, no fooling. So stuff that only works in IE6 will, as you can imagine, probably be around for a while. That's why the article says it will be a sticking point.
Honestly, though, such software would probably run just fine in emulation on today's hardware. Just roll out an IE6 VirtualPC image and Bob is very likely your uncle. I really don't see why IE9 needs ActiveX support at all.
Just make the sign-up form ask a question that all the locals can easily answer.
If I were doing one for Galion, and non-locals or bots were making a nuisance of themselves, I'd probably ask something like "The haunted house that the Jaycees run every October is on the corner of which two streets?" or "What's the name of the drug store on the square?" or perhaps "Who teaches US Government at Galion High School?" That would let in people who used to live here or have family here even if they aren't located here currently, but on the whole that's probably a good thing.
Of course, it's not worth messing with this at all unless you are having an actual problem. In most cases I would not expect any significant number of non-locals to find or bother molesting a site of purely local interest.
> The conclusions: Windows 7 is, overall, faster than both Vista and XP
.zip files. I don't know how the Windows Explorer team managed to make that particular task so ridiculously slow in XP. Info-zip can actually do -9 compression in less time than it'll take Windows just to extract it. So I suppose Seven could actually beat XP at that particular task, and they'll probably find a couple of other corner cases to bookmark, probably involving new kinds of hardware acceleration that normal applications don't use.
Thanks, I'm going to have to spend the next hour and a half winding the needle on my bogometer back around to zero.
Every version of Windows is always said to be overall faster than the previous. One of the selling points for Windows 95 was that it would make your computer faster (as compared, presumably, to DOS 6 and Windows 3.11). Windows 98 was faster than 95. Windows 2000 was faster again, and XP was faster than that.
Except, if you do a side-by-side comparison on identical hardware, it's extremely obvious that in fact exactly the opposite is true. If you run Windows 95 on a 233 MHz system with 64MB of RAM, it performs well. Try that with Seven!
Vista was *theoretically* supposed to be faster than XP, except nobody believes that because it's system requirements are SO much higher, mostly because of the large number of years that passed while it was being developed. Seven will be more similar to Vista than Vista was to XP, because not as many years have passed and not as many changes were made. Nonetheless, it's officially going to be faster than Vista and faster than XP, but I'm pretty confident that if you run it side-by-side with XP on identical hardware, XP will come out faster for most tasks.
Although, it wouldn't be at all hard to beat XP at extracting large
> > Why should synchronous writes be the default?
> Because it's what the user wants and expects
I suppose I should elaborate on this a bit...
Old Macs used to allow the user to drag a diskette icon to the trash and eject it even if there were files on that diskette that were currently open. Naturally, when you then tried to do anything at all with the windows in question, including close them, the Mac would ask for the disk back, because it wanted to cleanly close the file, write any pending changes, and so on. Well, guess what? The disk is gone. Why do you think the user ejected it in the first place? They were leaving the computer lab and they wanted to take their disk with them. The disk has left the building. You may as well ask for a billion dollars as ask for the disk back.
Fast forward to 2002. Windows XP has built-in support for creating data CDs, if you've got a CD-R drive. You just open My Computer, and you drag the files you want to put on the data CD, and you drop them on the little CD icon in the My Computer window. And then, if you're the sort of person who understands the relevance of the number 9660 to this workflow and knows that the phrase "random access" does *not* apply, you double-click the CD icon to get a window with the "write these files to disk" option in the sidebar, which you click. But normal users don't do this. Normal users drag the files onto the CD icon and drop them there, and then when the operation appears to have completed they eject the disk, take it, and leave. This is why Windows XP computers always have "files waiting to be written to the CD". Later the users are absolutely mystified that the files they put on the disk are not there. Did they bring the wrong disk? They *thought* this was the right one...
