Since users tend to treat Firefox as sort of a mini-desktop environment of its own (at least I do), I guess the "right" behavior when extension XYZ crashes would be for Firefox to toss up a simple alert saying, "Extension XYZ just hurled." Possible options:
Keep going without it
Try to restart the extension (is this even feasible with current codebase?)
Restart Firefox (I know this capability is built in)
E-mail a traceback to the developer (this could actually work really well if it was integrated into the extension system)
For memory limits, I would think that should be handled by the OS, or Firefox talking directly to the OS, not the user. So when malloc() finally fails, it would be nice for Firefox to squash the extension that called it (with the above notification), rather than seppuku. I guess. Because who knows, maybe there is or will be an extension that actually should be allowed to use insane amounts of memory -- like an image or video editor, maybe.
Expanding on the FF-as-OS idea, it would be nice to have extensions that work like kill, nice and top, as another poster mentioned.
Yep, there are lots of ways they could have spared themselves a few years of C++-induced pain. The root cause of the problem, IMHO, is that when they scrapped the original unusably bloated and crufty codebase, for the rewrite they restricted themselves to a very archaic subset of C++ -- for portability's sake, apparently. The restrictions went too far, though, so things like exceptions, some of the STL (auto_ptr!), even C++-style comments (// instead of/* */) were forbidden in the Moz Dev guidelines. (IIRC.) There are plenty of GC libraries out there for C++; the Boost libraries include that and many other very helpful tools -- but introducing that would be a big change to an enormous codebase. Recently they've seen the value in using newer features, particularly exceptions, and there's an effort towards simplifying the code base using these features now.
Historical note: One of the biggest ways Netscape originally doomed itself was by attempting to rewrite it all in Java. There's no silver bullet, but you don't have to shoot yourself in the foot, either.
I think any of the lighter Gecko browsers would be decent replacements for Firefox. Kazehakase or Epiphany are pretty zippy, and they might have the added benefit of being harder for the user to screw up (I'm bracing myself for the day when Firefox extensions become a serious vector for malware). Fluxbuntu, when it's done, says it's going to use K. by default, but with the user-agent string set to Mozilla to avoid trouble on sites that try to be clever. In my experience, Gnome or KDE run reasonably well even on 256 MB of RAM, but Firefox requires a multiple of that before its responsiveness becomes sane. 512 might be the bare minimum for a decent FF experience.
On the other hand, I think the whole point of gOS is to let the browser be the entire desktop. They rely on Google apps instead of OpenOffice etc., so the rest of the system is simply a minimal environment for using Firefox -- which would certainly ease the transition for Windows users who already use Firefox.
This is terrible. Do you know how web servers work? These are the steps:
You type wikipedia.org/slashdot into your address bar.
The browser formats the input in a reasonable way ("http://wikipedia.org/slashdot") then splits the request into parts -- here, it uses the HTTP protocol to send a request to whatever server the DNS says resides at wikipedia.org. The rest of the URL is sent as an input string to the server.
The server receives that raw string -- "slashdot" -- then parses it and figures out what to do with it based on the contents of the string. Some options are (configurable):
If the string looks like "index.html" or "blog/recent.html", assume that's a relative path to a file on the server. Load that file and spit it out with an appropriate header.
If the string is formatted like "comments.pl?sid=1234&op=Reply&...", that's a standardized form for a CGI query, and everything after the ? can be reliably parsed and treated as arguments to a script. Note that "http://www.google.com/search?q=slashdot" has this same form. There is absolutely no difference between a script that searches a database about local pages (e.g. PHP's or Wikipedia's), and one that searches a database about external pages (e.g. Google's). They're both websites, and they work exactly the same way -- it doesn't matter what's in the database that's being queried. It's a script, extracting input from the URL, querying a local database and returning the results.
If the string doesn't correspond to a local file, and it doesn't look like a reasonable CGI query either, the server can take some default action. If the site doesn't want users monkeying around with URLs, they're return a 404 error (file not found). Slightly nicer is to return a 404 page with some helpful info on it, like a site map. Another widely used default option is to take that string as input to another script -- some sites might print the same faulty string or address in the help text, like "Did you mean to search for 'slashdot'?" Or, if they don't mind surprising the user, they can just run a search using that string. That last option, widely used already, is what's being patented here.
