> Right, and we all know how much world class software has been > written by accountants, HR and marketing people. And how many VB > jockeys even know who Donald Knuth is. Spare me.
Accounting and HR are bad examples, because those jobs appeal to a completely different personality type than programming. Mixing a marketing job with programming might work out better, or some other field that actually interests you. If your education is sufficiently general, quite a few fields are open. Some are not; teaching pretty much requires either a degree in education (for primary or secondary) or in the field you want to teach (for post-secondary), for example. But there are many fields that are not so closed.
IT appeals to people, because it's fun work; it used to be that it paid very well anyway, because it was highly skilled; as the level of knowledge required to operate a computer and the level of computer knowledge possessed by people in other fields get closer together, it may be that IT work, being something that appeals to a lot of people, will not end up paying all that well. In ten years, it may be that plumbers make more than systems administrators. *shrug*. If that happens, you make your own decision about whether you'd rather follow the money or do IT work because you enjoy it.
> (How do you expect to broaden your knowledge in many subject > areas in 12 credit hours worth of work?)
It does depend very much on your major. My advise is to take a major that leaves room for some electives, so that you _can_ supplement the curriculum with stuff that got left out of the requirements. I chose Math (not Math Ed, just Math, the kind that involves modern algebra) and had plenty of room for electives. I minored in computer science, but I had room in my schedule to take extra computer courses, so that I ended up with as many credits toward my minor than toward my major, _plus_ still had room to supplement my education with two semesters of Koine Greek, a drawing class from the art department, an elective from the theology department, and a semester of astronomy. I could have chosen other electives instead -- extra history, for example -- but felt that the ones I chose did a good job of ballancing out the otherwise-lopsided gen-ed core, which was lacking in these areas. My point is, I was _able_ to do that ballancing, plus boost my minor a notch up, because my major left me room for some electives. If I'd majored in Math Ed, for example, I would have been stuck with a scheduled loaded with classes in my major and little flexibility.
It also helps to select a college where the core gen-ed classes are not a joke. I had a small handful of joke classes (College Life being the worst), but most of the gen-ed core at my school was good stuff, or at least pretty decent. (Of course, at _any_ school you get out what you put in to a large extent; slack off and shoot for a C- average, and NO school can give you the education that a motivated student extracts from an average school.)
Trying to deliver food to starving people in the third world is mostly a losing proposition -- not because we don't have food to spare, and not because they don't need food, but for more practical reasons that vary somewhat from area to area but start to look depressingly similar after a while. Mostly it has to do with what Bill Cosby calls "Brain Dammage".
The US government tried it in Somolia not very many years ago. Almost none of the food got to actual starving people; local thugs confiscated it so they could feed the armies they were using to oppress the people. (This was entirely predictable, for people who understand the third world.) We ended up getting involved militarily (yeah, more US forces in the third world, that sure makes us popular in the UN), but that didn't work so well either, and the instant our forces pulled out everything went back like it was. This was during the Clinton administration, and it was well-intentioned, but it just plain didn't work.
The US government isn't the only entity to ever try it, not by a long shot. Any number of church denominations have tried to set up an infrastructure for taking food to starving people; these experiments have all failed, and not for lack of food to take over.
GBIM (a missions organisation) concluded decades ago that providing education is okay, but providing physical goods brings out the worst in the people they are trying to help. They now have a standing policy against giving people physical stuff that is out of proportion to what they could get on their own. So they build church buildings out of local materiels now, instead of importing a nice one, and they don't hand out a lot of stuff. The reason providing education works better? Nobody's sure _exactly_. It's not because the people need it more than they need food and stuff; they need both. Mostly it's because starving people don't _fight_ over education. The really interesting thing is, it's something they want almost as much as they want food (in some places), but they behave differently to acquire it. The theory is that you can't steal or horde education because it takes too long to acquire, but others say it's because it isn't lost when shared. Whatever, it works: people behave more decently when you give them information than when you try to give them food.
Now, I'm not sure where computers would fall in. It's worth trying to see, but there's a distinct possibility they're going to fall into the same category with food, and that giving them out is going to prove to be impracticable. Of course, if that turns out to be the case you could retain the computers at the organisation and use them to provide training or whatnot.
If you want to avoid helping Microsoft, just make sure you train them on OSS.