And while we're at it, we may as well talk about USB mass storage devices, because, again, as soon as the operation appears to complete, they yank the thing out of the port immediately. Always. "Remove safely"? I've NEVER seen a normal user bother with that. Ever. (*I* do it, but I have also been known to compile Emacs from source because binaries were not yet available of the version I wanted, so obviously I don't count as a normal user.)
Users expect the data to be on the disk when the save or copy or whatever appears to complete. This is ALWAYS what they expect. It's always what they WANT. Under no circumstances would they ever prefer for the actual write to happen later in the background for improved performance.
> Why should synchronous writes be the default?
Because it's what the user wants and expects 99.9987849782% of the time, and almost no applications bother with it every time they should.
Of course, Unix systems generally have a sync mount option, which forces *all* writes to be done synchronously. But then that includes stuff like the cron job that updates locatedb, which SHOULD be asynchronous.
Personally, I still consider write cacheing to be a fundamentally bad idea. The process doing the writing should block until the data is ON THE DISK, physically (or until an error comes back saying it can't be written).
Other applications, on the other hand, should be able to continue about their merry way as if nothing were happening. This is where most current operating systems completely fall down, IMO. If a background process is flogging the disk, the whole system becomes unresponsive. That shouldn't happen. Only the application that's doing all that heavy disk work should be slow. (Frankly, I don't care how slow that background process gets; let the thing that updates locatedb run for a month and a half, so what?) Other processes should still get their normal time slices AND should have access to the disk during that time.
The other thing is that the window manager should have a way (a privileged way not available to other programs) to inform the kernel which process has keyboard focus at the moment, because that process should get a significantly larger portion of the available CPU time (if it needs it) than anything in the background.
> And if you happen to be running an Alpha, you can still run FreeBSD 6.3
What if I have a Vax? What version of BSD can I get that will run on that?
(This is largely theoretical, although I *do* have a MicroVax 3100-40 sitting around doing nothing in particular...)
Also, you forgot to mention ARM processors, which are pretty common, and 8-bit and 16-bit x86 systems, which used to be pretty common, pre-PowerPC Macs (6800 processor), and the Z80, among other things.
> To the degree that Norton does not exist except as a legal construct by state and federal law
...).
That's absurd. Norton exists in the physical world. They have tangible assets (buildings, equipment,
Additionally, state and federal law also allow plenty of other things that are, nonetheless, not part of the government. 501(c)3 non-profit organizations, including churches, have a legal existence granted by the law, even though of course churches have existed, de facto, for centuries before said laws were written. Come to think of it, commercial private-sector entities have also existed, de facto, for centuries before the current laws under which they exist today were ever written. The law allows labor unions to exist, but they are not part of the government. The law (indeed, the constitution, ostensibly the supreme law of the land) allows citizens to assemble peaceably in an organized fashion to protest government policies, and yet, we do not consider such protests to be actions of the government. The "first sale" doctrine -- a legal concept bound up in federal law in the US -- allows me, as a private individual, to resell an item that I have purchased, and yet, by doing so, I am not therefore part of the government.
> Examples abound-- they cannot refuse to sell you products or hire you because of race or gender.
That's because there are specific laws that say they can't.
> Similarly, they don't have to create a forum, but they have* to be fair in letting people post there.
No. Freedom of the press has always belonged to whoever owns the press. The newspaper does not have to publish your letter to the editor, or your press release. They can pick and choose and publish whichever ones they like, because it's their paper, and their printing press.
Book publishers do not have to publish every book sent to them. If they did, they'd all go out of business, because they all get approximately five hundred bajillion manuscripts sent to them day, most of which are completely unpublishable drivel. They pick and choose and publish the ones they think will actually sell, and the decision of whether to publish or not publish a specific book is entirely the purview of the publisher (and, actually, specific people working for the publisher, who do screen manuscripts for a living, a job I can assure you I do NOT want, no matter how much it pays). If Random House turns down your manuscript, you are free, as the author, to take it to Penguin and try to get *them* interested in publishing it, and if they turn you down flat you can go to another publisher, and another, and another -- and authors do exactly this. And when you get fed up with all that, you can go to a print-on-demand publisher and pay them up front for five hundred copies, which you can cart around to every bookstore and library you can find and try to interest them in it -- good luck with that, because they don't have shelf space for stuff nobody will want to read, which as a general rule tends to include most self-published material.
