Seriously, most people will continue using whatever was already on the computer when they bought the thing, until the day they buy a new one. This is generally a Bad Thing(TM) for all concerned, but it's what people think they want.
> Well, it's probably subjective but I've seen the opposite in many > instances. From what I've seen a single processor G4 Mac is > approximately equal to a 50% faster P4 Windows machine
Well, for one thing, the other poster's PC was a PIII, which is faster than a P4 at the same clock rate. (The advantage of P4 is that the clock rates go higher.)
It also will depend a great deal on what else is running in the background on each. If you have, for example, instant messaging software installed on your PC, that will slow it down.
> The default pager for the "man" program is a program called "more" Huh? I could've sworn it was less. [checks on 10.1.5] Huh, nope, you were right, it _is_ more. How braindead is that? Though I guess if it matters you can always man tar | less Except you probably don't want to use plain old tar on OS X; you probably want hfstar, or zip and unzip. tar will lose information for you. (Apple really should fix this, by bundling a version of tar that does what hfstar does. Whether it should be the default is the only real question.)
If G3s perform so doggone well, how come the performance on year-old iMac G3 systems is so lousy? Is it the lack of RAM, the OS, or what? Because the indigo iMac out behind the circ desk (G3/400, OS9.1) performs a great deal worse than my PII/233 at home, which I bought back in 1998 (though admittedly I added RAM when the prices on SDRAM dropped through the floor a while back).
Now, those G4 iMacs, those are another story. If you can get past the undeniable fact that they are the dorkiest looking computers EVER, they perform way beyond what I would have expected from their CPU clock rate numbers. If I were going to buy an Apple notebook (which I'm currently not planning on doing... my PII system has another year of life in it at least), I would save for an extra few months and get one of the G4 systems. It would last longer that way, and I hate migrating to a new system, spending months getting everything set up the way you want it and getting comfortable with the new system; my goal is to do that as seldom as possible.
> Why bother? The point of vi is that it loads in a microsecond
DOS loads faster than Linux too, but, with Emacs and Linux, making a big deal out of load time would be silly, because you don't need to reload it very often. A near complete list of times you'd need to do that would be roughly as follows: when there's a power outage and you didn't have a UPS, when you upgrade to a new version (though with Emacs, you'd have to restart it if you upgraded Emacs, if you upgraded your X server, or if you upgraded your Linux kernel), and when you have to turn off your computer to install some new hardware. (Okay, I also reboot now and again to use another operating system for a few minutes for one reason or another...)
A while back, I compared the sizes of source tarballs for Emacs 21.0.105 versus the then-current milestone of Mozilla. Emacs was larger. (I did this comparison because someone had filed a request in Bugzilla that the editor have an Emacs emulation mode. I suggested that it would be less work to have a Mozilla emulation mode in Emacs.)
> I find that edit.com is all the editor i really need for most > tasks... why doesn't someone come out with something that simple > for Linux?
Dude, I didn't even use edit.com as an editor when I was using DOS 5 for an operating system. Are there text editors that simple for Linux? Yes, a couple dozen in most distributions plus another several dozen that aren't in the distributions, but nobody gets very emotional about the wonderfulness of any particular one, because they all have pretty much the same complete lack of features, so one's as good as another, if you're into that sort of thing (or as bad if you're not).
Anyway, it would be easy for RedHat to make Emacs behave like vi in their default distribution; just put this in the site-lisp: (viper-mode)
Until we get the elisp backend for gcc working, we still need an OS. Once we get that in place, though, we can compile C and C++ apps to elisp and then byte-compile them, and get things like Mozilla and OpenOffice running inside of Emacs. Once we reach that point, we won't need an OS anymore. (Actually, there's one other thing we need, the multithreading stuff planned for Emacs 0.22.x or 0.23.x, but I was taking that for granted because I don't think the gcc integration will be landing until more like 0.24.x or 0.25.x.)
I'm looking forward to the day when we can _everything running inside of Emacs, including the boot loader. Then we can have a big party and set the final goals for version 1.0...
Oh, and for you vi people, you have nothing to fear, since we already have had vi running inside of Emacs (M-x viper) for some time now.
> 1) have to buy [...] to be able to produce these documents No, you don't. You can "print" from any application to a postscript file using a standard driver and then convert to a PDF. Easy. The only question is, why would you _want_ to do that? The original document was more useful in most cases than the resulting PDF.
> 2) where the documents get positively huge They are larger than they need to be, but this is not the issue IMO.
> 3) where the documents' compression is so varied... you compress, I really don't care about the compression.
My problems with PDF are more in terms of its supposed portability (which is fundamentally a _weakness_ of the format, not a strength; the only thing _less_ portable would be a proprietary word processing format) and the lack of any decent viewing software. (By decent, I mean the ability to use my colour preferences (because blinding white backgrounds are more evil than Bill Gates), the ability to search, to copy text to the clipboard, to scroll off the bottom of one page and onto the top of the next, and to do all the other usual things you can normally do with regular documents that are a royal pain with PDF if they are even possible.) I've tried xpdf, Acrobat Reader, ghostview, GSView, and a couple of others, and _none_ of them can manage the basic level of functionality of Netscape 3 or Word Perfect 6 for Windows 3, to say nothing of the functionality of a modern browser or word processing application. When I have to use a PDF, I feel like I'm stuck in 1985, technologically. Basically, putting your content in a PDF is saying to me, "I want you to really have to be desparate for this information, so I'm going to make you jump through a lot of HOOPS to get it! If you're not truly desparate to see this information, go away."
> Those bastards! Not supporting Windows 95! How dare them end-of-life > a 9 year old piece of software! Nevermind the fact that Apple no > longer supports anything other than OS X
First, the previous poster wasn't talking about Windows 95, but about the pre-SP3 release of WinXP. More importantly, there's a distinct difference here; the difference between OS 9 and OS X is more significant than the difference between W95 and WXP. (Yes, I know XP is NT under the hood, but OS X is still even more different from OS 9.)