Wikipedia does users the courtesy of printing a 404 page before redirecting to a page with search results for the faulty URL. Normally, though, a site prefers to let users (and bots) know when a given URL doesn't correspond to a real page. Notice that amazon.com/slashdot returns a 404 page instead of a search!
Just because this has been done for a long time doesn't mean it's good practice. Normally, the browser on the client side does the task of letting the user run a search using a default (or selected) search engine, then diligently passing a well-formed URL on to the internet, saving the destination server some effort. But it's always been possible to do the extra work on the server side, and there are existing standards indicating how to do it.
And why don't the major web search engines already do this, if it's so user-friendly? Because they're web search engines! They exist because URLs aren't a very user-friendly way to get around online! Google doesn't want you to use the address bar, they want you to download their toolbar and rely on their web apps to find what you're looking for. They're training their users.
In Google's case, they also probably don't want so many people accidentally stumbling across their own pages while searching for things like "ig" and "search".
Windows installations are known to decay over time. Doing a complete wipe and reinstallation of Windows usually brings the computer back into usable condition, but most users have reservations about doing that. It's easier to just get a new computer -- or install the latest Ubuntu.
Why does Windows rot?
- Package management: Even when you run a program's designated uninstaller, or go through Add/Remove Programs, almost everything will leave some trace of itself behind in the registry. The new, approved way to do configuration on Windows is the same as the *nix way, with config files in the user's home folder, but nobody can force vendors to behave. Installing a new program can sometimes overwrite an existing DLL used by other programs. Plus, as part of the Windows Experience (TM), programs tend to try to modify or at least detect each other on the system, and adapt accordingly.
Sprinkle bugs into each of these areas, and you get a system that gradually develops quirks over time, slows down while looking for items that no longer exist or have been moved, and loads services that don't need to be there. Macs have a cute, tidy package system that resolves this somewhat; *nix systems usually have very powerful and thorough package managers; and both use flat text files instead of a single, binary, monolithic registry for handling program configuration.
- Commercial software: Linus is right in saying that open source is a better engineering tactic. Software for Windows is sometimes great, but usually terrible. Not only is it written hastily due to market pressure, it's dictated by managers with impure motives (e.g. marketing, add flash at the expense of quality, add a pet feature), sometimes obfuscated by coders who either don't know any better or don't want anyone else touching their code, occasionally dependent on Windows bugs or deprecated features, and frequently designed to induce vendor lock-in and force frequent upgrades.
Microsoft has always tried to provide a flexible playground for commercial developers, and until very recently would allow completely pathological things to happen on Windows if that's what was required for existing important applications to run. All of these things are limited in open-source software.
- Too much magic: The Windows Experience (TM) is marked by discoverability combined with endless kludges. There are no man pages. Instead, if you want to find out how to change some property of the system, right-click around the thing you want to change. Poke through the context menus, bring up a tabbed dialog, click around, check some boxes, click OK a few times to get out of your layers of dialogs. Job well done.
To make this work, there are multiple separate program-defined settings affecting each property of the system. Examples: sound, web browser settings, task bar items, quick launch bar, system tray. Things appear and get modified, but it's not clear what the cause was. So opening some random app for the first time in 6 months might cause the internet access settings to be updated, change the default browser from Firefox to IE, and add an item to the quick launch bar. Norton Antivirus will then detect some of this, vomit on your screen, and begin blocking Firefox from accessing the internet.
On *nix, modularity takes precedence, and configurations are managed individually. So, if you do screw up your system, you generally have just one or two places to check to see what happened (the config file, and possibly a GUI front-end for it), and it's less likely you'll change a setting by accident.
Good point. I think Intel feared Linux until just a few years ago, because the WinTel situation was such a cash cow, but then came to appreciate the idea of being partially independent of Microsoft for its business model. Plus, Linux brought in the idea of companies running critical server operations on Intel's x86 processors, rather than big iron or SPARC.
Starting with "parallelepiped" probably wasn't the best way to explain the determinant to non-mathematicians.
Fortunately, it works in two dimensions, too. Draw a parallelogram:
A_____B
/ *** / C_____D
Get the coordinates of each of the vertices: A = (1, 3) B = (5, 3) C = (0, 0) D = (4, 0)
The area of the parallelogram is the base (4) times the height (3) = 12. Geometry.