Sun is different, for three reasons. Perhaps most obviously, Sun has been making server-grade systems since Apple was associated with green screens and keyboards built into the computer chassis, so they have a long-standing reputation for what Apple has only recently been trying to start doing. But traditions don't last forever, and the other two differences are more important. The second difference is that Sun doesn't just make their money only on hardware per se but on complete "solutions" (similar to what IBM does, but to perhaps a lesser extent). Sun (like DEC used to do) also works with vendors who base their "solutions" on Sun hardware and Solaris. (Think in terms of highly-industry-specific proprietary application suites that are sold as a package: hardware, software, training, and the maintenance contract all go together.) Apple AFAIK has to date not attempted to break into this market. (Microsoft has, and has gained a bit at the expense of Compaq (their DEC stuff) and others.) The third thing that keeps Apple from being quite comparable to Sun is that Apple has a much better relationship with Microsoft. Sun has been called a loose cannon, but Apple kisses whatever of Microsoft's anatomy is required to get some minimal cooperation. This is an attitude thing, but it is probably the most important of the three differences. Apple doubtless resents Microsoft's dominant position, but they work with it, calmly playing their cards and positioning themselves; Sun is more... ruthless. That's why you can get Solaris.x86, but Mac OS X86 is a rumour.
> MacOS aliases are far superior to symlinks IMHO.
In principle, maybe, but they are less well-supported. symlinks are supported universally. (Apps that don't understand what symlinks are just get the file itself from the filesystem, rather than the symlink; that includes apps ported from other platforms.)
> the alias has a much higher chance of identifying the proper > file should the original file move
Unix geeks don't mv files when there are symlinks pointing to them. Worrying about that is like worrying about whether a Mac user who is learning Linux/Gnome might screw something up by using the incorrect octal numbers while trying to chmod/var
> Wouldn't it be easier to buy a PC and run Windows?
In a word, no. It is easier to get _started_ that way, but over the long term it is not easier. And don't start yammering about file format incompatibilities, because I've had pretty rotten experiences trying to convert documents from one proprietary format to another for use on a different Windows system than the one on which they were created. I've been down that path, and I'm not going back. All my data now are in accessible formats and _staying_ that way.
> I'm on a Mac and yet my mouse has two buttons and a scroll wheel
Indeed: at the cost of Mac hardware, another few bucks for a decent mouse is like nothing, so do yourself a favour and get one. Logitech makes nice ones, and they even come in designs that look appropriately trendy (a.k.a. stupid) to match an iMac or eMac. I've always found it odd that Macs, which have such a good foothold in the graphics arts market don't come with a decent mouse; both Gimp and Photoshop are basically useless with a single-button mouse. I guess it's because graphics arts professionals have strong pointing-device preferences (I have a Mac-using artist and typesetting friend who swears by trackballs, for example) whereas the other portion of the Mac user segment would never use the second button and might find it confusing. *shrug*. It's not a big deal; the price of a mouse is nothing compared to the price of a Mac, so just get one.
Or this, which in addition to being very
portable also will revolutionise your maintainability. (Java is also
better than C/C++ in this regard, but IMO it doesn't go far enough
in moving away from the pitfalls of C/C++.) C-based langauges (C,
objective C, C++, C#, ++C, C^2, C ad nauseam) are fundamentally
outmoded, because the langauge tools place burdens on the developers
that in any modern development environment ought to be handled
automatically.
So you end up with buffer overruns, dereferencing pointers that no
longer point anywhere sensible, and other programming errors that
have plagued us since the sixties and are totally avoidable if you
just use a modern, high-level language.
(Note to Python fans: I used Perl as my example because it happens
to be a language I know and use; I did not say anything bad about
Python, only about C and its ilk, which you don't like either.)
> You think that's a side at which you're looking?
No, it's a surface. The other poster is correct: a sphere has no edges. In geometry, "edge" is techspeak for a line, arc, curve, or segment of one of those (I _think_ I covered the bases there) along which a surface meets a surface. A sphere does not need any edges because it only has one surface. (It is possible to devise an object that has only one surface yet has an edge, but it would not be a sphere, because the surface has to meet itself. It would also have vertices, at the ends of the edge. The edge could be either straight or curved. Either way, the surface would appear somewhat reminiscent of a cone near each vertex. I believe it is possible to have any number of these pinch-type edges on a single surface, provided no set of them link end-to-end in a complete circuit.)
Did you see that? 1968, Douglas Engelbart demonstrates computer windows and a wooden stylus he calls a mouse. 1968. Can you say "Microsoft vs Lindows trademark lawsuit"? How about 1968, can you say that? (I knew the concept was old, but I didn't know it was that old.)
> To a packed house at a computer conference in San Francisco, > Stanford Research Institute's Douglas Engelbart made a dramatic > presentation that included first-time demonstrations of onscreen > "windows," teleconferencing and a wooden stylus device he called > a "mouse." Engelbart didn't see much value in the peripheral, and > neither did Stanford Research, which owned the patent and later > licensed it to companies like Apple Computer for a $45,000 > one-time fee. Two decades later, Engelbart's in-vention was the > PC standard.
> I bet you didn't know you can still format XML tags with CSS
Yes, I did know; that's why in the demo I wrote this: > (Presumably, this is so the rendering engine for HTML and XHTML > can share a lot of information with the one for general XML.)
