Depends what you search for. If you search for GTK+, you won't get a lot of ads. If you search for CPU, you get two at the top and a trainload of them down the side. Of course, the page still loads in notime flat over dialup.
Actually, go look at Yahoo today. Really. It's noticeably better than last week, already. Okay, it's not Google, not yet anyway, but websites are much easier to change than a brick-and-mortar business. Yahoo's been around the web for most of its history, so presumably they can figure out what they're doing. The article shows that they've realised what works about Google's approach; presumably they'll now set about emulating it.
This is a good thing for all concerned. Yahoo still has the best heirarchical index, even though all the major search engines have been busily trying to copy it, and they've got it fully integrated with their search engine. There's potential there. There are some things they need to work on, of course...
What Google really has in its favour is the Groups. I'd like to see their main search automatically go to the Groups whenever the web doesn't turn up much.
As far as userbase, they're about even. Google may be used for more searches, but just as many people visit Yahoo as Google, so if the Yahoo search gets good, people will notice. They're not coming from obscurity. Right now, people search with Google because they like it. It's simple, it's easy to use, and the ads are all properly targeted, related to words you actually typed in, so they don't seem like an irrelevant waste of time. From the article:
# Companies have found that offering their products in small # text ads that are related to the topic of Web searches is a # great way to find customers who actually want to buy whatever # they sell.
The key there is "actually want to buy". Nobody ever did a Google search for "tropical fish" and saw an underwear advertisement. But that's a relatively easy thing for Yahoo to fix, at least in theory. It means revamping their ad program, but most ad programs get continually revised anyway.
What's really cool is, if Yahoo goes to text-based targeted ads and is successful, the other search engines will start to get the hint too. We could see the death of stupid irrelevant ads that suck a billion baud for animations that advertise products we don't care about or want. If all the advertisements I see are for products at least tangentially related to stuff I'm at least somewhat interested in, I'll consider that an improvement.
Ellison was probably not aware of this (somewhat uncommon) word,
and was probably intending "decimate" as a backformation from the
adjective "decimated" (which means roughly the same thing as
anihilated or eradicated).
That said, a prediction to the effect that Windows will be
anihilated or eradicated by Linux is grossly premature; I
believe several major distros are ready for the desktop and
will shortly begin to cut into Windows' market share (in a
more tangible way than they have to date -- i.e., people
besides utter geeks (e.g., myself) will be involved), but
Windows is not going away for the forseeable future. The
literal meaning of "decimate" might actually be closer to
reality. Even if a supermajor OEM like HP were to start
shipping OSS-based systems exclusively as of next week, it
would still be a decade before Windows would even lose its
market majority, much less cease to be profitable for Microsoft.
Ellison is full of hot air as usual. He's more likely to be
made irrelevant by PostgreSQL and MySQL than Microsoft is by
Linux and OO (both of which I use myself, BTW, so no whining
about FUD).
Then there's what he said about IIS, which is so stupid it
doesn't really even deserve a response. Market dominance?/me is incredulous.
I've never understood the rational behind Quicken and its ilk. A special app just for finances... because, after all, finances are so different from the rows and columns of numbers that spreadsheets were designed to handle. It's like the people who want a "resume program" to use to create a resume, because a resume is so different from the formatted text that a word processing application was made to handle.
Huh?
I don't understand people. I understand computers pretty well, but I don't understand people. Probably never will.
Hey, I wish the folks at MonkeyDance luck. If they're half as successful as the Quicken people, I guess they'll be doing pretty okay for themselves, though I certainly won't understand how.
What I'd like to see as (after a penguin logo) the very first thing that comes up wanting user input is a simple question about the user's level of expertise:
How experienced are you?
( ) Please don't ask me a lot of questions I can't answer,
just get me on the internet.
( ) I can answer simple questions like what kind of tasks
I want to perform with this computer, what my email
address is, and so forth.
( ) I want to select what to install, at least to some
extent. I know the difference between Gnome and KDE
and have a preference.
( ) I have very specific ideas about what I want installed
and how I want it configured. I know what xinetd is.
( ) When the gurus have trouble, they ask me for help.
You know, with pretty icons showing a baby holding a blanket for the first choice and a bunch of hexadecimal numbers for the last choice, or something along those lines.
If they choose the first choice, you ask no avoidable questions. You give them big launchers on the panel that say stuff like "type a document", "get email", and "surf the internet". You bury stuff like the terminal app and control panels under "System Administration" which is under "Advanced" which is under "Other" in the menu; the toplevel items are stuff like "play a cd" and "turn off computer". You only install one application in each category, so if you install OO you don't install KOffice or Gnumeric, and if you install kcalc you don't need the gnome calculator, and if you install gnome-terminal you don't install konsole or eterm. i.e., the distro makes all the decisions and the user just uses whatever you give them. Oh, and you don't ask them for a username and password; you generate one and set it to autologin. (The root password you should generate too, give it to them, and tell them to write it down and keep it safe in case they ever have to get help, the person who helps them may need it.)