Now pick any two sides and call them your vectors -- say, CA and CD. These two vectors span the parallelogram. CA = [1 3] CD = [4 0]
Then the area of the parallelogram ABCD is also the (absolute value of the) determinant of the matrix created by placing those two vectors in columns: | 1 4 | | 3 0 |
det = | 1*0 - 3*4 | = | -12 | = 12
Same number. Linear algebra.
To get to what the previous poster said, picture exactly the same thing in three dimensions instead of two. A parallelepiped is a three-dimensional parallelogram, more or less -- a block, optionally tilted, is an example. So, given a 3x3 matrix, turn the three columns into vectors, and picture the space that the vectors span. The determinant of the 3x3 matrix is the area of that block.
So now you know better than most beginning math majors what a determinant is.
Treating everything as a file (directories, links, devices, services) is Unixy. Having lots of small programs that each do one thing well is also Unixy. However, for the sake of security, it's better for the operating system to treat the web browser like a kleptomaniac leper. The local machine is a safe environment, and the web is not, so it's generally bad if there's no clear (visible) distinction between safe and unsafe places for the user to be.
On the other hand, the browser feels like a very natural environment now, so it also makes sense to let users "browse" the local machine the same way. You can put anything in an emacs buffer, so sure, make it possible to put anything into a browser tab. But I think it's weird to put everything in the browser by default.
It differs between schools. At the various University of California schools, engineering degrees are 180 units (4 years @ 15 units of core classes each quarter), while music degrees float around 80 units of core classes, and communications degrees weigh in at a mere 65 units (UC Davis). Viticulture and Enology is also a 180-unit major at UC Davis -- it's a big school for wine. My music minor cost about 20 units.
I'm impressed that Iowa State requires 200 semester units for a music degree. I thought that would translate to 300 units at a quarterly school, but that sounds unreasonable, so the conversion must be something else.
Could the idea be to confuse the average consumers (and buzz-word obsessed manager types) into thinking Microsoft when they hear 'Open Source'?
Not just to confuse consumers and managers, but politicians, too. There has been a lot of legislation flying around lately regarding open standards, government use of open source, etc. (especially around the OpenDocument/OOXML dispute), and at least some politicians are beginning to like the sound of the word "Open." It sounds a bit like cleaning up corruption, something many politicians are looking for ways to demonstrate (quickly and cheaply, of course). Especially outside the U.S., governments have been pushing legislation to favor "open" platforms.
Microsoft wants to be able to cast itself as having all the advantages of open-source alternatives, but with the bonus of being backed by a huge, well-known company. By blurring the lines between open-source, especially Free software, and Microsoft's own protected, heavily patented offerings, it's easy to throw some smooth lobbyists at politicians who don't fully understand the issue and convince them that Microsoft products fit the government's need for openness and standards compliance.
And for business buyers, the reflex for the past decade or more has been: "I need a ___ product. What does Microsoft have?" So now there's an answer when the blank is "open source."
Nope, the article lists 3 directors of the foundation and Theo isn't one of them. I expect this foundation to mostly leave Theo alone to do his thing, while independently drumming up bigger donations for OpenBSD projects.
I remember a Slash article a few months ago discussing how Theo was super broke, at that discussion led to how OpenBSD's financial problems would be much more tractable if they'd just set up a foundation people or businesses could donate to. So, apparently other OpenBSD devs thought the same thing, and did it. This is good news.
Yep, I think Epiphany is exactly what the article is imagining -- strip out XULRunner, use native widgets instead to build the GUI. Plus, the framework for making Epiphany extensions is much different from that for Firefox add-ons, so it takes more effort to add bloat to your browser. Unfortunately, the add-ons are what I like about Firefox...
I've also toyed with the idea of building a browser with WebKit and Guile (with gtk+, maybe), on the theory that the extension system is the most important thing for a browser outside the rendering engine itself. Or hell, just hook emacs up to WebKit or Gecko.
We're talking about two dimensions here -- civil rights, and gun ownership. The Democratic party in the U.S. supports civil rights and opposes widespread gun ownership, usually, while the Republican party makes up the other half, currently. Not every country works this way. In particular, you're confusing Republicans with Libertarians, and Democrats with Socialists.
In the U.S., the out-of-control government is Republican, and that's the team with the most guns. As a previous poster pointed out, the ones with guns are supporting the same government that's running the constitution through a paper shredder one page at a time. Those red-blooded Texans that I actually referred to as the potential revolutionaries in my earlier post are the ones who would be suppressed by a majority of fellow Texan soldiers who strongly support the government. My point is that you can't depend on the militant portion of the population to be the resistance against an overreaching government.