But in theory, if we were being strictly specification-complaint, that would only work in XML. The demo is served as text/html and does not have an xml version declaration (one of those funny things with the question marks beginning and end before the doctype (which also isn't there in the demo)). So it ought to be treated as HTML (or SGML), not XML. In theory.
> Both are originally ancient Greek, not Latin as one might think.
But alpha is a plural suffix in Greek, too... neuter nom/acc...
> The plural of schema is schemata
Oh, duh, I see it now; it's third declension, and the a isn't a suffix at all; the root ends in t, which drops off in the nominative singular where there's no ending. Why didn't I see that before?
You can have wellformed XML without them, but there must be a DTD or Scheme in order to have _valid_ XML.
> The url used in a namespace declaration doesn't need to > correspond to a real document
Or, more to the point, the document at that URL can be an inside joke from the movie Ghostbusters, rather than having any actual declarations. (Those of you who think I am kidding on this point have never tried to access the document that the XUL namespace declaration points to.) This, however, is not really important.
> Even in case the document used a DTD or Schema To be valid it has to... anyway, even if it doesn't, there is one implicit.
> that DTD or Scheme were available The availability of the DTD or Schema[1] is really not important. It would be easy enough to write a program that analyses documents that are known to be valid and keeps track of which tags contain data, and which ones contain PCDATA, and which other tags they have nested in them. Analyse enough documents, and you have a subset of the original DTD that's good enough for creating documents that are guaranteed to be compatible and can use all the features used by the documents you analysed.
> you still don't know what the hell the tags mean
Of all the points you made, this is the important one. XML is by its very nature a very flexible standard. It's not like HTML where a formal standard specifies that <p> is a paragraph and that it is a block-level element with certain amounts of white space top and bottom and so on and so forth. The tags and attributes an the format can be interpreted in whatever way the application sees fit.
In practice, that means another word-processing app can with relative ease use the same format in such a way that tools for searching and indexing will work on documents created by both apps, and it means that if you open a Word document in whatever other app that uses that format you can make minor changes (such as wording changes) and save it, and when Word opens it again it won't be munged (assuming the other app does things in a sane manner that preserves whatever markup it doesn't understand). But it does NOT mean that the doc will necessarily look the same in the other app as it does in Word.
[1] And when did "schema" become singular, anyhow?
> I think an analogy to Frontpage is appropriate here. Sure, > it produces HTML
No, it doesn't. It produces something that looks vaguely similar to HTML, perhaps, but HTML it is not. You look at a FrontPage document's source closely, and you see a mishmash of deprecated HTML3 markup, newer markup that didn't exist in HTML3 but was introduced later, plus the occasional attribute that never existed in _any_ version of HTML, thrown in for good measure.
It is only because of the long-standing practice of browsers since Mosaic (possibly before) to ignore any tag or attribute they don't understand that a FrontPage document will display at all in any browser. (This is fun to try sometime: make up a tag, completely out of thin air, and use it in a webpage, and see how various browsers handle the page.)
<voice id="Linus" rate="slow">I pronounce Linux as Linux</voice> Any browser will display the quote as if the voice tags weren't there at all -- does that make it HTML?
> User's need to be in the habit of locking the workstation when > they leave it.
That would be good for security, certainly...
> A good IT department will audit this (at least for the users > that reside in the office... that goes for plain-view passwords, > etc) and penalize users who do not (give them a slow POS or > something with a ton of dead pixels).
The IT department does not always have the authority to do this. For that matter, the IT department doesn't always have the authority to require passwords to be changed annually (or to change them), let alone penalising anyone in any way.
> Then you have the frequency the signal is broadcasted on > randomly shuffled based on the current time.
You have assumed that the repeaters can't just blindly repeat all frequencies. (I'm not sure how they'd do that, but if they did, it would foil your frequency-switching encryption.)
If you need the thing to transmit a signal via radio waves, then I think you're probably right. However, there _are_ passive (as in non-powered, like the other poster was asking) one-way cryptographic devices. I read an article once (possibly on/., possibly elsewhere) about some people who had taken bits of glass and embedded them in a card-shaped slab of something-or-another, so that shining light through it from various angles would produce various patterns. It was said to be impossible to work backwords from the speckle-patterns it produced and create a copy of the thing, so if the authenticator picked a different angle each time to shine the light through, it could be assured that the correct pattern could only be produced by the original "key".
However, any object-key system like this doesn't prevent somebody from just stealing the key object along with the thing it unlocks. It's fine for things like the article discussed (preventing random people in a hospital from reading patient records when the doctor steps away from the computer), but it would not work in a case involving someone actively seeking your data (e.g., espionage).
This won't put any effective pressure on Apple; I'm pretty sure they don't even care that Gnu-Darwin exists. What it will do is make life painful for normal users and reduce the amount of positive exposure some people get to open software. Gnu-Darwin is the equivalent for the Apple platform of Cygwin on the Windows side. Dropping it is like saying "you can't use our free software unless you switch operating systems". Huh? I thought it was free?