If they choose the second choice, you ask them stuff like "Do you want to install the basic development packages so that you can download and install software that is distributed as source code?", whether they want to share their internet connection with other computers on the local network, and stuff like that. You still don't ask about specific apps, and you pick one app that your distro feels is best for users at this level wanting to perform a given task. The user tells you they want to type documents, and you give them OO (or whatever). The menus should be structured by task, but they can have the name of the app. "Advanced" (or whatever you call it) can be a toplevel menu item, with stuff like the terminal app and control panels just inside it (thus, only two levels deep). At this level you can ask for a username and password and whether autologin should be used.
The third choice should give you roughly what the Mandrake install has been like if you don't turn on individual package selection. You can get multiple apps in the same category this way (e.g., both Gnome Office and KOffice if you select both Gnome and KDE).
The fourth choice turns on individual package selection, so you can wade through trees full of packages and individually choose which text editor(s) to install and stuff. This is what I would choose:-)
The fifth choice is full geek mode, where you get to do everything, including configure your kernel, and the default wallpaper says the name of the distro in ASCII binary, if you choose to install X.
What we need is more radical than a fork of XFree
on
XFree86 Politics
·
· Score: 3, Insightful
What we actually need is a replacement for X11, redesigned from the ground up, with compatability libs to allow normal, window-bound X11 apps to work on it. We'd lose existing "special" apps (window managers, screensavers, panels,...), which is a shame, but it's what needs to be done to allow for future improvements.
Don't get me wrong, I mostly like XFree... but the design is (gradually) reaching the end of its useful lifespan. There are a number of improvements I'd like to see that are fairly impractical for a design based on X11. Resizing windows is nice, but I also want to be able to scale them. (This implies that bitmapped fonts should die, among other things.) Being able to grab a bitmap of the desktop and use it as a window background is one thing, but I really want a full alpha channel for every window (controlled by the application for each widget in the window, or for each pixel in an image canvas widget) plus an overall opacity setting (controlled by the user) for the whole window. And so on.
Uniformitarianism as it is generally applied in practice requires things to stay the same that... don't. The gradualism we were all taught in junior high school relies on such things being constant as climate (mean temperature in an area, mean wind speed in an area, mean rainfall in an area,...), Earth's magnetic field, and so on. It falls over under even moderate scrutiny, yes, but that doesn't make it a straw man; I was really taught this stuff, in "science" class, by men who were convinced it was true.
Similarly, radioactive decay dating only works if rocks are closed systems. This is easy to believe if you don't know better, because rocks *look* like a closed system at first glance. And on a time scale of a few days, they mostly are. (A bit of water gets through, but water is a pretty small molecule.) But on a timescale of decades, rocks are as open a system as trees and lakes; stuff comes in and goes out and the contents change -- repeatedly. But to set a date for the rock by measuring the ratios of pre-decay and post- decay elements you have to assume that it has been a closed system since its formation, centuries or millenia ago; otherwise, the pre-decay and post-decay elements have come and gone any number of times and the ratio is meaningless.
I *wish* people were on the whole intelligent enough to view uniformitarianism as a straw man. Also the half-baked notion that the the earth revolves around the sun. (We've known since Newton that they revolve around eachother and around every other nearby object of substantial mass; every astronomer knows this, and can tell you about the measurable effect of (e.g.) Jupiter on the Earth's motion, but most people still believe the Earth revolves around the Sun.)
You ever try to buy RAM for an eight-year-old Compaq? You know, the kind that stores its BIOS stuff on the hard drive and predates things like PCI and (more pertinently) SDRAM?
The easiest way would be to buy a lot of two dozen such systems on eBay for roughly the cost of shipping and hope one or two of them might have compatible RAM, attempt to figure out which if any do by trial and error, then use them all as geeky decorations or bookends or stepstools or something.
486s are even *worse* to buy RAM for; some of them use pre-EDO types of RAM, which are about as easy to work with as vacuum tubes, and the motherboards are about as well-documented as obfuscated code competition entries. Oh, and they usually max out at a whole lot less RAM than you'd need to copy the contents of a CD-ROM.
Incidentally, with uniformitarianism thoroughly discredited, that means radioactive-decay dating methods are unreliable. But we already knew that, too.
Sheesh, this is NEWS? The earth is an open system: that's been established _repeatedly_ now. The energy coming in from the Sun (and trace amounts from other sources) is not without effect, duh. So of _course_ stuff changes. Yeah, the earth's rotation changes, its inclination to the eccliptic changes, it's orbit changes, its mass changes, the distance to the moon changes, the composition of the atmosphere changes, the chemical content of any given rock changes, et cetera. Uniformitarianism is an interesting idea, but it doesn't jive with the real world.
Next they'll be reporting that the English language changes too...