For a contrasting example, see Ireland in 1920. There, the IRA was able to hold some power against the British by stockpiling arms among those who resisted the oppressive government. It's an interesting case for gun rights as a safeguard against foreign tyranny, but it doesn't bear any resemblance to the situation in the U.S.
(Naturally, fixing the problems from within the existing system is preferable to violent revolution, since even under Bush, it's really not that bad yet, compared to, say, Russia.)
If the extra oxygen (or whatever) is there to keep gamblers more alert and energetic, that's fine if it affects the dealers too, as long as it's not a health hazard.
The last time we saw something at all like what you're describing was in Beijing in 1989. The Chinese government was clearly wrong, other communist countries around the world were throwing the bums out, and so the students, intellectuals, and other members of the tiny minority of people who are ever paying attention to politics enough to try to change things started a revolt. It was so successful that much of the rest of the city joined in, and even the local military mostly turned to the people's side.
So, they brought in military from outside the city. Those guys didn't care; it wasn't their city. Most people don't care enough about the constitution to put their own lives on the line for it, even when the negative effects of its erosion are obvious. It's the prisoner's dilemma -- if the government is killing anyone who goes against the party line, I'll live if I stay quiet, and I'll die if I fight... it doesn't matter that the fight's feasible if the entire civilian population joins in, because individually we're each strongly motivated to stay out of the fight.
If some thinkers in Texas or California decided to make an assault on Washington, D.C., they'd be tagged out by the national guard before they reached the Mississippi River. The Bush Administration would have international support in suppressing this local insurgency, in stamping out a potential civil war.
The U.S. can't hold Iraq at the moment, but Saddam could -- why? The U.S. isn't trying to maintain totalitarian rule in Iraq; they'd rather be rid of it. On home turf, the rules are different, and the results would be the same as we saw in every other totalitarian country -- that is, nothing changes until it's in the ruling party's economic interest to loosen up.
PJ at Groklaw ranted about this grudge already. Presumably on account of the release of GPLv3, InformationWeek is now on a rampage of FUD and misinformation against open-source software and particularly Linux. Last week they declared Open Source / Linux "Dead". The author of that one has since written a correction, so it appears everyone at InformationWeek had a few synapses fuse this month and we're still sifting through the fallout of that.
This troll didn't need to make it to the front page.
I'm not going to try to continue with that analogy, but the theoretical goal here is to eventually discover and fix every bug in a program -- not just reduce the number of bugs being found. For a contrast, see qmail or any of the many other programs whose creators offer a bounty for any exploits found. Security bugs in popular products are worth much more than most bounties, though, and Microsoft in particular doesn't seem to offer any bounties at all.
Your analogy would be relevant if there were an effectively unlimited number of bugs in the software in question, and the goal was not to find the bugs and fix them, but to avoid ever learning about the bugs at all. So, by failing to pay up for security exploits, Microsoft is essentially stating that: (1) There is an inexhaustible supply of security holes in their software. (2) They're only concerned about addressing vulnerabilities that have already been publicized, not exploits that are quietly making the rounds on the black market.
For Internet Explorer, both of these are obviously true. IE has to withstand all the evils of the Internet, and the product doesn't even make money for Microsoft -- security is hard, and the thing had to be hacked together quickly enough to squash Netscape, ergo it's infinitely bug-ridden; and making IE safer won't make it any more profitable that $0/year, so the best tactic for the company is to spend just enough resources on it to avoid embarrassment, and no more.
For any product that has to stand on its own, that feedback loop is a good thing, putting steadily more hacker pressure on a product as it gains popularity, eventually leading to a "tried-and-true" program with a solid track record. Clearly Microsoft has enough productive developers on staff -- evidenced by the insane number of developer tools and new application frameworks they churn out every year -- but security in mainstream products like IE is not part of the business plan, and they clearly have an additional mental block at the management level that prevents them from shifting resources away from "fire and motion" and towards security, except in the very most critical products.
Anyway, at least the "wrong way" takes more effort in Python. Except for the missing colon -- but colons and whitespace vs "end" vs curly braces is an argument nobody wins.
The biggest flaw in Python is that Pythonistas get too defensive when talking about the lesser flaws. Python's sweet, but it's no silver bullet.