Now, if Gnu-Darwin was relying on some libraries that weren't properly license-compatible, then that needed to be fixed... but _dropping_ Gnu-Darwin entirely isn't the way to do that. They could have temporarily pulled it, if they'd announced it in a way that said, basically, "we goofed and are fixing the problem, it was a licensing issue; we were linking against something that isn't license-compatible", but trying to blame this on Apple is like blaming the weather for making you cold when you forgot to wear a coat.
> As an added bonus, I'm running it swiftly and happily on > an old PII/300.
This is impressive _how_?
I'm running on a PII/233, and regardless of which browser I use (Mozilla, Netscape 7, Opera, Phoenix, Konqueror, Amaya, Arachne, Netscape 4 (ick), K-Meleon, Galeon,... whatever) the speed is pretty much exactly the same -- and unless I'm doing something in the background that's lot more CPU-intensive than web browsing, my CPU-utilisation-meter almost never goes past 50%.
Web browsing speed depends almost 100% on three things: bandwidth, RAM, and latency (in that order). CPU speed, unless you're trying to use a 486 (or worse), is a complete non-issue.
The whole "Opera is fast" argument just doesn't fly with me. Opera loads pages in the same amount of time as any other browser. The only way to speed it up substantially would be to not retrieve some of the content (such as images and plugins, perhaps), but almost any browser can do that if that's the effect you're after.
I have Opera, and I use it from time to time (mostly for testing pages to see how they look in it), and I'm unable to perceive any increase in speed over other browsers.
Opera does have a smaller footprint than the big boys, but that's a separate issue; smaller footprint only means faster if you're so low on system resources that you're using a swapfile, and if that's the case you've got bigger problems than your web browser.
> Yes, users can choose their own wallpaper etc. - but what's > wrong with that?
Depends, but if you have an account where you don't want anything changed (such as a guest account), that's easy too: set it up like you want, make a tarball of the user's home directory, and set up a cron job that untars it overtop of whatever is there.
> Yeah, I'd say that learning Windows (aside from administration) > is really just learning an application: explorer.
Um, have you ever tried to administer a Windows box? Knowing Explorer is what you take for granted; it's the undocumented stuff that you have to know to survive. You're dead in the water if you aren't comfortable with the registry, for example. First time any problem crops up, you'd best know how to work with cabinets, and which undocumented batch files that get created by install processes are run on startup and, if broken, have to be deleted in order to restore the system to a bootable state. (And no, I'm not talking about AUTOEXEC.BAT; if you thought that was what I meant, you'll end up formatting the drive the first time anything goes wrong, but not until after you pull out your hair first.)
The difference between Windows and Linux is not one of complexity; Windows and Linux have roughly the same amount of complexity. The difference is one of documentation: Linux has some. (The other difference is consistency in terms of the visual appearance of UI widgets; almost all Windows apps use the same widget set. (That's a good thing.) RedHat is working on this problem, but their solution is incomplete at this time.)
As long as they have unsupervised physical access to the system, they can always circumvent it. Ultimately, if there is no other way, they can set the BIOS-forget jumper to wipe any CMOS password, set it to boot from a removable drive, and then have their way with the MBR and the boot sector of the boot partition. In almost all cases, there's a much easier way that doesn't involve opening the case. DeepFreeze is, from what I'm told, good enough that if you have no bootable removable drives, set the BIOS password, and can keep them from opening the case you won't have much trouble -- but you are always taking the risk that the teacher or lab assistant will step out of the room for too long and some clown will set the BIOS jumper and have his little fun. (Having no removable drives goes a long way toward making this harder, but that isn't always practical.)
The better solution is to go with thin clients. Then all they can do is steal the thin clients, but without getting into the server room, that's the limit. You hook up a new thin client, and it's as if nothing happened. (This assumes the thin-client server is secure from network-based attacks; I suggest not using a Microsoft solution on the server end, and don't use your thin client server for serving other things like mail, either; spend the $50 on ebay and get yourself an old system you can make into a separate mail server, if it comes to that.)
Seriously: a thin-client solution takes more setup, but once you have it in place, your headaches are greatly reduced. The only downside is a Single Point Of Failure, which is another reason you don't use a Microsoft solution on the server end.
Slashdot Mirror
> Right, and we all know how much world class software has been
> written by accountants, HR and marketing people. And how many VB
> jockeys even know who Donald Knuth is. Spare me.
Accounting and HR are bad examples, because those jobs appeal to a
completely different personality type than programming. Mixing a
marketing job with programming might work out better, or some other
field that actually interests you. If your education is sufficiently
general, quite a few fields are open. Some are not; teaching pretty
much requires either a degree in education (for primary or secondary)
or in the field you want to teach (for post-secondary), for example.
But there are many fields that are not so closed.