My Pentium/90 dialup router is running 2.2.17-21mdk, just because I haven't ever bothered to upgrade it. Just about the only change I've made to it in the last year is set it to forward port 80 to my desktop, so I can get to my personal Bugzilla and stuff from work. I used to do a lot of stuff on the router, when my desktop was running Windows, but with my desktop running Linux/Gnome now I no longer have to shell into the router to do stuff, except to redial the ppp link, and that's largely automated.
In another year or so I'll probably replace my desktop, at which point my current desktop (a PII/233) will probably become the new dialup router, but in the meanwhile there's not a lot of point in messing with my current setup. Other things to do with my time, and all that. Installing a new kernel on my networking hardware rates right up there with dusting under the bed and flossing the dog's teeth and watching television: I'm not that bored.
If you can't use email for time-sensitive information, there's something wrong with your setup. email is delivered just as quickly as IM, and with a good biff you are aware whenever you get email; the difference is in the behavior of the client, with regard to how it treats other apps on the system and how it consumes system resources: most email clients are well- behaved, but I've never seen a well-behaved IM client.
> They recited the reasons we should not have it, how it was > dangerous, etc. but no alternatives.
email is flexible enough to do all the things you describe doing with IM, but it won't crash your system.
> There's no requirement to justify why it has to > be installed
There sure ought to be. IM is *horrible* for system stability. There's no way I'd approve it for installation on any system I have to administer. Especially not on Windows systems. You've got email, and you've got a phone system, and you've all got mailboxes: you don't need IM.
I'll support Windows, and I'll even support junkware like Acrobat Reader and Flash, but I draw the line at IM. Oh, I don't support Bonzi Buddy either.
When a server process runs as root, every vulnerability anyone finds in it is a root vulnerability. For extra bonus insecurity points, write it in a language that doesn't protect you from memory managment errors, and then have a security philosophy that says, in effect, "if the environment isn't exactly what we want it to be, any insecurities aren't our fault".
I've been saying for months that this would happen. It will happen again, too. It's high time to retire sendmail and adopt other solutions.
> > Unfortunately the critical mass for this to really work is > > very, very large.
Yes, it is large.
> I don't think this is necessarily true. As the article points > out, setting it up on a few servers would be sufficient to get > things started provided those few servers were the right ones.
Let me guess: Yahoo's several dozen, AOL's however many, and the ones at Earthlink, demon.co.uk, and MSN -- and I close?
That's a very large critical mass, not in terms of the number of servers, but in terms of the amount of mail handled (and, therefore, the amount of server beef needed to implement any such measures).
> I don't think they should be doing this in Java though. Java is > not a text parsing language and this thing really requires some > text parsing muscle. Cross platform ability isn't as important.
No need to sacrifice the cross-platformness. Perl is a GREAT text processing language, performs faster than Java, and as an added bonus is much more cross-platform (provided you don't need a GUI (which for this you don't)). It does use quite a bit of RAM sometimes, but so does Java. And doing SMTP stuff in Perl is really easy. (Net::SMTP rocks in a significant way.) And any operating system that's remotely appropriate for use as a mail server probably comes with Perl out of the box these days.
1. Obtain sheet music for complete works of Bach
2. Scan it all in
3. Write special open-source music-OCR software that
transcribes sheet music into MIDI format, and run
your scanned sheet music through
4. Use timidity or something to make WAV files out of it
5. ???
6. Profit!
Of course, you'll need to hire a bunch of employees if you want to complete step 2 in a reasonable timeframe...
Indeed. I knew it when I was ten, and I'd never even met an actual locksmith.
The solution is equally simple: if security actually matters, you sacrifice the convenience of having a single master key and install locks that use a completely different key in the places that matter. Your "master key" is then a whole ring of keys, but hey.
Next they'll start talking about how the social engineering technique used by computer crackers can be used in the real world too... just phone up the front desk and ask 'em to unlock the side door and let in the plumber...
> you can't put root.root files in a user's $HOME and chmod 644 > them and hope the user can't modify them. As they are in the > user directory, in which the user can do whatever she wants, > the can be unlinked/moved elsewhere.
So set up the Mozilla launcher to run Mozilla suid as another user. (The user can, of course, launch Mozilla another way, or install another copy of it (if there's room in their home dir), or another browser altogether (if there's room), but it's possible at least to ensure that the default install can't be eliminated.) The only reason I can think of to do such a thing would be if you wanted to maintain a guest user account, which is usually a bad idea. If it's a question of the user screwing up his install by mistake and wanting it back, you can just put a tarball in a protected place and give the user a "restore defaults" button that runs a script that untars it in the user's home directory. Voila.
If all they did was prevent you from changing the prefs from within Netscape (e.g., using the CCK), then a quick Google search will tell you how to get around it with nothing more complicated than a text editor. If they really want to keep you from changing anything, they have to protect the settings at the filesystem level by denying you write access (which they may have done, but that's a separate topic, since it really has nothing to do with Netscape or Mozilla per se; the same technique could protect the settings for any application or even prevent you from altering a document).