Apple is stopping support for an old version that has no real multitasking, no memory protection, no cli (therefore, a pain to support, especialy remotely), and in any number of other ways is generally horrible under the hood (sure, it looks nice to users), not to mention fundamentally incompatible with the current release at the API level, to say nothing of the ABI... it would be unrealistic to expect them to support both at once, because of the enormous differences.
That's not quite the same as stopping support for the previous service pack that is for practical purposes the same, but has a slightly different license agreement and one or two other minor differences. (Whether MS is indeed killing support for pre-SP3 releases of WinXP is another matter; the other posters allege that that is the case, but I personally was unaware of it. My point is that _if_ that _were_ the case, it would not be the same as what Apple is doing.)
* The real kind, where a runaway process CANNOT lock up the
whole system by monopolising the CPU. (It can still crash
everything with a pointer error, but that's a lack of memory
protection talking; OS 9 lacks that _also_, and OS X has it.)
If it's an app you use a lot, it's worth finding their primary bug database and getting familiar with it. Do a quick Google search with the name of the software (OpenOffice, Mozilla, Gimp, whatever), and in most cases you'll find the primary website for that project. (This may be on sourceforge, as you cite, and some programs are hosted at redhat, but many projects have their own site.) Small projects may have an email address where you should send bug reports. Larger projects usually have some kind of bug database. Bugzilla is the one I like best, but there are others (Jitterbug for example). Like I said, if it's an app you use constantly, you may be interested in more than just reporting your bug -- searching to see if it's already been reported, perhaps even already been triaged or even fixed, or how soon that is likely to happen, and so on. Some projects also include feature requests in the same database, so you can track the upcoming features that interest you. If the site is using one of the better issue tracking packages, you can even add your name to a list to be notified when the bug is changed or fixed.
If it's an app you use only infrequently, you may not want to go to quite so much trouble as all that.
The article says that the cheaters can have an impact on the results
of the whole project. I read that, and I did a doubletake: What,
the SETI project has results?
It doesn't have anything to do with what browser we use; it has to do
with having deleted Flash in order to restore sanity to our web
experience.
I also disable window.open() calls during load and unload events,
limit animated GIFs to one cycle through (i.e., no looping forever),
disable scripts from messing with my status bar, restrict websites from
removing my toolbars and resizing my windows, and if I don't like
the colour scheme on a website I turn that off too. (I love
the preferences
toolbar. It rocks.) I tried Flash out briefly, and
I have no reservations about skipping any site that relies on it.
> "Just install Flash" is a fair enough comment...
"Go bite a sidewalk" would be a fair enough reply.
> it's just a case of tracking where the main plugin directory > is for each particular combination of distro and Mozilla.
That part's easy. I know where the plugins go. The problem is, convincing myself to tollerate having Flash installed. I can't. Last time I had it installed, it didn't make it 24 hours before going into a deletium tube. I'm one of those guys who deletes the default plugin just so I won't have to keep getting the "install flash" dialog boxes. (The plugins I _want_, such as Java, I already have.) I don't _want_ my browser to flash. I want it to lay the page out once and then leave it that way so I can read it. If the page author has a problem with that, there are millions of other sites out there, and I won't have time to read even just the interesting ones anyway.
> "Even" for the Vatican? That's like saying the estimate was too much > "even for Cleveland". The Vatican is just a city made out of a museum, > with a really famous person living there. It's only rich culturally.
I suspect the Vatican rakes in more annual revenue than Cleveland on tourism alone. However, it's true that the once-vast riches of the Roman Catholic empire are largely depleted.
> It's common knowledge that speed limits are arbitrarily low
It's also common knowledge that the _reason_ they're set low is because you can't stop people from doing 5mph over, because the tickets are too easy to fight and cost too little. The correct solution IMO (well, _part_ of the correct solution) is to have a 5mph minimum, so that if you're 1mph over it costs the same as if you're 5mph over (obviously, 30mph over would cost more), and to write the law so that it's not any easier to fight the ticket if you're 1mph over than if you're 30mph over. If we can stop people from doing 5 over, we can raise the limit to whatever speed is safe. (Note: that's what speed is determined by the authorities to be safe, not whatever speed makes the driver comfortable.)
> The 85th percentile speed of these roads is much higher than > the posted speed limit, so either an 85th percentile study was > never made (or is woefully out of date), or the speed limit was > intentionally set low to attract speeding ticket money
This is plain ludicrous. The 85th percentile speed (i.e., the speed 85% of people go less than) will rise if you raise the speed limit, as has been tested repeatedly. People with very few exceptions are comfortable going a good deal faster than is safe. (If you don't believe this, try to find a newspaper without a report of a traffic accident in it.) People like you, who believe that whatever speed you are comfortable going is safe, are the reason speed limits exist in the first place.
> That means that I don't know I'm guilty. I do not feel I'm > guilty. I don't feel ashamed, either.
Repeat offenders never do, for any crime. People who feel remorse are people who only got caught maybe once or perhaps twice, if at all. People who have been caught three or more times sometimes are sorry they got _caught_, but they aren't sorry they did it. That's true for everything from speeding all the way up to murder.
Which is why the laws need to be harder on repeat offenders. The ticket system I favour goes something like this...
k is some constant set by lawmakers, and is mostly unimportant.
One dollar, ten dollars, doesn't matter. It doesn't even need
to change with inflation, IMO.
l is the posted speed limit.
s is the speed you were going.
o is the offense number (1 for first offense, 2 for 2nd,...)