IT appeals to people, because it's fun work; it used to be that it
paid very well anyway, because it was highly skilled; as the level
of knowledge required to operate a computer and the level of computer
knowledge possessed by people in other fields get closer together,
it may be that IT work, being something that appeals to a lot of
people, will not end up paying all that well. In ten years, it may
be that plumbers make more than systems administrators. *shrug*.
If that happens, you make your own decision about whether you'd
rather follow the money or do IT work because you enjoy it.
> (How do you expect to broaden your knowledge in many subject
> areas in 12 credit hours worth of work?)
It does depend very much on your major. My advise is to take a major
that leaves room for some electives, so that you _can_ supplement the
curriculum with stuff that got left out of the requirements. I chose
Math (not Math Ed, just Math, the kind that involves modern algebra)
and had plenty of room for electives. I minored in computer science,
but I had room in my schedule to take extra computer courses, so that
I ended up with as many credits toward my minor than toward my major,
_plus_ still had room to supplement my education with two semesters
of Koine Greek, a drawing class from the art department, an elective
from the theology department, and a semester of astronomy. I could
have chosen other electives instead -- extra history, for example --
but felt that the ones I chose did a good job of ballancing out the
otherwise-lopsided gen-ed core, which was lacking in these areas.
My point is, I was _able_ to do that ballancing, plus boost my minor
a notch up, because my major left me room for some electives. If
I'd majored in Math Ed, for example, I would have been stuck with
a scheduled loaded with classes in my major and little flexibility.
It also helps to select a college where the core gen-ed classes are
not a joke. I had a small handful of joke classes (College Life
being the worst), but most of the gen-ed core at my school was good
stuff, or at least pretty decent. (Of course, at _any_ school
you get out what you put in to a large extent; slack off and shoot
for a C- average, and NO school can give you the education that a
motivated student extracts from an average school.)
Trying to deliver food to starving people in the third world is mostly
a losing proposition -- not because we don't have food to spare, and
not because they don't need food, but for more practical reasons that
vary somewhat from area to area but start to look depressingly similar
after a while. Mostly it has to do with what Bill Cosby calls "Brain
Dammage".
The US government tried it in Somolia not very many years ago.
Almost none of the food got to actual starving people; local thugs
confiscated it so they could feed the armies they were using to
oppress the people. (This was entirely predictable, for people
who understand the third world.) We ended up getting involved
militarily (yeah, more US forces in the third world, that sure
makes us popular in the UN), but that didn't work so well either,
and the instant our forces pulled out everything went back like
it was. This was during the Clinton administration, and it was
well-intentioned, but it just plain didn't work.
The US government isn't the only entity to ever try it, not by a
long shot. Any number of church denominations have tried to set
up an infrastructure for taking food to starving people; these
experiments have all failed, and not for lack of food to take over.
GBIM (a missions organisation) concluded decades ago that providing
education is okay, but providing physical goods brings out the
worst in the people they are trying to help. They now have a
standing policy against giving people physical stuff that is out
of proportion to what they could get on their own. So they build
church buildings out of local materiels now, instead of importing
a nice one, and they don't hand out a lot of stuff. The reason
providing education works better? Nobody's sure _exactly_. It's
not because the people need it more than they need food and stuff;
they need both. Mostly it's because starving people don't _fight_
over education. The really interesting thing is, it's something
they want almost as much as they want food (in some places), but
they behave differently to acquire it. The theory is that you can't
steal or horde education because it takes too long to acquire, but
others say it's because it isn't lost when shared. Whatever, it
works: people behave more decently when you give them information
than when you try to give them food.
Now, I'm not sure where computers would fall in. It's worth trying
to see, but there's a distinct possibility they're going to fall
into the same category with food, and that giving them out is going
to prove to be impracticable. Of course, if that turns out to be the
case you could retain the computers at the organisation and use them
to provide training or whatnot.
If you want to avoid helping Microsoft, just make sure you train
them on OSS.
> Sun is another example of this, I believe.
Sun is different, for three reasons. Perhaps most obviously, Sun
has been making server-grade systems since Apple was associated
with green screens and keyboards built into the computer chassis,
so they have a long-standing reputation for what Apple has only
recently been trying to start doing. But traditions don't last
forever, and the other two differences are more important. The
second difference is that Sun doesn't just make their money only on
hardware per se but on complete "solutions" (similar to what IBM
does, but to perhaps a lesser extent). Sun (like DEC used to do)
also works with vendors who base their "solutions" on Sun hardware
and Solaris. (Think in terms of highly-industry-specific proprietary
application suites that are sold as a package: hardware, software,
training, and the maintenance contract all go together.) Apple
AFAIK has to date not attempted to break into this market.