But, you know, that would be your problem. The book stores and libraries don't have to put your book on their shelves. Because, you know the shelves are THEIR shelves, not yours. And the publishers don't have to print your book on their press, because it's THEIR press, not yours. You can go out and get your OWN press, if you can afford it, or you can try to *build* your own, out of corrugated cardboard and popsicle sticks if you want -- and if you can get your handmade press to work, Slashdot will probably run a story on it, because that's the kind of story they like to feature; but, of course, that's up to the Slashdot editors.
Similarly, Norton can publish on their website whatever they want (as long as it doesn't break any laws, e.g., it's not outright fraudlent and it's not child porn and they have permission from the copyright holder and so on). And they can *NOT* publish anything they *don't* want to publish. Because, you know,
Just give some astronauts a whole lot of iced tea to drink, in those little bubble pouches, then fly them in a slightly higher orbit, and let them have contests to knock the debris out of orbit!
> By their argument, `cd /; rm -rf .' still ought to work. Sigh.
.. should all be treated (equally) specially. I'm not sure I buy that, but the whole thing is kind of silly anyhow. I mean, really, who entrusts root access, on any system that matters, to somebody who thinks combining -f with -r is a good idea? Seriously, why is this even an issue? It's not like the combination of recursion, coersion, the root directory, *and* doing the whole operation as root is something you can put together accidentally when you're trying to do something totally innocuous that wouldn't generally require caution otherwise. There are at least three different things going on there that *individually* should be treated with respect and seriousness. If you manage to combine all of them without extreme caution, it's because you're fundamentally a careless person.
/foo as root on a production system, and he hit the enter key prematurely. Or the user was trying to do rm -rf / as root, for kicks, on a system that was about to be decommissioned, but he got the wrong terminal window that was shelled into a more important system. Meh. If the sysadmin is that much of a clown, he *WILL* find a way to hose the system sooner or later, no matter how much rm is neutered.
Actually, I think their argument is that / and . and
The horror stories you hear about people "accidentally" messing up a production system this way always involve combining several inherently unsafe behaviors, and then inadvertently (or, a cynic might suggest, "inadvertently" on purpose) getting just one more unsafe thing into the mix. For instance, the user was *trying* to do rm -rf
I mean, burning magnesium reacts violently with nitrogen. Does this mean nitrogen gas -- which makes up more than half of our atmosphere -- is an unsafe and reactive chemical? Personally, I think it means burning magnesium should be handled with caution and by qualified people. In terms of rm, I think the -r and -f arguments are both fairly serious things that should be handled with caution by qualified people. And then there's root access to a production system, which *definitely* should not be handed out like candy to careless idiots. Putting all three together... you're playing with fire in a magnesium warehouse. If you're careless and burn the place down, it's not fair to blame the atmosphere.
> But at least you get the source to rm(1) so you can fix that bug - or write your own, it's not that hard.
You mean like this?
perl -e ' sub e { my $file = $_; -d $file ? ((map { e $_ } <$file/*>), $file) : $file } unlink for e "/"; '
I've seen some of the HTML these tools (Frontweaver, Dreampage, HotMetal, etc) produce, and I Do Not Want It.
I use Emacs and w3schools, and my HTML is clean, scalable, efficient, reasonably accessible, and very maintainable, and honestly I don't spend that much time on it. HTML is, fundamentally, very easy, once you know what you're doing.
In terms of keeping all the pages on a site updated with side-wide changes, I mostly use a combination of keyboard macros, custom elisp, Perl, regular expressions, chewing gum, and bailing wire. But it works, and it works the way I *want* it to work.