This starts at 1 when the system goes into effect; your extant
priors are in the past, since the old system did not adequately
deter speeding. (If this sounds like a break, it is, but when
you see the formula you won't exactly be thinking "great, I can
speed more now". At least, not if you know any arithmetic.)
$ is the amount you owe. You have to pay it to get your license
back, which was confiscated by the officer who stopped you.
(He'll be happy to call you a cab. Sorry about making you late,
but that's what happens when you speed. Have a nice day.)
n is an acceleration constant, also set by lawmakers. Low
values of n (2 or so) allow for a larger number of offenses
in a lifetime; higher values (10 or more) make it painful
to get your fourth of fifth ticket.
if (s<=l) { $=0 }
else {
if (s-l<5) { d=5 }
else { d=s-l }
$ = d * k * n^o
}
The first part (if (s<=l) { $=0 }) is questionable, since perhaps weather conditions should matter, but that's a separate discussion. Anyway, the real solution to that would be to calculate l differently, and have computerised signs that _post_ it differently according to the conditions. That reduces the potential for court arguments over whether visibility was really that low, etc -- you just go whatever speed is posted (or less), and the traffic cop just stops anybody going over what's posted, and you both let the highway people worry about what speed should be posted. (People such as yourself will of course claim that the highway people are in cahoots with the police to make you speed so they can collect revenue, but it won't stand up in court, hopefully.) We wouldn't have the infrastructure to make all signs dynamic, so the static ones could just have fixed limits based on "normal" weather conditions, and you could drive the limit during abnormal weather at your peril. Anyplace where accidents happen in bad weather could get the fancy signs.
This has a couple of effects. First, repeat offenders have a limited number of offenses (inversely proportional to n) to shape up, or they won't be able to afford the tickets. (Wealthy people can afford a couple more tickets than poor people, but that's all; With n=10, you need another digit on your salary to afford each extra ticket. Even with an n of 2, twice the salary means you can afford _one_ more ticket than the guy with half your salary.) k can be set arbitrarily low, so that the first ticket costs basically nothing (say five bucks if you're doing five over) but after a few tickets (again, inversely proportional to n) every offense _hurts_, and it gets worse every time. This would keep the most dangerous maniacs off the road. (If you don't pay the ticket, you forfeit your license to drive until you do pay it. Obviously, we'd still have to enforce the laws against driving without a license, but we have that now.)
The lower limit on d is the solution to your low-limit problem. The 85th-percentile speed will rapidly converge on a point just _below_ the speed limit (as it should be), and the speed limit can then be treated as a _limit_ (rather than a goal) and set to whatever speed would be considered safe.
It would take everyone a year or so to get used to this, so n should probably start at about 2, and everyone's slate wiped clean (once only, as a special measure of grace) when it is raised after about five years.
The nice thing about this system is, it's _firm_, and it's fair: it applies the same standard to everyone. (Well, everyone who isn't exempt on some grounds, such as diplomatic immunity or being on the way to a session of congress or having a lawyer who can get the court to throw out the ticket.) Sure, you can call it draconian, but it only hurts if you speed.
As far as using ticket money to generate revenue, if it's a problem, just raise the excise tax on gas; it's not like most people _really_ think gas costs too much, or they wouldn't be buying so many SUVs.
> However, some staff members had already publicly admitted finding > the crack to journalists
Admitting something to a journalist doesn't make it admissible in court. Not that I'm sympathetic to people who pay lawyers to work the system like that and get them off the hook on technicalities...
> Further proof that if you're rich or powerful criminal law simply > doesn't apply.
Oh, it applies. Sort of. You get penalised by having to pay money to lawyers. The worse your offense, the more money you have to pay. (Consider the amounts O.J. had to spend.) It's not quite the same as doing jail time, though...
The problem with freenet is that its ideology gets in the way of any practical use anyone might want to put it to. You can agree with the ideology all you like, but fundamentally freenet is so concerned about providing free anonymous speech that in practice what it's going to provide is the ability to shout in the forest where nobody hears.
I'll explain. Because they want everything to be anonymous, they made sure content gets spread across all nodes (flooding) and can not be (easily) traced to the given originating node. Consequently, there's no reliable addressing mechanism. You cannot, therefore, create content and make it available at a certain address all the time. All you can do is create the content and watch it get mixed with all the other content.
Survivable? Sure, if you mean by that that as long as people run nodes they'll be sharing _something_, but if you want a particular piece of content to remain available, the only way to ensure that is to keep injecting it again and again and again -- like the way spammers use email. Otherwise, it goes through each node once, in the midst of whatever other content is being injected, and soon is gone. That model is _anything but_ survivable in practice.
Sure, it may work now, when everyone running a freenet node is genuinely concerned about free speech and wants the system to work, but if it ever catches on, it will rapidly devolve into a shouting match, where injecting your content only a few times will ensure no one can find it in the sea of _stuff_ that gets repeatedly injected.
> No slower than redhat maybe, with its convoluted init scripts..
When I was using Windows primarily, and just booting to Linux for experimentation and learning purposes, the long boot time bothered me a little, but I shrugged and said, "it's not that much longer than Windows". After I gave up Pegasus Mail (for Gnus), I started using my Mandrake/Linux/Gnome system for regular work and discovered that I didn't need to boot into Windows anymore... I no longer care about the boot time; it could be twenty minutes, because I almost never have to do it anymore.
FWIW, I'm not a Linux advocate per se[1], and I understand that WinXP has longer uptimes than Win95. I'm just explaining why reboot time doesn't matter in some contexts. I suspect in a financial institution such as a bank, they'd just leave everything on all the time probably, so reboot time would be mostly a non-issue.
[1] Rather than any specific OS, I advocate cross-platform tech
(Gecko, OpenOffice, and the like), and argue for quality.
Linux, for example, really needs to implement automatic file
versioning, like VMS has and like ITS had. It's shameful that
Linux lacks this feature. (Of course, several other major
OSes don't have it either, but nevertheless Linux should.)