Yeah, Gecko is big. It has to be, to get all the layouts correct. Understand, it's designed to lay out and render, correctly, anything from non-wellformed pre-W3C HTML on the one end of the scale up through XSLT at the other end, plus XUL. That's a tall order. Konqueror doesn't handle quite as wide a spectrum.
That said, KHTML handles more of MSIE's proprietary non-W3C extensions to the DOM than Gecko does, which _may_ be part of why Apple chose it.
There are two types of coasters: those designed to be enjoyable and give you fond memories of the wonderful time you had so you come back, and those designed for bragging rights. This is the latter type. It is made to be the tallest and fastest, so that Cedar Point can once again bost the tallest and fastest coaster. The Mean Streak was built for similar reasons: they got tired of PKI being able to claim the longest wooden coaster. (The Beast is still a much better ride than the Mean Streak -- not that it really matters; PKI only has two or three really good coasters, so it's not in the same league with the Point; if it weren't so close physically it wouldn't even be considered as a form of competition.)
> Want a longer ride? Buy another ticket
Huh? I've never heard of an amusement park charging for _each_ ride. That's a non-issue. But you do have to stand in line, and in any event, longer rides are known for being more enjoyable. The Magnum is an excellent coaster that a lot of people still want to ride, not because it was once the tallest and fastest in the world but because the whole ride is enjoyable, not just the first hill. (The Magnum was one of the rare coasters built for both bragging rights _and_ memorable rider experience.)
It's more like this: want an enjoyable, memorable ride? There are no shortage of coasters at the Point, so you won't have any trouble finding such a ride. But one thing they didn't have (anymore) was the tallest, steepest, fastest. And they haven't had that for too many years; they _had_ to build a tallest-steepest-fastest before very much longer, or their reputation would flag. You don't keep a reputation for being the coaster capital of the world easily. There are too many other places that want to catch up with that (though no more than at most half a dozen places are really even contenders, and none of them are within a day's drive).
> The incremental find text is a bit nicer than ctrl-F
It is; I just haven't got used to being able to do that in my web browser yet. I hit Ctrl-F out of habbit. The other thing is, the incremental search in Mozilla (typeahead find) is lacking some of the features that would make it really useful. In particular, with Emacs incremental search you can type the first part of what you're looking for, hit Ctrl-S a couple of times to move forward through a couple of instances, type a little more onto your search string, hit Ctrl-S a couple more times, realise you went past the instance you were interested in, hit Ctrl-R to go back to it... The feature in Mozilla is not quite so mature in its implementation.
> since it doesn't drag up a little dialog that (in Mozilla 1.2 > under OSX) you can't close without using your mouse.
That's a Mac thing. You can't do Jack Squat(TM) on a Mac without using a mouse. In fact, you can generally use a Mac better with no keyboard than with no mouse. This has been true at _least_ since the days of System 6, and it's only any different with OS X if you intend to spend most of your time in Terminal.app or using X11 apps that have been ported over. On any other platform, _all_ dialogs can be closed without using a mouse. (On Win32 or Gnome, there are no fewer than three distinct ways to close any dialog with the keyboard.) This is nice; in the summer around here (though not this last summer; we had a dry year) it often gets sufficiently humid that the mouse is basically not usable, and I have a tendency during August to set the mouse on a shelf and use the keyboard pretty much exclusively until fall comes and dries things up. The _only_ thing I've discovered I can't do effeciently without a mouse is image editing. (Though optical mice are getting pretty cheap these days; this year when summer rolls around I might just get one of those, and then if I need to use Gimp I can...)
Re:You misunderstand completely
on
E ~ mc^2
·
· Score: 2
> Do you believe cosmology is not a science because we can't > recreate the big bang?
> No type-ahead? Let's see.. typing "sl".. ooh.. completes to > "slashdot.org"
You're typing in the location bar. He was talking about type-ahead find, which is roughly analogous to the incremental search feature in Emacs (albeit less powerful in several major ways). With your cursor in the page, you just start typing, and it searches forward and finds the first occurrance of what you typed -- either in a link (if you just start typing) or anywhere (if you start with a/ before typing what you want to find). It's fairly useful, although I must confess that I haven't gotten in the habbit yet of using it in Mozilla; I still find myself hitting Ctrl-F and Ctrl-G a lot.
Slashdot Mirror
Depends what you search for. If you search for GTK+, you won't
get a lot of ads. If you search for CPU, you get two at the top
and a trainload of them down the side. Of course, the page still
loads in notime flat over dialup.
Actually, go look at Yahoo today. Really. It's noticeably better
than last week, already. Okay, it's not Google, not yet anyway,
but websites are much easier to change than a brick-and-mortar
business. Yahoo's been around the web for most of its history,
so presumably they can figure out what they're doing. The article
shows that they've realised what works about Google's approach;
presumably they'll now set about emulating it.