A Singaporean told me that although Americans eat amazingly large portions at a meal, we also eat less often. I suppose three meals a day is theoretically the standard, but in practice Americans will normally skip breakfast, then have one large meal and one comparatively small meal (lunch or dinner, varying), and graze on snacks and drinks like soda or coffee throughout the day. The portion sizes served at restaurants are not what we eat at every meal; they're a matter of marketing and presentation. If an American is still hungry after finishing his meal at a restaurant, he won't come back again -- but leftovers for the next day's lunch are always appreciated.
I'm calling out the constant grazing for a large portion of the obesity problem, in addition to the content of the food (probably the biggest issue), and it's also likely that irregular meal sizes tend to confuse the body's metabolism as well.
Dunno, I thought the panting hordes of turgid bloggers were more responsible for this superhype than Apple was. As far as I can tell, all Apple did was let Steve Jobs deliver a killer keynote address in January, feed table scraps to journalists and bloggers for six months, and then release phone on time alongside some comparatively subdued-looking TV commercials with plinky music. Correct me if I'm mistaken, but I think Apple fans and spectators just took this basic idea and ran a marathon.
For an appropriate image, picture Steve Jobs rubbing honey all over himself, then running through a field while thousands of people throw wads of loose cash at him and watch it stick. That's called marketing.
When Mark Shuttleworth won the corporate version of the lottery, he did exactly that for about 5 years. Then he started Canonical, that rare kind of company where the business model isn't a relevant question, hired a few dozen developers on his own dime, and started releasing Ubuntu in apparently a sustained effort to eventually kill Microsoft. I don't know what the moral of that story is, but I imagine it means that a future of perfectly satisfied humans means I'll someday be able to file my taxes electronically without Windows.
Shuttleworth: You're right, I did lose a million dollars last year. I expect to lose a million dollars this year. I expect to lose a million dollars next year. You know, Mr. Thatcher, at the rate of a million dollars a year, I'll have to close this place in... 200 years.
No, not that kind of wealth. Debtors (especially students and recent grads) may have a negative net worth, but they still have most of the luxuries of modern living -- often a small but reliable income, housing, heat and air conditioning, cheap food, computer access, e-mail accounts, cell phones, etc. Debt is really what makes that possible: without debt to support you, if you run out of resources, you starve.
The economy doesn't particularly have emotions or values (part of what the book seems to be about). Banks support the culture of debt, and overall, they profit from it, incidentally crushing some imprudent consumers in the process. Some societies make this economic issue into a moral one as well, and that's where death/prison/slavery come in. But here, technology has made widespread debt possible by making credit as readily available as it is today. Without credit cards (for example), you'd have to bargain with an individual each time you borrowed money, and it'd be difficult for a lender to set up a safe situation with interest, monthly billing, and almost entirely automated accounting.
As the grandparent insightfully described, modern life -- even with debt -- is full of luxuries where 500 years ago the equivalent comfort would have only been available to a small number of aristocrats. If you're reading Slashdot, you're probably not a starving subsistence farmer. If you're railing against the mechanization of modern life, especially without having taken the time to read and cite Pirsig's "Zen and the Art of Motorcycle Maintenance," then it's clear you've reached such a level of comfort in your life that you can contemplate actually giving up some of these modern luxuries in exchange for a closer look at an unnameable Quality threading through humanity.
OK, that's marketing. Marketing talking points are designed almost completely independently of actual company motivations in most companies, especially larger ones. Their pitch is that Apple products are better than Microsoft products in all ways, ergo, Safari must be better than any Windows browser, even on Windows.
Clearly, we should be observing that this browser was released on Windows just as Apple is ramping up to release the iPhone, and 3rd-party apps on the iPhone will run through Safari. As a side effect, giving web developers in general another reason to test their sites with Safari could result in a better web-browsing experience for Mac users. But it's mostly for the iPhone -- there's a reason Apple dropped "Computer" from their name earlier this year.
That's why Safari completely flouts Windows UI guidelines as well -- the idea isn't to make Safari popular on Windows, it's to let Windows developers preview their work exactly as it would render on an iPhone (or Mac, incidentally).
Interesting. This actually makes the higher-ups look much more inept than the author of this report. I'm noticing the following:
For a people which is nearly universal in its opposition to being occupied, attacking the occupier is a natural reaction and is widely accepted, even by those who are friendly to us. "It is nothing personal," one businessman told me, "I like you and believe you could be bringing us a better future, but I still sympathize with those who attack the coalition because it is not right for Iraq to be occupied by foreign military forces."