If they continue to allow trade secrets like this to leak out,
who knows what could happen. I mean, if the world knows that
MS Office uses XML-based file formats, that could be a huge
disaster! If MS doesn't act quickly to stifle this leak,
cross-platform software
developers might copy
this innovation and take away their competitive advantage!
> I think if you save up karma you can actually post something at "+2" > right off the bat in exchange for that karma.
This surprised me the first time it happened. I posted something, and then I got the message that it had been moderated +1, and was now scored 3. I went, "huh?" because previously my posts had started out at 1.
> Open Office is avaliable for Windows too, so that's like saying > "ford cars run on gas" as a selling point.
This is true. I deploy OpenOffice on the Windows systems at work, at every opportunity.
> Slow reboot time? No slower than any linux going straight to X.
Yes, but in context he directly juxtaposed the reboot time against mention of frequent crashes, implying that NT _needs_ to be rebooted, and that doing so is slow. If I set up a Linux system for a newbie, I'd point to the power switch on the monitor and explain that that's how to turn it on and off. (I'd also make darn sure the filesystem is journaled, because you know there'll be a couple power outages a year.)
> And I won't even start on the virus checker.
My family runs one (yes, that one), because they do unsafe stuff. No, they don't use Outlook (I installed Pegasus Mail), so they're not getting anything by email (Pegasus issues a very ominous warning before allowing the user to launch any executable attachments), but my youngest sister downloads and installs all sorts of #@$!, and I don't trust her to know what's safe. So they run the AV.
If the user in question is afraid to try to install anything, then you can probably dispense with the AV if you can Outlook in favour of a decent email client and don't have a floppy drive. Especially if you make something other than IE the default browser. (I'm making the assumption here that an end-user system won't have IIS, because whoever installed everything will have had the good sense to nix that and any other infamous superfluous services (upnp and such).)
> What are the differences betweeen WinXP (home) and Win98? > Don't even tell me it's more stable - its still Windows
Actually, it isn't the same Windows. Windows XP is NT. I know the version numbers are confusing, but you can visualise the two distinct product lines like this:
The first line can be collectively called "Windows NT" or simply "NT" if you like to abbreviate. Microsoft is working on the next major release (codenamed longhorn), but given the timeframe they've set for that, I expect to see at least one incremental release based on XP before that. This product line has an underlying architecture based partly on XENIX, with concepts borrowed from other systems (including VMS). The GUI is an implementation of the Win32 API.
Note that the second line stops with WinMe. There will be no more versions of that product. This line can be collectively called "Consumer Windows" or "Win9x" but is usually just called "Windows" for simplicity. This product had an architecture based mostly on a re-implementation of DOS with some important changes, most notably multitasking, introduced to bring it into the 90s. (Some claim that Win9x actually _contains_ DOS under the hood, but this is dubious in a technological sense; the "DOS" that is under the hood (MS-DOS 7.0) in Win9x is _a_ DOS but is a reimplementation rather than an incorporation or direct upgrade of the earlier product and so is not 100% compatible with DOS 6. In particular, it is less compatible with DOS 6 than DOS 6 is with DOS 3.) The GUI is an implementation of the Win32 API.
In summary, Win98 and WinXP have two major things in common:
* They are made by the same company. (So are OpenVMS and
digital Unix, for that matter.)
* They both implement the Win32 API. (Yet both BeOS
and Linux implement the POSIX API, and nobody in his
right mind would call them the same OS.)
Oh, and they're both available for x86 hardware. Whoopee, so is NetBSD, and you don't see anyone saying _that's_ the same as Win98. They look a little bit similar (well, they can if WinXP is set up with the "classic" look and feel), but KDE looks like Win98 too, if it's configured that way. They have binary compatibility, but FreeBSD has binary compatibility with SCO and Linux, without being accused of being the same OS as either of them.
The thing is, Microsoft _wants_ you (well, not you individually but everyone in general) to view Windows XP as the successor to Windows 98. Because Windows 98 was their most popular product. Their marketing department will do _anything_ to get you to think that Windows XP is the next version of Windows 98. They tried that with Windows 2000 (by naming it that), and it didn't fly, but Windows XP is doing somewhat better by most accounts.
> The celcius scale is great; I'm just not used to it. I like the > fact that it's based on the solid and gaseous states on water. > If water freezes, it's 0, and if it boils, it's 100. Logical.
It's just as arbitrary as any other scale. You've assumed standard atmospheric pressure, which pretty much negates any benefit your logic might have had for science, and as for ordinary people, the vast bulk of all conversation involving temperature has to do with the weather, not with water boiling. The _only_ argument for the switch to Celsius that makes _any_ sense is, that's what the rest of the world uses. This argument doesn't go over well in the US, for attitude reasons. Consequently, no switch to Celsius is very likely to happen terribly soon.
Fahrenheit is actually great for weather, because 100 is about as warm as you would ever want it to be, and 0 is about as cold as most people would ever want it to be (though personally I don't mind if it's cooler than that). Sure, it's arbitrary, but it connects well with the way most people think. Celsius would work alright if you were used to it, but... we're not.
Slashdot Mirror
s/automatic//;
Seriously, most people will continue using whatever was already on
the computer when they bought the thing, until the day they buy a
new one. This is generally a Bad Thing(TM) for all concerned, but
it's what people think they want.
> Well, it's probably subjective but I've seen the opposite in many
> instances. From what I've seen a single processor G4 Mac is
> approximately equal to a 50% faster P4 Windows machine
Well, for one thing, the other poster's PC was a PIII, which is faster
than a P4 at the same clock rate. (The advantage of P4 is that the
clock rates go higher.)