This is a good thing for all concerned. Yahoo still has the best
heirarchical index, even though all the major search engines have
been busily trying to copy it, and they've got it fully integrated
with their search engine. There's potential there. There are
some things they need to work on, of course...
What Google really has in its favour is the Groups. I'd like to
see their main search automatically go to the Groups whenever the
web doesn't turn up much.
As far as userbase, they're about even. Google may be used for
more searches, but just as many people visit Yahoo as Google, so
if the Yahoo search gets good, people will notice. They're not
coming from obscurity. Right now, people search with Google
because they like it. It's simple, it's easy to use, and the
ads are all properly targeted, related to words you actually
typed in, so they don't seem like an irrelevant waste of time.
From the article:
# Companies have found that offering their products in small
# text ads that are related to the topic of Web searches is a
# great way to find customers who actually want to buy whatever
# they sell.
The key there is "actually want to buy". Nobody ever did a Google
search for "tropical fish" and saw an underwear advertisement.
But that's a relatively easy thing for Yahoo to fix, at least
in theory. It means revamping their ad program, but most ad
programs get continually revised anyway.
What's really cool is, if Yahoo goes to text-based targeted ads
and is successful, the other search engines will start to get
the hint too. We could see the death of stupid irrelevant ads
that suck a billion baud for animations that advertise products
we don't care about or want. If all the advertisements I see
are for products at least tangentially related to stuff I'm at
least somewhat interested in, I'll consider that an improvement.
Ellison was probably not aware of this (somewhat uncommon) word, and was probably intending "decimate" as a backformation from the adjective "decimated" (which means roughly the same thing as anihilated or eradicated).
This is a much more common usage in practice, and thesaurus.com seems to agree with it.
That said, a prediction to the effect that Windows will be anihilated or eradicated by Linux is grossly premature; I believe several major distros are ready for the desktop and will shortly begin to cut into Windows' market share (in a more tangible way than they have to date -- i.e., people besides utter geeks (e.g., myself) will be involved), but Windows is not going away for the forseeable future. The literal meaning of "decimate" might actually be closer to reality. Even if a supermajor OEM like HP were to start shipping OSS-based systems exclusively as of next week, it would still be a decade before Windows would even lose its market majority, much less cease to be profitable for Microsoft.
Ellison is full of hot air as usual. He's more likely to be made irrelevant by PostgreSQL and MySQL than Microsoft is by Linux and OO (both of which I use myself, BTW, so no whining about FUD).
Then there's what he said about IIS, which is so stupid it doesn't really even deserve a response. Market dominance? /me is incredulous.
I've never understood the rational behind Quicken and its ilk. A
special app just for finances... because, after all, finances are
so different from the rows and columns of numbers that spreadsheets
were designed to handle. It's like the people who want a "resume
program" to use to create a resume, because a resume is so different
from the formatted text that a word processing application was made
to handle.
Huh?
I don't understand people. I understand computers pretty well,
but I don't understand people. Probably never will.
Hey, I wish the folks at MonkeyDance luck. If they're half as
successful as the Quicken people, I guess they'll be doing pretty
okay for themselves, though I certainly won't understand how.
What I'd like to see as (after a penguin logo) the very first thing
:-)
that comes up wanting user input is a simple question about the
user's level of expertise:
How experienced are you?
( ) Please don't ask me a lot of questions I can't answer,
just get me on the internet.
( ) I can answer simple questions like what kind of tasks
I want to perform with this computer, what my email
address is, and so forth.
( ) I want to select what to install, at least to some
extent. I know the difference between Gnome and KDE
and have a preference.
( ) I have very specific ideas about what I want installed
and how I want it configured. I know what xinetd is.
( ) When the gurus have trouble, they ask me for help.
You know, with pretty icons showing a baby holding a blanket
for the first choice and a bunch of hexadecimal numbers for
the last choice, or something along those lines.
If they choose the first choice, you ask no avoidable questions.
You give them big launchers on the panel that say stuff like
"type a document", "get email", and "surf the internet". You
bury stuff like the terminal app and control panels under
"System Administration" which is under "Advanced" which is
under "Other" in the menu; the toplevel items are stuff like
"play a cd" and "turn off computer". You only install one
application in each category, so if you install OO you don't
install KOffice or Gnumeric, and if you install kcalc you don't
need the gnome calculator, and if you install gnome-terminal
you don't install konsole or eterm. i.e., the distro makes
all the decisions and the user just uses whatever you give them.
Oh, and you don't ask them for a username and password; you
generate one and set it to autologin. (The root password you
should generate too, give it to them, and tell them to write
it down and keep it safe in case they ever have to get help,
the person who helps them may need it.)