Wildly in conflict with the Bush Administration's talking points, especially in 2004. Iraq isn't the colony of bloodthirsty lunatics that right-wing news makes them out to be, and it's not a matter of a few opium-mad terrorists ruining it for everyone else. They're all fighting a foreign occupation, and everyone on the ground there, U.S. soldiers included, appears to know it.
A number of individuals have expressed satisfaction at the announcement of the new political calendar [...] What they have gotten wrong is the idea that the military will be leaving Iraq in June, which one individual said he was sure was a major factor in the diminishing attacks.
So, the attacks may have dropped because the Iraqis thought they won. Interesting. So, the key to peace is a scheduled withdrawal from the southern areas of Iraq? This may be true, but it's easy to see why it would be redacted once it passed through the Pentagon's hands.
There is also the sense among many Sunnis that we are clearly a temporary presence and other Iraqis are the real enemy so in the long term it might be better to begin to gear up and position themselves for the fight to come. This would not in the end make the reduction in attacks a good news story. In short there is plenty of fight left in the Sunnis here and plenty of weapons to carry on that fight.
I.e., it's acknowledged in the Sunni areas that the will be immediate civil war as soon as the U.S. leaves; they're in a holding pattern until then. This is an addendum to the previous point, which effectively reverses it.
The only positive remark I see in this memo is Project Money Flowing: The occupation may be causing as much violence and madness as it's preventing, but public works projects still work, getting people back to work rebuilding their country and pulling an income. So, what solution does this suggest? Dump as much money into Iraq public works and infrastructure as the U.S. did during the Great Depression; split the country between Kurd/Shi'a and Sunni (Turkey will never stand for a 3-way split) to curb the impending civil war and possible genocide; and schedule a withdrawal. Is any of this politically feasible for the Bush administration? In 2007, is anything?
But yes, it's obvious why this was mostly redacted, and I think we can all agree that the military should be using plain text editors for important information.
Slashdot Mirror
For memory limits, I would think that should be handled by the OS, or Firefox talking directly to the OS, not the user. So when malloc() finally fails, it would be nice for Firefox to squash the extension that called it (with the above notification), rather than seppuku. I guess. Because who knows, maybe there is or will be an extension that actually should be allowed to use insane amounts of memory -- like an image or video editor, maybe.
Expanding on the FF-as-OS idea, it would be nice to have extensions that work like kill, nice and top, as another poster mentioned.
Yep, there are lots of ways they could have spared themselves a few years of C++-induced pain. The root cause of the problem, IMHO, is that when they scrapped the original unusably bloated and crufty codebase, for the rewrite they restricted themselves to a very archaic subset of C++ -- for portability's sake, apparently. The restrictions went too far, though, so things like exceptions, some of the STL (auto_ptr!), even C++-style comments (// instead of /* */) were forbidden in the Moz Dev guidelines. (IIRC.) There are plenty of GC libraries out there for C++; the Boost libraries include that and many other very helpful tools -- but introducing that would be a big change to an enormous codebase. Recently they've seen the value in using newer features, particularly exceptions, and there's an effort towards simplifying the code base using these features now.
Historical note: One of the biggest ways Netscape originally doomed itself was by attempting to rewrite it all in Java. There's no silver bullet, but you don't have to shoot yourself in the foot, either.
I think any of the lighter Gecko browsers would be decent replacements for Firefox. Kazehakase or Epiphany are pretty zippy, and they might have the added benefit of being harder for the user to screw up (I'm bracing myself for the day when Firefox extensions become a serious vector for malware). Fluxbuntu, when it's done, says it's going to use K. by default, but with the user-agent string set to Mozilla to avoid trouble on sites that try to be clever. In my experience, Gnome or KDE run reasonably well even on 256 MB of RAM, but Firefox requires a multiple of that before its responsiveness becomes sane. 512 might be the bare minimum for a decent FF experience.
On the other hand, I think the whole point of gOS is to let the browser be the entire desktop. They rely on Google apps instead of OpenOffice etc., so the rest of the system is simply a minimal environment for using Firefox -- which would certainly ease the transition for Windows users who already use Firefox.
Wikipedia does users the courtesy of printing a 404 page before redirecting to a page with search results for the faulty URL. Normally, though, a site prefers to let users (and bots) know when a given URL doesn't correspond to a real page. Notice that amazon.com/slashdot returns a 404 page instead of a search!