As far as Drupal, though, I thought that was a CMS. Do people really try to use it as an HTML editor? Ugh.
> G+R+E+E+D
That, and a basic lack of discernment.
I mean, fundamentally, if the fed funds rate is 1%, and somebody offers you a AAA-rated safe investment that pays 4% per annum, shouldn't your brain be screaming "TOO GOOD TO BE TRUE" so hard it blows steam out your ears? You don't have to understand what's wrong with it. You don't have to analyze it in depth. All you need to know is, it's paying out quadruple the fed funds rate in interest, and there's GOTTA be a reason.
Yeah, that's no moon. (Wait for it...) It's a space station!
Oh, dear. Your patented method for trademarking a copyright violates my intellectual property rights under the PRO-IP act, the DMCA, and also the Berne Convention. I'm afraid I'm going to have to ask you to cease and desist all use of the method covered by your copyright trademark patent.
Right, the upper limit is not the big problem.
Obviously, if you've got two things orbiting together, the larger one isn't going to be classed as a satellite of the smaller (although, under standard Newtonian physics, the relationship is actually mutual, even with a quite large disparity in size).
There's some question about where to draw the line between twin-planets that aren't quite exactly the same size, versus a planet with a fairly large moon, but there are various ways to write a definition for that that doesn't rely on arbitrary numbers. (For instance, you could say that if the center of mass of the combined system lays outside of either body they are companions, and otherwise the smaller object is a satellite of the larger. This has the consequence of making Pluto and Charon companion minor planets, but we've already taken the big step of demoting Pluto to minor-planet status in the first place.)
The harder question is where to draw the line between a small moon and a mere particle/rock/meteor/asteroid that happens to be in orbit around a much larger body. Personally, I'd vote for gravitationally-self-rounded, because I never felt that the asteroids around Mars were large enough to be proper moons anyway. This would also greatly simplify and reduce the moon counts for the gas giant planets. Smaller orbiting objects would still be natural satellites, of course; the definition would only apply to the word "moon". But the astronomical community has never reached any significant degree of consensus on this issue.
> Firms who discriminate against people who aren't ashamed of their
...
> life and like to talk openly about it will wind up full of drones
I'm thinking they'll wind up full of people who use pseudonyms and/or whose names are along the lines of Jennifer Miller or Robert Anderson.
But yeah, if you really believe that this practice does not help the firm get better employees, and may even hinder them from doing so, then it logically follows that there's no point in worrying much about it, because employers who don't engage in the practice will do well and compete well against the ones who do, and so there will always be jobs available for people who don't worry about the issue.
I really am not sure I see what the big deal is. If I search for myself on Google (by my name, not the pseudonym), the first result is some dude's facebook profile who clearly isn't me, but after that you get my contributions to the Open Clip Art Library, my CPAN profile, the reviews I've written on Amazon,
> Needy Window huh?
As long as they only changed the *default* number of blinks from three to seven, so that you can still configure it down to zero, I don't see a problem there. Because anybody who would be annoyed by seven was probably annoyed by three (the default since WinXP) anyway.
> They don't give you the choice of when to reboot after an update
Yeah, you' think they'd give the system administrator the option to configure a delay into that, so that it wouldn't bug the users for the first N hours, on the assumption that by then the system will be shut down anyway. A lot of office PCs routinely get shut down every night, so if the sysadmin set the upgrade-reboot-delay (or whatever they want to call it) to twenty hours, the logged-in user would never be bothered, and upgrades would still all get installed in reasonably a timely fashion. And if for some reason a user *doesn't* shut down the PC at night, it'll then force the reboot when the delay does expire.
> Now that I think about it, why does an OS need to reboot after
> every installation / removal of an application or update?
Because NTFS doesn't have inodes to abstract away the stored information from the directory entry. As a result, it's not safe to update a file that is currently in use. For some kinds of files, the only way for it to not be in use is if the system isn't fully booted up.