If they choose the second choice, you ask them stuff like
"Do you want to install the basic development packages so
that you can download and install software that is distributed
as source code?", whether they want to share their internet
connection with other computers on the local network, and
stuff like that. You still don't ask about specific apps,
and you pick one app that your distro feels is best for
users at this level wanting to perform a given task. The
user tells you they want to type documents, and you give
them OO (or whatever). The menus should be structured by
task, but they can have the name of the app. "Advanced"
(or whatever you call it) can be a toplevel menu item, with
stuff like the terminal app and control panels just inside it
(thus, only two levels deep). At this level you can ask
for a username and password and whether autologin should
be used.
The third choice should give you roughly what the Mandrake
install has been like if you don't turn on individual
package selection. You can get multiple apps in the same
category this way (e.g., both Gnome Office and KOffice if
you select both Gnome and KDE).
The fourth choice turns on individual package selection, so
you can wade through trees full of packages and individually
choose which text editor(s) to install and stuff. This is
what I would choose
The fifth choice is full geek mode, where you get to do everything,
including configure your kernel, and the default wallpaper says
the name of the distro in ASCII binary, if you choose to install X.
What we actually need is a replacement for X11, redesigned from the ...), which is a shame, but it's
ground up, with compatability libs to allow normal, window-bound
X11 apps to work on it. We'd lose existing "special" apps (window
managers, screensavers, panels,
what needs to be done to allow for future improvements.
Don't get me wrong, I mostly like XFree... but the design is
(gradually) reaching the end of its useful lifespan. There are a
number of improvements I'd like to see that are fairly impractical
for a design based on X11. Resizing windows is nice, but I also
want to be able to scale them. (This implies that bitmapped fonts
should die, among other things.) Being able to grab a bitmap of
the desktop and use it as a window background is one thing, but
I really want a full alpha channel for every window (controlled
by the application for each widget in the window, or for each
pixel in an image canvas widget) plus an overall opacity setting
(controlled by the user) for the whole window. And so on.
Uniformitarianism as it is generally applied in practice requires ...), Earth's magnetic field,
things to stay the same that... don't. The gradualism we were
all taught in junior high school relies on such things being
constant as climate (mean temperature in an area, mean wind speed
in an area, mean rainfall in an area,
and so on. It falls over under even moderate scrutiny, yes, but
that doesn't make it a straw man; I was really taught this stuff,
in "science" class, by men who were convinced it was true.
Similarly, radioactive decay dating only works if rocks are
closed systems. This is easy to believe if you don't know
better, because rocks *look* like a closed system at first
glance. And on a time scale of a few days, they mostly are.
(A bit of water gets through, but water is a pretty small
molecule.) But on a timescale of decades, rocks are as open
a system as trees and lakes; stuff comes in and goes out
and the contents change -- repeatedly. But to set a date
for the rock by measuring the ratios of pre-decay and post-
decay elements you have to assume that it has been a closed
system since its formation, centuries or millenia ago;
otherwise, the pre-decay and post-decay elements have come
and gone any number of times and the ratio is meaningless.
I *wish* people were on the whole intelligent enough to
view uniformitarianism as a straw man. Also the half-baked
notion that the the earth revolves around the sun. (We've
known since Newton that they revolve around eachother and
around every other nearby object of substantial mass; every
astronomer knows this, and can tell you about the measurable
effect of (e.g.) Jupiter on the Earth's motion, but most
people still believe the Earth revolves around the Sun.)
> Easy solution, Get a little more RAM
You ever try to buy RAM for an eight-year-old Compaq? You know,
the kind that stores its BIOS stuff on the hard drive and predates
things like PCI and (more pertinently) SDRAM?
The easiest way would be to buy a lot of two dozen such systems
on eBay for roughly the cost of shipping and hope one or two of
them might have compatible RAM, attempt to figure out which if
any do by trial and error, then use them all as geeky decorations
or bookends or stepstools or something.
486s are even *worse* to buy RAM for; some of them use pre-EDO
types of RAM, which are about as easy to work with as vacuum
tubes, and the motherboards are about as well-documented as
obfuscated code competition entries. Oh, and they usually max
out at a whole lot less RAM than you'd need to copy the contents
of a CD-ROM.
Incidentally, with uniformitarianism thoroughly discredited,
that means radioactive-decay dating methods are unreliable.
But we already knew that, too.
Sheesh, this is NEWS? The earth is an open system: that's been
established _repeatedly_ now. The energy coming in from the Sun
(and trace amounts from other sources) is not without effect, duh.
So of _course_ stuff changes. Yeah, the earth's rotation changes,
its inclination to the eccliptic changes, it's orbit changes, its
mass changes, the distance to the moon changes, the composition
of the atmosphere changes, the chemical content of any given
rock changes, et cetera. Uniformitarianism is an interesting
idea, but it doesn't jive with the real world.
Next they'll be reporting that the English language changes too...