Just because this has been done for a long time doesn't mean it's good practice. Normally, the browser on the client side does the task of letting the user run a search using a default (or selected) search engine, then diligently passing a well-formed URL on to the internet, saving the destination server some effort. But it's always been possible to do the extra work on the server side, and there are existing standards indicating how to do it.
And why don't the major web search engines already do this, if it's so user-friendly? Because they're web search engines! They exist because URLs aren't a very user-friendly way to get around online! Google doesn't want you to use the address bar, they want you to download their toolbar and rely on their web apps to find what you're looking for. They're training their users.
In Google's case, they also probably don't want so many people accidentally stumbling across their own pages while searching for things like "ig" and "search".
Windows installations are known to decay over time. Doing a complete wipe and reinstallation of Windows usually brings the computer back into usable condition, but most users have reservations about doing that. It's easier to just get a new computer -- or install the latest Ubuntu.
Why does Windows rot?
- Package management: Even when you run a program's designated uninstaller, or go through Add/Remove Programs, almost everything will leave some trace of itself behind in the registry. The new, approved way to do configuration on Windows is the same as the *nix way, with config files in the user's home folder, but nobody can force vendors to behave. Installing a new program can sometimes overwrite an existing DLL used by other programs. Plus, as part of the Windows Experience (TM), programs tend to try to modify or at least detect each other on the system, and adapt accordingly.
Sprinkle bugs into each of these areas, and you get a system that gradually develops quirks over time, slows down while looking for items that no longer exist or have been moved, and loads services that don't need to be there. Macs have a cute, tidy package system that resolves this somewhat; *nix systems usually have very powerful and thorough package managers; and both use flat text files instead of a single, binary, monolithic registry for handling program configuration.
- Commercial software: Linus is right in saying that open source is a better engineering tactic. Software for Windows is sometimes great, but usually terrible. Not only is it written hastily due to market pressure, it's dictated by managers with impure motives (e.g. marketing, add flash at the expense of quality, add a pet feature), sometimes obfuscated by coders who either don't know any better or don't want anyone else touching their code, occasionally dependent on Windows bugs or deprecated features, and frequently designed to induce vendor lock-in and force frequent upgrades.
Microsoft has always tried to provide a flexible playground for commercial developers, and until very recently would allow completely pathological things to happen on Windows if that's what was required for existing important applications to run. All of these things are limited in open-source software.
- Too much magic: The Windows Experience (TM) is marked by discoverability combined with endless kludges. There are no man pages. Instead, if you want to find out how to change some property of the system, right-click around the thing you want to change. Poke through the context menus, bring up a tabbed dialog, click around, check some boxes, click OK a few times to get out of your layers of dialogs. Job well done.
To make this work, there are multiple separate program-defined settings affecting each property of the system. Examples: sound, web browser settings, task bar items, quick launch bar, system tray. Things appear and get modified, but it's not clear what the cause was. So opening some random app for the first time in 6 months might cause the internet access settings to be updated, change the default browser from Firefox to IE, and add an item to the quick launch bar. Norton Antivirus will then detect some of this, vomit on your screen, and begin blocking Firefox from accessing the internet.
On *nix, modularity takes precedence, and configurations are managed individually. So, if you do screw up your system, you generally have just one or two places to check to see what happened (the config file, and possibly a GUI front-end for it), and it's less likely you'll change a setting by accident.
Good point. I think Intel feared Linux until just a few years ago, because the WinTel situation was such a cash cow, but then came to appreciate the idea of being partially independent of Microsoft for its business model. Plus, Linux brought in the idea of companies running critical server operations on Intel's x86 processors, rather than big iron or SPARC.
Starting with "parallelepiped" probably wasn't the best way to explain the determinant to non-mathematicians.
Fortunately, it works in two dimensions, too. Draw a parallelogram:
A_____B
/ *** /
C_____D
Get the coordinates of each of the vertices:
A = (1, 3)
B = (5, 3)
C = (0, 0)
D = (4, 0)
The area of the parallelogram is the base (4) times the height (3) = 12. Geometry.
Now pick any two sides and call them your vectors -- say, CA and CD. These two vectors span the parallelogram.
CA = [1 3]
CD = [4 0]
Then the area of the parallelogram ABCD is also the (absolute value of the) determinant of the matrix created by placing those two vectors in columns:
| 1 4 |
| 3 0 |
det = | 1*0 - 3*4 | = | -12 | = 12
Same number. Linear algebra.