It also will depend a great deal on what else is running in the
background on each. If you have, for example, instant messaging
software installed on your PC, that will slow it down.
> The default pager for the "man" program is a program called "more"
Huh? I could've sworn it was less. [checks on 10.1.5] Huh, nope,
you were right, it _is_ more. How braindead is that? Though I guess
if it matters you can always man tar | less Except you probably don't
want to use plain old tar on OS X; you probably want hfstar, or zip
and unzip. tar will lose information for you. (Apple really should
fix this, by bundling a version of tar that does what hfstar does.
Whether it should be the default is the only real question.)
If G3s perform so doggone well, how come the performance on year-old
iMac G3 systems is so lousy? Is it the lack of RAM, the OS, or what?
Because the indigo iMac out behind the circ desk (G3/400, OS9.1)
performs a great deal worse than my PII/233 at home, which I bought
back in 1998 (though admittedly I added RAM when the prices on SDRAM
dropped through the floor a while back).
Now, those G4 iMacs, those are another story. If you can get past
the undeniable fact that they are the dorkiest looking computers EVER,
they perform way beyond what I would have expected from their CPU
clock rate numbers. If I were going to buy an Apple notebook (which
I'm currently not planning on doing... my PII system has another year
of life in it at least), I would save for an extra few months and get
one of the G4 systems. It would last longer that way, and I hate
migrating to a new system, spending months getting everything set up
the way you want it and getting comfortable with the new system; my
goal is to do that as seldom as possible.
> Why bother? The point of vi is that it loads in a microsecond
DOS loads faster than Linux too, but, with Emacs and Linux, making a
big deal out of load time would be silly, because you don't need to
reload it very often. A near complete list of times you'd need to
do that would be roughly as follows: when there's a power outage and
you didn't have a UPS, when you upgrade to a new version (though with
Emacs, you'd have to restart it if you upgraded Emacs, if you upgraded
your X server, or if you upgraded your Linux kernel), and when you
have to turn off your computer to install some new hardware. (Okay,
I also reboot now and again to use another operating system for a few
minutes for one reason or another...)
A while back, I compared the sizes of source tarballs for Emacs
21.0.105 versus the then-current milestone of Mozilla. Emacs was
larger. (I did this comparison because someone had filed a request
in Bugzilla that the editor have an Emacs emulation mode. I suggested
that it would be less work to have a Mozilla emulation mode in Emacs.)
> I find that edit.com is all the editor i really need for most
> tasks... why doesn't someone come out with something that simple
> for Linux?
Dude, I didn't even use edit.com as an editor when I was using DOS 5
for an operating system. Are there text editors that simple for Linux?
Yes, a couple dozen in most distributions plus another several dozen
that aren't in the distributions, but nobody gets very emotional about
the wonderfulness of any particular one, because they all have pretty
much the same complete lack of features, so one's as good as another,
if you're into that sort of thing (or as bad if you're not).
Anyway, it would be easy for RedHat to make Emacs behave like vi in
their default distribution; just put this in the site-lisp:
(viper-mode)
Until we get the elisp backend for gcc working, we still need an OS.
Once we get that in place, though, we can compile C and C++ apps to
elisp and then byte-compile them, and get things like Mozilla and
OpenOffice running inside of Emacs. Once we reach that point, we
won't need an OS anymore. (Actually, there's one other thing we need,
the multithreading stuff planned for Emacs 0.22.x or 0.23.x, but I was
taking that for granted because I don't think the gcc integration will
be landing until more like 0.24.x or 0.25.x.)
I'm looking forward to the day when we can _everything running inside
of Emacs, including the boot loader. Then we can have a big party
and set the final goals for version 1.0...
Oh, and for you vi people, you have nothing to fear, since we already
have had vi running inside of Emacs (M-x viper) for some time now.
> Am I the only person who absolutely hates PDF's
No. I loathe them. However...
> 1) have to buy [...] to be able to produce these documents
No, you don't. You can "print" from any application to a postscript
file using a standard driver and then convert to a PDF. Easy. The
only question is, why would you _want_ to do that? The original
document was more useful in most cases than the resulting PDF.
> 2) where the documents get positively huge
They are larger than they need to be, but this is not the issue IMO.
> 3) where the documents' compression is so varied... you compress,
I really don't care about the compression.
My problems with PDF are more in terms of its supposed portability
(which is fundamentally a _weakness_ of the format, not a strength;
the only thing _less_ portable would be a proprietary word processing
format) and the lack of any decent viewing software. (By decent, I
mean the ability to use my colour preferences (because blinding white
backgrounds are more evil than Bill Gates), the ability to search, to
copy text to the clipboard, to scroll off the bottom of one page and
onto the top of the next, and to do all the other usual things you
can normally do with regular documents that are a royal pain with
PDF if they are even possible.) I've tried xpdf, Acrobat Reader,
ghostview, GSView, and a couple of others, and _none_ of them can
manage the basic level of functionality of Netscape 3 or Word Perfect
6 for Windows 3, to say nothing of the functionality of a modern
browser or word processing application. When I have to use a PDF,
I feel like I'm stuck in 1985, technologically. Basically, putting
your content in a PDF is saying to me, "I want you to really have to
be desparate for this information, so I'm going to make you jump
through a lot of HOOPS to get it! If you're not truly desparate to
see this information, go away."
> Those bastards! Not supporting Windows 95! How dare them end-of-life
> a 9 year old piece of software! Nevermind the fact that Apple no
> longer supports anything other than OS X
First, the previous poster wasn't talking about Windows 95, but
about the pre-SP3 release of WinXP. More importantly, there's a
distinct difference here; the difference between OS 9 and OS X
is more significant than the difference between W95 and WXP.
(Yes, I know XP is NT under the hood, but OS X is still even
more different from OS 9.)