(Microsoft has, and has gained a bit at the expense of Compaq (their
DEC stuff) and others.) The third thing that keeps Apple from being
quite comparable to Sun is that Apple has a much better relationship
with Microsoft. Sun has been called a loose cannon, but Apple
kisses whatever of Microsoft's anatomy is required to get some
minimal cooperation. This is an attitude thing, but it is probably
the most important of the three differences. Apple doubtless resents
Microsoft's dominant position, but they work with it, calmly playing
their cards and positioning themselves; Sun is more... ruthless.
That's why you can get Solaris.x86, but Mac OS X86 is a rumour.
> MacOS aliases are far superior to symlinks IMHO.
/var
In principle, maybe, but they are less well-supported. symlinks
are supported universally. (Apps that don't understand what
symlinks are just get the file itself from the filesystem, rather
than the symlink; that includes apps ported from other platforms.)
> the alias has a much higher chance of identifying the proper
> file should the original file move
Unix geeks don't mv files when there are symlinks pointing to them.
Worrying about that is like worrying about whether a Mac user who
is learning Linux/Gnome might screw something up by using the
incorrect octal numbers while trying to chmod
> Wouldn't it be easier to buy a PC and run Windows?
In a word, no. It is easier to get _started_ that way, but over the
long term it is not easier. And don't start yammering about file
format incompatibilities, because I've had pretty rotten experiences
trying to convert documents from one proprietary format to another
for use on a different Windows system than the one on which they
were created. I've been down that path, and I'm not going back.
All my data now are in accessible formats and _staying_ that way.
> I'm on a Mac and yet my mouse has two buttons and a scroll wheel
Indeed: at the cost of Mac hardware, another few bucks for a
decent mouse is like nothing, so do yourself a favour and get one.
Logitech makes nice ones, and they even come in designs that look
appropriately trendy (a.k.a. stupid) to match an iMac or eMac. I've
always found it odd that Macs, which have such a good foothold in the
graphics arts market don't come with a decent mouse; both Gimp and
Photoshop are basically useless with a single-button mouse. I guess
it's because graphics arts professionals have strong pointing-device
preferences (I have a Mac-using artist and typesetting friend who
swears by trackballs, for example) whereas the other portion of the
Mac user segment would never use the second button and might find
it confusing. *shrug*. It's not a big deal; the price of a mouse
is nothing compared to the price of a Mac, so just get one.
Or this, which in addition to being very portable also will revolutionise your maintainability. (Java is also better than C/C++ in this regard, but IMO it doesn't go far enough in moving away from the pitfalls of C/C++.) C-based langauges (C, objective C, C++, C#, ++C, C^2, C ad nauseam) are fundamentally outmoded, because the langauge tools place burdens on the developers that in any modern development environment ought to be handled automatically. So you end up with buffer overruns, dereferencing pointers that no longer point anywhere sensible, and other programming errors that have plagued us since the sixties and are totally avoidable if you just use a modern, high-level language. (Note to Python fans: I used Perl as my example because it happens to be a language I know and use; I did not say anything bad about Python, only about C and its ilk, which you don't like either.)
> You think that's a side at which you're looking?
No, it's a surface. The other poster is correct: a sphere has no
edges. In geometry, "edge" is techspeak for a line, arc, curve,
or segment of one of those (I _think_ I covered the bases there)
along which a surface meets a surface. A sphere does not need any
edges because it only has one surface. (It is possible to devise
an object that has only one surface yet has an edge, but it would
not be a sphere, because the surface has to meet itself. It would
also have vertices, at the ends of the edge. The edge could be
either straight or curved. Either way, the surface would appear
somewhat reminiscent of a cone near each vertex. I believe it is
possible to have any number of these pinch-type edges on a single
surface, provided no set of them link end-to-end in a complete
circuit.)
Did you see that? 1968, Douglas Engelbart demonstrates computer
windows and a wooden stylus he calls a mouse. 1968. Can you say
"Microsoft vs Lindows trademark lawsuit"? How about 1968, can you
say that? (I knew the concept was old, but I didn't know it was
that old.)
> To a packed house at a computer conference in San Francisco,
> Stanford Research Institute's Douglas Engelbart made a dramatic
> presentation that included first-time demonstrations of onscreen
> "windows," teleconferencing and a wooden stylus device he called
> a "mouse." Engelbart didn't see much value in the peripheral, and
> neither did Stanford Research, which owned the patent and later
> licensed it to companies like Apple Computer for a $45,000
> one-time fee. Two decades later, Engelbart's in-vention was the
> PC standard.
> I bet you didn't know you can still format XML tags with CSS
Yes, I did know; that's why in the demo I wrote this:
> (Presumably, this is so the rendering engine for HTML and XHTML
> can share a lot of information with the one for general XML.)
But in theory, if we were being strictly specification-complaint,
that would only work in XML. The demo is served as text/html and
does not have an xml version declaration (one of those funny things
with the question marks beginning and end before the doctype (which
also isn't there in the demo)). So it ought to be treated as HTML
(or SGML), not XML. In theory.