Microsoft has needed for several years now to introduce a successor to NTFS, for a variety of reasons. (Windows should continue to support both, just like it still supports FAT, although after a version or two the main filesystem on which the OS is installed might be required to be the new fs, or certainly it would at least be strongly encouraged.) However, their original concept for what that would mean was a database-like filesystem built on top of NTFS, and when it eventually became clear that that wasn't actually the way to go, they pulled that feature from Longhorn/Vista and sort of set the issue aside for a while, because Vista was already overdue and they needed to get it out the door, with or without a new filesystem. Seven is a minor upgrade built mostly on Vista technology, so a new filesystem isn't in the cards there either. Perhaps for blackcomb (or whatever they're going to call the release after Seven)? Who knows. But sooner or later they will need to do it.
There are several things they really ought to build into their next filesystem. The aforementioned inodes are one. Versioning is another. Optional per-user automatic encryption, tied to login, is another.
> #7: Programs COULD install themselves to your gnome-panel. Wouldn't be that hard.
It would be highly undesirable for a lot of Gnome users, though. We tend to want things they way we want them, set them up that way, and expect them to STAY THAT WAY unless and until we change them. We don't want software making changes to our setup without our permission. If we wanted that, we'd probably use Windows.
> #15: Have you seen the new GNOME network manager?
No, I'm still on etch here, so if Gnome has a *new* feature, I haven't seen it yet.
Come to think of it, though, aren't new features against Gnome's religion these days?
> I finally read the manpage for interfaces and just use ifupdown without the network manager now.
I set up the network settings the way I want them when I installed, and haven't needed to change them since.
> #23 is EVIL!
I never liked Windows Media Player. It always seemed to be too focused on disrupting the user's experience with undesirable features nobody would ever want, things that actively got in the way of playing music (or video, or whatever you were trying to play). Instead of potentially-useful features, like looking up the CD you just inserted in the cddb and filling in the track titles, it instead chose to do stupid annoying stuff, like hide the tracklist and instead show stupid screen-saver-esque effects, advertisements, and other malarkey. Also, the interface appeared to have been designed by a first-year art student with no sense of aesthetics and even less concern for function. When I had a version of Windows that came with WMP, I initially replaced WMP with the old equivalent applications that had come with Windows 95, because they were significantly better. Later I downloaded some freeware media player or another. These days I mostly use xmms. It does what I want and mostly stays out of my way otherwise. If I were using Windows, I'd probably be using a different media player than xmms, but I can tell you this for free: it wouldn't be Windows Media Player.
> #24 is WEIRD! You couldn't resume playback after coming back from sleep mode?
> That seems like a bug fix, not a feature enhancement.
Indeed.
> I sure hope that one day we get some decent open video drivers
> for ATI, because that's all that's keeping me from switching.
I realize this won't help you right *now* (unless you have money to spend on the switch, which is not usually how it works for most of us), but you can always keep this in mind next time you upgrade your hardware, and consider getting a different brand of video card.
Personally, I won't buy ATI or nVidia cards. I don't have money to spend on overpriced wonky hardware that's only good for playing games. Depending on the needs for any given system, I either go with cheap onboard video (typically, Intel chipset), which is plenty good enough for systems with only very basic video needs (servers, and systems that are only used for the web and email and office documents), or a Matrox video card if it actually matters (e.g., for a workstation).
But maybe that's just me.
> It is not human nature, to repeat a failed action in the hopes that the results change...
That depends on the human. Intelligent people try to figure out why it didn't work and do something about that. But a double-digit percentage of the population just plain isn't that bright.
You know what would be really cool, and make network administration a lot more fun? If the print spool would do a checksum or something and, if the print job is exactly the same as the last one, sound an audible alarm and refuse ("Error: Duplicate Print Job. This document has already just been sent to the print spool."), or at least require administrator approval.