Apple is stopping support for an old version that has no real
multitasking, no memory protection, no cli (therefore, a pain to
support, especialy remotely), and in any number of other ways is
generally horrible under the hood (sure, it looks nice to users),
not to mention fundamentally incompatible with the current release
at the API level, to say nothing of the ABI... it would be
unrealistic to expect them to support both at once, because of
the enormous differences.
That's not quite the same as stopping support for the previous
service pack that is for practical purposes the same, but has
a slightly different license agreement and one or two other minor
differences. (Whether MS is indeed killing support for pre-SP3
releases of WinXP is another matter; the other posters allege
that that is the case, but I personally was unaware of it. My
point is that _if_ that _were_ the case, it would not be the
same as what Apple is doing.)
* The real kind, where a runaway process CANNOT lock up the
whole system by monopolising the CPU. (It can still crash
everything with a pointer error, but that's a lack of memory
protection talking; OS 9 lacks that _also_, and OS X has it.)
If it's an app you use a lot, it's worth finding their primary bug
database and getting familiar with it. Do a quick Google search with
the name of the software (OpenOffice, Mozilla, Gimp, whatever), and in
most cases you'll find the primary website for that project. (This
may be on sourceforge, as you cite, and some programs are hosted at
redhat, but many projects have their own site.) Small projects may
have an email address where you should send bug reports. Larger
projects usually have some kind of bug database. Bugzilla is the
one I like best, but there are others (Jitterbug for example). Like
I said, if it's an app you use constantly, you may be interested in
more than just reporting your bug -- searching to see if it's already
been reported, perhaps even already been triaged or even fixed, or
how soon that is likely to happen, and so on. Some projects also
include feature requests in the same database, so you can track the
upcoming features that interest you. If the site is using one of the
better issue tracking packages, you can even add your name to a list
to be notified when the bug is changed or fixed.
If it's an app you use only infrequently, you may not want to go
to quite so much trouble as all that.
The article says that the cheaters can have an impact on the results of the whole project. I read that, and I did a doubletake: What, the SETI project has results?
It doesn't have anything to do with what browser we use; it has to do with having deleted Flash in order to restore sanity to our web experience. I also disable window.open() calls during load and unload events, limit animated GIFs to one cycle through (i.e., no looping forever), disable scripts from messing with my status bar, restrict websites from removing my toolbars and resizing my windows, and if I don't like the colour scheme on a website I turn that off too. (I love the preferences toolbar. It rocks.) I tried Flash out briefly, and I have no reservations about skipping any site that relies on it.
I let animated GIFs play through once, as long as they don't repeat.
> "Just install Flash" is a fair enough comment...
"Go bite a sidewalk" would be a fair enough reply.
> it's just a case of tracking where the main plugin directory
> is for each particular combination of distro and Mozilla.
That part's easy. I know where the plugins go. The problem is,
convincing myself to tollerate having Flash installed. I can't.
Last time I had it installed, it didn't make it 24 hours before going
into a deletium tube. I'm one of those guys who deletes the default
plugin just so I won't have to keep getting the "install flash"
dialog boxes. (The plugins I _want_, such as Java, I already have.)
I don't _want_ my browser to flash. I want it to lay the page out
once and then leave it that way so I can read it. If the page author
has a problem with that, there are millions of other sites out there,
and I won't have time to read even just the interesting ones anyway.
> "Even" for the Vatican? That's like saying the estimate was too much
> "even for Cleveland". The Vatican is just a city made out of a museum,
> with a really famous person living there. It's only rich culturally.
I suspect the Vatican rakes in more annual revenue than Cleveland on
tourism alone. However, it's true that the once-vast riches of the
Roman Catholic empire are largely depleted.
> It's common knowledge that speed limits are arbitrarily low
...)
It's also common knowledge that the _reason_ they're set low is
because you can't stop people from doing 5mph over, because the
tickets are too easy to fight and cost too little. The correct
solution IMO (well, _part_ of the correct solution) is to have a
5mph minimum, so that if you're 1mph over it costs the same as
if you're 5mph over (obviously, 30mph over would cost more), and
to write the law so that it's not any easier to fight the ticket
if you're 1mph over than if you're 30mph over. If we can stop
people from doing 5 over, we can raise the limit to whatever speed
is safe. (Note: that's what speed is determined by the authorities
to be safe, not whatever speed makes the driver comfortable.)
> The 85th percentile speed of these roads is much higher than
> the posted speed limit, so either an 85th percentile study was
> never made (or is woefully out of date), or the speed limit was
> intentionally set low to attract speeding ticket money
This is plain ludicrous. The 85th percentile speed (i.e., the
speed 85% of people go less than) will rise if you raise the speed
limit, as has been tested repeatedly. People with very few
exceptions are comfortable going a good deal faster than is safe.
(If you don't believe this, try to find a newspaper without a report
of a traffic accident in it.) People like you, who believe that
whatever speed you are comfortable going is safe, are the reason
speed limits exist in the first place.
> That means that I don't know I'm guilty. I do not feel I'm
> guilty. I don't feel ashamed, either.
Repeat offenders never do, for any crime. People who feel remorse
are people who only got caught maybe once or perhaps twice, if at
all. People who have been caught three or more times sometimes are
sorry they got _caught_, but they aren't sorry they did it. That's
true for everything from speeding all the way up to murder.
Which is why the laws need to be harder on repeat offenders. The
ticket system I favour goes something like this...
k is some constant set by lawmakers, and is mostly unimportant.
One dollar, ten dollars, doesn't matter. It doesn't even need
to change with inflation, IMO.
l is the posted speed limit.
s is the speed you were going.
o is the offense number (1 for first offense, 2 for 2nd,
This starts at 1 when the system goes into effect; your extant
priors are in the past, since the old system did not adequately
deter speeding. (If this sounds like a break, it is, but when
you see the formula you won't exactly be thinking "great, I can
speed more now". At least, not if you know any arithmetic.)