My Pentium/90 dialup router is running 2.2.17-21mdk, just because
I haven't ever bothered to upgrade it. Just about the only change
I've made to it in the last year is set it to forward port 80 to
my desktop, so I can get to my personal Bugzilla and stuff from
work. I used to do a lot of stuff on the router, when my desktop
was running Windows, but with my desktop running Linux/Gnome now
I no longer have to shell into the router to do stuff, except to
redial the ppp link, and that's largely automated.
In another year or so I'll probably replace my desktop, at which
point my current desktop (a PII/233) will probably become the new
dialup router, but in the meanwhile there's not a lot of point in
messing with my current setup. Other things to do with my time,
and all that. Installing a new kernel on my networking hardware
rates right up there with dusting under the bed and flossing the
dog's teeth and watching television: I'm not that bored.
> email for non-time sensitive information
If you can't use email for time-sensitive information, there's
something wrong with your setup. email is delivered just as
quickly as IM, and with a good biff you are aware whenever you
get email; the difference is in the behavior of the client,
with regard to how it treats other apps on the system and how
it consumes system resources: most email clients are well-
behaved, but I've never seen a well-behaved IM client.
> They recited the reasons we should not have it, how it was
> dangerous, etc. but no alternatives.
email is flexible enough to do all the things you describe
doing with IM, but it won't crash your system.
> There's no requirement to justify why it has to
> be installed
There sure ought to be. IM is *horrible* for
system stability. There's no way I'd approve
it for installation on any system I have to
administer. Especially not on Windows systems.
You've got email, and you've got a phone system,
and you've all got mailboxes: you don't need IM.
I'll support Windows, and I'll even support
junkware like Acrobat Reader and Flash, but I
draw the line at IM. Oh, I don't support Bonzi
Buddy either.
When a server process runs as root, every vulnerability anyone finds
in it is a root vulnerability. For extra bonus insecurity points,
write it in a language that doesn't protect you from memory managment
errors, and then have a security philosophy that says, in effect,
"if the environment isn't exactly what we want it to be, any
insecurities aren't our fault".
I've been saying for months that this would happen. It will happen
again, too. It's high time to retire sendmail and adopt other
solutions.
> > Unfortunately the critical mass for this to really work is
> > very, very large.
Yes, it is large.
> I don't think this is necessarily true. As the article points
> out, setting it up on a few servers would be sufficient to get
> things started provided those few servers were the right ones.
Let me guess: Yahoo's several dozen, AOL's however many, and
the ones at Earthlink, demon.co.uk, and MSN -- and I close?
That's a very large critical mass, not in terms of the number of
servers, but in terms of the amount of mail handled (and, therefore,
the amount of server beef needed to implement any such measures).
> I don't think they should be doing this in Java though. Java is
> not a text parsing language and this thing really requires some
> text parsing muscle. Cross platform ability isn't as important.
No need to sacrifice the cross-platformness. Perl is a GREAT
text processing language, performs faster than Java, and as an
added bonus is much more cross-platform (provided you don't need
a GUI (which for this you don't)). It does use quite a bit of
RAM sometimes, but so does Java. And doing SMTP stuff in Perl
is really easy. (Net::SMTP rocks in a significant way.) And
any operating system that's remotely appropriate for use as a
mail server probably comes with Perl out of the box these days.
1. Obtain sheet music for complete works of Bach
2. Scan it all in
3. Write special open-source music-OCR software that
transcribes sheet music into MIDI format, and run
your scanned sheet music through
4. Use timidity or something to make WAV files out of it
5. ???
6. Profit!
Of course, you'll need to hire a bunch of employees if you
want to complete step 2 in a reasonable timeframe...
> Nope, the master key in this situation is the minimal set of keys
> that it takes to get into the place where the whole ring is stored.
On someone's person, I would hope. Preferably someone who is never
alone on the job.
None of this prevents anyone from drilling the locks, but if they
do that you'll know about it.
> Everybody knows that.
Indeed. I knew it when I was ten, and I'd never even met an actual
locksmith.
The solution is equally simple: if security actually matters, you
sacrifice the convenience of having a single master key and install
locks that use a completely different key in the places that matter.
Your "master key" is then a whole ring of keys, but hey.
Next they'll start talking about how the social engineering technique
used by computer crackers can be used in the real world too...
just phone up the front desk and ask 'em to unlock the side door
and let in the plumber...
You might need to use -R also.
> you can't put root.root files in a user's $HOME and chmod 644
> them and hope the user can't modify them. As they are in the
> user directory, in which the user can do whatever she wants,
> the can be unlinked/moved elsewhere.
So set up the Mozilla launcher to run Mozilla suid as another
user. (The user can, of course, launch Mozilla another way, or
install another copy of it (if there's room in their home dir),
or another browser altogether (if there's room), but it's possible
at least to ensure that the default install can't be eliminated.)
The only reason I can think of to do such a thing would be if you
wanted to maintain a guest user account, which is usually a bad
idea. If it's a question of the user screwing up his install by
mistake and wanting it back, you can just put a tarball in a
protected place and give the user a "restore defaults" button that
runs a script that untars it in the user's home directory. Voila.