> Both are originally ancient Greek, not Latin as one might think.
But alpha is a plural suffix in Greek, too... neuter nom/acc...
> The plural of schema is schemata
Oh, duh, I see it now; it's third declension, and the a isn't a
suffix at all; the root ends in t, which drops off in the nominative
singular where there's no ending. Why didn't I see that before?
I learned something today.
> You don't need a DTD or Schema to have XML
You can have wellformed XML without them, but there must be a
DTD or Scheme in order to have _valid_ XML.
> The url used in a namespace declaration doesn't need to
> correspond to a real document
Or, more to the point, the document at that URL can be an inside
joke from the movie Ghostbusters, rather than having any actual
declarations. (Those of you who think I am kidding on this point
have never tried to access the document that the XUL namespace
declaration points to.) This, however, is not really important.
> Even in case the document used a DTD or Schema
To be valid it has to... anyway, even if it doesn't, there
is one implicit.
> that DTD or Scheme were available
The availability of the DTD or Schema[1] is really not important.
It would be easy enough to write a program that analyses documents
that are known to be valid and keeps track of which tags contain
data, and which ones contain PCDATA, and which other tags they
have nested in them. Analyse enough documents, and you have a
subset of the original DTD that's good enough for creating
documents that are guaranteed to be compatible and can use all
the features used by the documents you analysed.
> you still don't know what the hell the tags mean
Of all the points you made, this is the important one. XML is
by its very nature a very flexible standard. It's not like HTML
where a formal standard specifies that <p> is a paragraph and
that it is a block-level element with certain amounts of white
space top and bottom and so on and so forth. The tags and
attributes an the format can be interpreted in whatever way
the application sees fit.
In practice, that means another word-processing app can with
relative ease use the same format in such a way that tools for
searching and indexing will work on documents created by both apps,
and it means that if you open a Word document in whatever other app
that uses that format you can make minor changes (such as wording
changes) and save it, and when Word opens it again it won't be
munged (assuming the other app does things in a sane manner that
preserves whatever markup it doesn't understand). But it does NOT
mean that the doc will necessarily look the same in the other app
as it does in Word.
[1] And when did "schema" become singular, anyhow?
Heh. That was me too. Actually, I _do_ use CUPS at work, but only
the client part of it; I never turned it on as a server, so...
I've worked up an even better demonstration
> I think an analogy to Frontpage is appropriate here. Sure,
> it produces HTML
No, it doesn't. It produces something that looks vaguely similar
to HTML, perhaps, but HTML it is not. You look at a FrontPage
document's source closely, and you see a mishmash of deprecated
HTML3 markup, newer markup that didn't exist in HTML3 but was
introduced later, plus the occasional attribute that never
existed in _any_ version of HTML, thrown in for good measure.
It is only because of the long-standing practice of browsers since
Mosaic (possibly before) to ignore any tag or attribute they don't
understand that a FrontPage document will display at all in any
browser. (This is fun to try sometime: make up a tag, completely
out of thin air, and use it in a webpage, and see how various
browsers handle the page.)
<voice id="Linus" rate="slow">I pronounce Linux as Linux</voice>
Any browser will display the quote as if the voice tags weren't
there at all -- does that make it HTML?
> User's need to be in the habit of locking the workstation when
> they leave it.
That would be good for security, certainly...
> A good IT department will audit this (at least for the users
> that reside in the office... that goes for plain-view passwords,
> etc) and penalize users who do not (give them a slow POS or
> something with a ton of dead pixels).
The IT department does not always have the authority to do this.
For that matter, the IT department doesn't always have the authority
to require passwords to be changed annually (or to change them), let
alone penalising anyone in any way.
> Then you have the frequency the signal is broadcasted on
> randomly shuffled based on the current time.
You have assumed that the repeaters can't just blindly repeat
all frequencies. (I'm not sure how they'd do that, but if they
did, it would foil your frequency-switching encryption.)
If you need the thing to transmit a signal via radio waves, then I /., possibly elsewhere)
think you're probably right. However, there _are_ passive (as in
non-powered, like the other poster was asking) one-way cryptographic
devices. I read an article once (possibly on
about some people who had taken bits of glass and embedded them in
a card-shaped slab of something-or-another, so that shining light
through it from various angles would produce various patterns. It
was said to be impossible to work backwords from the speckle-patterns
it produced and create a copy of the thing, so if the authenticator
picked a different angle each time to shine the light through, it
could be assured that the correct pattern could only be produced
by the original "key".
However, any object-key system like this doesn't prevent somebody
from just stealing the key object along with the thing it unlocks.
It's fine for things like the article discussed (preventing random
people in a hospital from reading patient records when the doctor
steps away from the computer), but it would not work in a case
involving someone actively seeking your data (e.g., espionage).