$ is the amount you owe. You have to pay it to get your license
back, which was confiscated by the officer who stopped you.
(He'll be happy to call you a cab. Sorry about making you late,
but that's what happens when you speed. Have a nice day.)
n is an acceleration constant, also set by lawmakers. Low
values of n (2 or so) allow for a larger number of offenses
in a lifetime; higher values (10 or more) make it painful
to get your fourth of fifth ticket.
if (s<=l) { $=0 }
else {
if (s-l<5) { d=5 }
else { d=s-l }
$ = d * k * n^o
}
The first part (if (s<=l) { $=0 }) is questionable, since perhaps
weather conditions should matter, but that's a separate discussion.
Anyway, the real solution to that would be to calculate l differently,
and have computerised signs that _post_ it differently according to
the conditions. That reduces the potential for court arguments over
whether visibility was really that low, etc -- you just go whatever
speed is posted (or less), and the traffic cop just stops anybody
going over what's posted, and you both let the highway people worry
about what speed should be posted. (People such as yourself will
of course claim that the highway people are in cahoots with the
police to make you speed so they can collect revenue, but it won't
stand up in court, hopefully.) We wouldn't have the infrastructure
to make all signs dynamic, so the static ones could just have fixed
limits based on "normal" weather conditions, and you could drive
the limit during abnormal weather at your peril. Anyplace where
accidents happen in bad weather could get the fancy signs.
This has a couple of effects. First, repeat offenders have a limited
number of offenses (inversely proportional to n) to shape up, or they
won't be able to afford the tickets. (Wealthy people can afford a
couple more tickets than poor people, but that's all; With n=10, you
need another digit on your salary to afford each extra ticket. Even
with an n of 2, twice the salary means you can afford _one_ more
ticket than the guy with half your salary.) k can be set arbitrarily
low, so that the first ticket costs basically nothing (say five bucks
if you're doing five over) but after a few tickets (again, inversely
proportional to n) every offense _hurts_, and it gets worse every
time. This would keep the most dangerous maniacs off the road. (If
you don't pay the ticket, you forfeit your license to drive until you
do pay it. Obviously, we'd still have to enforce the laws against
driving without a license, but we have that now.)
The lower limit on d is the solution to your low-limit problem. The
85th-percentile speed will rapidly converge on a point just _below_
the speed limit (as it should be), and the speed limit can then be
treated as a _limit_ (rather than a goal) and set to whatever speed
would be considered safe.
It would take everyone a year or so to get used to this, so n should
probably start at about 2, and everyone's slate wiped clean (once
only, as a special measure of grace) when it is raised after about
five years.
The nice thing about this system is, it's _firm_, and it's fair: it
applies the same standard to everyone. (Well, everyone who isn't
exempt on some grounds, such as diplomatic immunity or being on the
way to a session of congress or having a lawyer who can get the court
to throw out the ticket.) Sure, you can call it draconian, but it
only hurts if you speed.
As far as using ticket money to generate revenue, if it's a problem,
just raise the excise tax on gas; it's not like most people _really_
think gas costs too much, or they wouldn't be buying so many SUVs.
> However, some staff members had already publicly admitted finding
> the crack to journalists
Admitting something to a journalist doesn't make it admissible in
court. Not that I'm sympathetic to people who pay lawyers to work
the system like that and get them off the hook on technicalities...
> Further proof that if you're rich or powerful criminal law simply
> doesn't apply.
Oh, it applies. Sort of. You get penalised by having to pay money
to lawyers. The worse your offense, the more money you have to pay.
(Consider the amounts O.J. had to spend.) It's not quite the same
as doing jail time, though...
The problem with freenet is that its ideology gets in the way of any
practical use anyone might want to put it to. You can agree with the
ideology all you like, but fundamentally freenet is so concerned about
providing free anonymous speech that in practice what it's going to
provide is the ability to shout in the forest where nobody hears.
I'll explain. Because they want everything to be anonymous, they
made sure content gets spread across all nodes (flooding) and can
not be (easily) traced to the given originating node. Consequently,
there's no reliable addressing mechanism. You cannot, therefore,
create content and make it available at a certain address all the
time. All you can do is create the content and watch it get mixed
with all the other content.
Survivable? Sure, if you mean by that that as long as people run
nodes they'll be sharing _something_, but if you want a particular
piece of content to remain available, the only way to ensure that
is to keep injecting it again and again and again -- like the way
spammers use email. Otherwise, it goes through each node once,
in the midst of whatever other content is being injected, and soon
is gone. That model is _anything but_ survivable in practice.
Sure, it may work now, when everyone running a freenet node is
genuinely concerned about free speech and wants the system to work,
but if it ever catches on, it will rapidly devolve into a shouting
match, where injecting your content only a few times will ensure no
one can find it in the sea of _stuff_ that gets repeatedly injected.
No, no, the correct response is complete incredulity:
"You agreed to WHAT ? [stunned silence] Oh, my. That is a problem." [more stunned silence]
> No slower than redhat maybe, with its convoluted init scripts..
When I was using Windows primarily, and just booting to Linux for
experimentation and learning purposes, the long boot time bothered
me a little, but I shrugged and said, "it's not that much longer
than Windows". After I gave up Pegasus Mail (for Gnus), I started
using my Mandrake/Linux/Gnome system for regular work and discovered
that I didn't need to boot into Windows anymore... I no longer
care about the boot time; it could be twenty minutes, because I
almost never have to do it anymore.
FWIW, I'm not a Linux advocate per se[1], and I understand that
WinXP has longer uptimes than Win95. I'm just explaining why
reboot time doesn't matter in some contexts. I suspect in a
financial institution such as a bank, they'd just leave everything
on all the time probably, so reboot time would be mostly a non-issue.