If all they did was prevent you from changing the prefs from within
Netscape (e.g., using the CCK), then a quick Google search will tell
you how to get around it with nothing more complicated than a text
editor. If they really want to keep you from changing anything,
they have to protect the settings at the filesystem level by denying
you write access (which they may have done, but that's a separate
topic, since it really has nothing to do with Netscape or Mozilla
per se; the same technique could protect the settings for any
application or even prevent you from altering a document).
Yeah, Gecko is big. It has to be, to get all the layouts correct.
Understand, it's designed to lay out and render, correctly, anything
from non-wellformed pre-W3C HTML on the one end of the scale up
through XSLT at the other end, plus XUL. That's a tall order.
Konqueror doesn't handle quite as wide a spectrum.
That said, KHTML handles more of MSIE's proprietary non-W3C extensions
to the DOM than Gecko does, which _may_ be part of why Apple chose it.
> but I don't see why they are so short
There are two types of coasters: those designed to be enjoyable and
give you fond memories of the wonderful time you had so you come back,
and those designed for bragging rights. This is the latter type. It
is made to be the tallest and fastest, so that Cedar Point can once
again bost the tallest and fastest coaster. The Mean Streak was built
for similar reasons: they got tired of PKI being able to claim the
longest wooden coaster. (The Beast is still a much better ride than
the Mean Streak -- not that it really matters; PKI only has two or
three really good coasters, so it's not in the same league with the
Point; if it weren't so close physically it wouldn't even be
considered as a form of competition.)
> Want a longer ride? Buy another ticket
Huh? I've never heard of an amusement park charging for _each_ ride.
That's a non-issue. But you do have to stand in line, and in any
event, longer rides are known for being more enjoyable. The Magnum
is an excellent coaster that a lot of people still want to ride,
not because it was once the tallest and fastest in the world but
because the whole ride is enjoyable, not just the first hill. (The
Magnum was one of the rare coasters built for both bragging rights
_and_ memorable rider experience.)
It's more like this: want an enjoyable, memorable ride? There are
no shortage of coasters at the Point, so you won't have any trouble
finding such a ride. But one thing they didn't have (anymore) was
the tallest, steepest, fastest. And they haven't had that for too
many years; they _had_ to build a tallest-steepest-fastest before
very much longer, or their reputation would flag. You don't keep
a reputation for being the coaster capital of the world easily.
There are too many other places that want to catch up with that
(though no more than at most half a dozen places are really even
contenders, and none of them are within a day's drive).
> The incremental find text is a bit nicer than ctrl-F
It is; I just haven't got used to being able to do that in my web
browser yet. I hit Ctrl-F out of habbit. The other thing is, the
incremental search in Mozilla (typeahead find) is lacking some of
the features that would make it really useful. In particular, with
Emacs incremental search you can type the first part of what you're
looking for, hit Ctrl-S a couple of times to move forward through
a couple of instances, type a little more onto your search string,
hit Ctrl-S a couple more times, realise you went past the instance
you were interested in, hit Ctrl-R to go back to it... The feature
in Mozilla is not quite so mature in its implementation.
> since it doesn't drag up a little dialog that (in Mozilla 1.2
> under OSX) you can't close without using your mouse.
That's a Mac thing. You can't do Jack Squat(TM) on a Mac without
using a mouse. In fact, you can generally use a Mac better with
no keyboard than with no mouse. This has been true at _least_
since the days of System 6, and it's only any different with OS X
if you intend to spend most of your time in Terminal.app or using
X11 apps that have been ported over. On any other platform, _all_
dialogs can be closed without using a mouse. (On Win32 or Gnome,
there are no fewer than three distinct ways to close any dialog with
the keyboard.) This is nice; in the summer around here (though not
this last summer; we had a dry year) it often gets sufficiently
humid that the mouse is basically not usable, and I have a tendency
during August to set the mouse on a shelf and use the keyboard
pretty much exclusively until fall comes and dries things up. The _only_ thing I've discovered I can't do effeciently without a mouse
is image editing. (Though optical mice are getting pretty cheap
these days; this year when summer rolls around I might just get
one of those, and then if I need to use Gimp I can...)
> Do you believe cosmology is not a science because we can't
> recreate the big bang?
In a word, yes.
> No type-ahead? Let's see.. typing "sl".. ooh.. completes to
/
> "slashdot.org"
You're typing in the location bar. He was talking about type-ahead
find, which is roughly analogous to the incremental search feature
in Emacs (albeit less powerful in several major ways). With your
cursor in the page, you just start typing, and it searches forward
and finds the first occurrance of what you typed -- either in a
link (if you just start typing) or anywhere (if you start with a
before typing what you want to find). It's fairly useful, although
I must confess that I haven't gotten in the habbit yet of using it
in Mozilla; I still find myself hitting Ctrl-F and Ctrl-G a lot.