This won't put any effective pressure on Apple; I'm pretty sure they
don't even care that Gnu-Darwin exists. What it will do is make
life painful for normal users and reduce the amount of positive
exposure some people get to open software. Gnu-Darwin is the
equivalent for the Apple platform of Cygwin on the Windows side.
Dropping it is like saying "you can't use our free software unless
you switch operating systems". Huh? I thought it was free?
Now, if Gnu-Darwin was relying on some libraries that weren't
properly license-compatible, then that needed to be fixed... but
_dropping_ Gnu-Darwin entirely isn't the way to do that. They
could have temporarily pulled it, if they'd announced it in a
way that said, basically, "we goofed and are fixing the problem,
it was a licensing issue; we were linking against something that
isn't license-compatible", but trying to blame this on Apple is
like blaming the weather for making you cold when you forgot to
wear a coat.
> As an added bonus, I'm running it swiftly and happily on
... whatever) the speed is
> an old PII/300.
This is impressive _how_?
I'm running on a PII/233, and regardless of which browser I use
(Mozilla, Netscape 7, Opera, Phoenix, Konqueror, Amaya, Arachne,
Netscape 4 (ick), K-Meleon, Galeon,
pretty much exactly the same -- and unless I'm doing something in
the background that's lot more CPU-intensive than web browsing,
my CPU-utilisation-meter almost never goes past 50%.
Web browsing speed depends almost 100% on three things: bandwidth,
RAM, and latency (in that order). CPU speed, unless you're trying
to use a 486 (or worse), is a complete non-issue.
The whole "Opera is fast" argument just doesn't fly with me. Opera
loads pages in the same amount of time as any other browser. The
only way to speed it up substantially would be to not retrieve some
of the content (such as images and plugins, perhaps), but almost any
browser can do that if that's the effect you're after.
I have Opera, and I use it from time to time (mostly for testing
pages to see how they look in it), and I'm unable to perceive any
increase in speed over other browsers.
Opera does have a smaller footprint than the big boys, but that's
a separate issue; smaller footprint only means faster if you're so
low on system resources that you're using a swapfile, and if that's
the case you've got bigger problems than your web browser.
> Yes, users can choose their own wallpaper etc. - but what's
> wrong with that?
Depends, but if you have an account where you don't want anything
changed (such as a guest account), that's easy too: set it up
like you want, make a tarball of the user's home directory, and
set up a cron job that untars it overtop of whatever is there.
> Yeah, I'd say that learning Windows (aside from administration)
> is really just learning an application: explorer.
Um, have you ever tried to administer a Windows box? Knowing
Explorer is what you take for granted; it's the undocumented stuff
that you have to know to survive. You're dead in the water if you
aren't comfortable with the registry, for example. First time any
problem crops up, you'd best know how to work with cabinets, and
which undocumented batch files that get created by install processes
are run on startup and, if broken, have to be deleted in order to
restore the system to a bootable state. (And no, I'm not talking
about AUTOEXEC.BAT; if you thought that was what I meant, you'll
end up formatting the drive the first time anything goes wrong, but
not until after you pull out your hair first.)
The difference between Windows and Linux is not one of complexity;
Windows and Linux have roughly the same amount of complexity. The
difference is one of documentation: Linux has some. (The other
difference is consistency in terms of the visual appearance of UI
widgets; almost all Windows apps use the same widget set. (That's
a good thing.) RedHat is working on this problem, but their
solution is incomplete at this time.)
As long as they have unsupervised physical access to the system, they
can always circumvent it. Ultimately, if there is no other way, they
can set the BIOS-forget jumper to wipe any CMOS password, set it to
boot from a removable drive, and then have their way with the MBR and
the boot sector of the boot partition. In almost all cases, there's
a much easier way that doesn't involve opening the case. DeepFreeze
is, from what I'm told, good enough that if you have no bootable
removable drives, set the BIOS password, and can keep them from
opening the case you won't have much trouble -- but you are always
taking the risk that the teacher or lab assistant will step out of
the room for too long and some clown will set the BIOS jumper and
have his little fun. (Having no removable drives goes a long way
toward making this harder, but that isn't always practical.)
The better solution is to go with thin clients. Then all they can
do is steal the thin clients, but without getting into the server
room, that's the limit. You hook up a new thin client, and it's as
if nothing happened. (This assumes the thin-client server is secure
from network-based attacks; I suggest not using a Microsoft solution
on the server end, and don't use your thin client server for serving
other things like mail, either; spend the $50 on ebay and get
yourself an old system you can make into a separate mail server, if
it comes to that.)
Seriously: a thin-client solution takes more setup, but once you
have it in place, your headaches are greatly reduced. The only
downside is a Single Point Of Failure, which is another reason
you don't use a Microsoft solution on the server end.
I'm holding out for residential T1. I'll get Cowboy Neal to pay for it.