[1] Rather than any specific OS, I advocate cross-platform tech
(Gecko, OpenOffice, and the like), and argue for quality.
Linux, for example, really needs to implement automatic file
versioning, like VMS has and like ITS had. It's shameful that
Linux lacks this feature. (Of course, several other major
OSes don't have it either, but nevertheless Linux should.)
If they continue to allow trade secrets like this to leak out, who knows what could happen. I mean, if the world knows that MS Office uses XML-based file formats, that could be a huge disaster! If MS doesn't act quickly to stifle this leak, cross-platform software developers might copy this innovation and take away their competitive advantage!
> I think if you save up karma you can actually post something at "+2"
> right off the bat in exchange for that karma.
This surprised me the first time it happened. I posted something, and
then I got the message that it had been moderated +1, and was now
scored 3. I went, "huh?" because previously my posts had started out
at 1.
> Open Office is avaliable for Windows too, so that's like saying
> "ford cars run on gas" as a selling point.
This is true. I deploy OpenOffice on the Windows systems at work,
at every opportunity.
> Slow reboot time? No slower than any linux going straight to X.
Yes, but in context he directly juxtaposed the reboot time against
mention of frequent crashes, implying that NT _needs_ to be rebooted,
and that doing so is slow. If I set up a Linux system for a newbie,
I'd point to the power switch on the monitor and explain that that's
how to turn it on and off. (I'd also make darn sure the filesystem
is journaled, because you know there'll be a couple power outages
a year.)
> And I won't even start on the virus checker.
My family runs one (yes, that one), because they do unsafe stuff.
No, they don't use Outlook (I installed Pegasus Mail), so they're
not getting anything by email (Pegasus issues a very ominous
warning before allowing the user to launch any executable
attachments), but my youngest sister downloads and installs all
sorts of #@$!, and I don't trust her to know what's safe. So they
run the AV.
If the user in question is afraid to try to install anything, then
you can probably dispense with the AV if you can Outlook in favour
of a decent email client and don't have a floppy drive. Especially
if you make something other than IE the default browser. (I'm making
the assumption here that an end-user system won't have IIS, because
whoever installed everything will have had the good sense to nix that
and any other infamous superfluous services (upnp and such).)
> What are the differences betweeen WinXP (home) and Win98?
- -- >
> Don't even tell me it's more stable - its still Windows
Actually, it isn't the same Windows. Windows XP is NT. I know
the version numbers are confusing, but you can visualise the two
distinct product lines like this:
---WinNT3.5-----WinNT4.0-----Win2000-----WinXP-
-----Win95---OSR2----Win98---Win98SE---WinMe
The first line can be collectively called "Windows NT" or simply
"NT" if you like to abbreviate. Microsoft is working on the next
major release (codenamed longhorn), but given the timeframe they've
set for that, I expect to see at least one incremental release
based on XP before that. This product line has an underlying
architecture based partly on XENIX, with concepts borrowed from
other systems (including VMS). The GUI is an implementation of
the Win32 API.
Note that the second line stops with WinMe. There will be no
more versions of that product. This line can be collectively
called "Consumer Windows" or "Win9x" but is usually just called
"Windows" for simplicity. This product had an architecture based
mostly on a re-implementation of DOS with some important changes,
most notably multitasking, introduced to bring it into the 90s.
(Some claim that Win9x actually _contains_ DOS under the hood,
but this is dubious in a technological sense; the "DOS" that is
under the hood (MS-DOS 7.0) in Win9x is _a_ DOS but is a
reimplementation rather than an incorporation or direct upgrade
of the earlier product and so is not 100% compatible with DOS 6.
In particular, it is less compatible with DOS 6 than DOS 6 is
with DOS 3.) The GUI is an implementation of the Win32 API.
In summary, Win98 and WinXP have two major things in common:
* They are made by the same company. (So are OpenVMS and
digital Unix, for that matter.)
* They both implement the Win32 API. (Yet both BeOS
and Linux implement the POSIX API, and nobody in his
right mind would call them the same OS.)
Oh, and they're both available for x86 hardware. Whoopee, so is
NetBSD, and you don't see anyone saying _that's_ the same as Win98.
They look a little bit similar (well, they can if WinXP is set up
with the "classic" look and feel), but KDE looks like Win98 too,
if it's configured that way. They have binary compatibility, but
FreeBSD has binary compatibility with SCO and Linux, without being
accused of being the same OS as either of them.
The thing is, Microsoft _wants_ you (well, not you individually
but everyone in general) to view Windows XP as the successor
to Windows 98. Because Windows 98 was their most popular product.
Their marketing department will do _anything_ to get you to think
that Windows XP is the next version of Windows 98. They tried
that with Windows 2000 (by naming it that), and it didn't fly,
but Windows XP is doing somewhat better by most accounts.
> The celcius scale is great; I'm just not used to it. I like the
> fact that it's based on the solid and gaseous states on water.
> If water freezes, it's 0, and if it boils, it's 100. Logical.
It's just as arbitrary as any other scale. You've assumed standard
atmospheric pressure, which pretty much negates any benefit your
logic might have had for science, and as for ordinary people, the
vast bulk of all conversation involving temperature has to do with
the weather, not with water boiling. The _only_ argument for the
switch to Celsius that makes _any_ sense is, that's what the rest
of the world uses. This argument doesn't go over well in the US,
for attitude reasons. Consequently, no switch to Celsius is very
likely to happen terribly soon.
Fahrenheit is actually great for weather, because 100 is about as
warm as you would ever want it to be, and 0 is about as cold as
most people would ever want it to be (though personally I don't
mind if it's cooler than that). Sure, it's arbitrary, but it
connects well with the way most people think. Celsius would work
alright if you were used to it, but... we're not.