To get to what the previous poster said, picture exactly the same thing in three dimensions instead of two. A parallelepiped is a three-dimensional parallelogram, more or less -- a block, optionally tilted, is an example. So, given a 3x3 matrix, turn the three columns into vectors, and picture the space that the vectors span. The determinant of the 3x3 matrix is the area of that block.
So now you know better than most beginning math majors what a determinant is.
Emacs wasn't originally written for Unix.
Treating everything as a file (directories, links, devices, services) is Unixy. Having lots of small programs that each do one thing well is also Unixy. However, for the sake of security, it's better for the operating system to treat the web browser like a kleptomaniac leper. The local machine is a safe environment, and the web is not, so it's generally bad if there's no clear (visible) distinction between safe and unsafe places for the user to be.
On the other hand, the browser feels like a very natural environment now, so it also makes sense to let users "browse" the local machine the same way. You can put anything in an emacs buffer, so sure, make it possible to put anything into a browser tab. But I think it's weird to put everything in the browser by default.
(Yes, I use Xfce and Vim.)
It differs between schools. At the various University of California schools, engineering degrees are 180 units (4 years @ 15 units of core classes each quarter), while music degrees float around 80 units of core classes, and communications degrees weigh in at a mere 65 units (UC Davis). Viticulture and Enology is also a 180-unit major at UC Davis -- it's a big school for wine. My music minor cost about 20 units.
I'm impressed that Iowa State requires 200 semester units for a music degree. I thought that would translate to 300 units at a quarterly school, but that sounds unreasonable, so the conversion must be something else.
Not just to confuse consumers and managers, but politicians, too. There has been a lot of legislation flying around lately regarding open standards, government use of open source, etc. (especially around the OpenDocument/OOXML dispute), and at least some politicians are beginning to like the sound of the word "Open." It sounds a bit like cleaning up corruption, something many politicians are looking for ways to demonstrate (quickly and cheaply, of course). Especially outside the U.S., governments have been pushing legislation to favor "open" platforms.
Microsoft wants to be able to cast itself as having all the advantages of open-source alternatives, but with the bonus of being backed by a huge, well-known company. By blurring the lines between open-source, especially Free software, and Microsoft's own protected, heavily patented offerings, it's easy to throw some smooth lobbyists at politicians who don't fully understand the issue and convince them that Microsoft products fit the government's need for openness and standards compliance.
And for business buyers, the reflex for the past decade or more has been: "I need a ___ product. What does Microsoft have?" So now there's an answer when the blank is "open source."
Nope, the article lists 3 directors of the foundation and Theo isn't one of them. I expect this foundation to mostly leave Theo alone to do his thing, while independently drumming up bigger donations for OpenBSD projects.
I remember a Slash article a few months ago discussing how Theo was super broke, at that discussion led to how OpenBSD's financial problems would be much more tractable if they'd just set up a foundation people or businesses could donate to. So, apparently other OpenBSD devs thought the same thing, and did it. This is good news.
Yep, I think Epiphany is exactly what the article is imagining -- strip out XULRunner, use native widgets instead to build the GUI. Plus, the framework for making Epiphany extensions is much different from that for Firefox add-ons, so it takes more effort to add bloat to your browser. Unfortunately, the add-ons are what I like about Firefox...
I've also toyed with the idea of building a browser with WebKit and Guile (with gtk+, maybe), on the theory that the extension system is the most important thing for a browser outside the rendering engine itself. Or hell, just hook emacs up to WebKit or Gecko.
We're talking about two dimensions here -- civil rights, and gun ownership. The Democratic party in the U.S. supports civil rights and opposes widespread gun ownership, usually, while the Republican party makes up the other half, currently. Not every country works this way. In particular, you're confusing Republicans with Libertarians, and Democrats with Socialists.
In the U.S., the out-of-control government is Republican, and that's the team with the most guns. As a previous poster pointed out, the ones with guns are supporting the same government that's running the constitution through a paper shredder one page at a time. Those red-blooded Texans that I actually referred to as the potential revolutionaries in my earlier post are the ones who would be suppressed by a majority of fellow Texan soldiers who strongly support the government. My point is that you can't depend on the militant portion of the population to be the resistance against an overreaching government.
For a contrasting example, see Ireland in 1920. There, the IRA was able to hold some power against the British by stockpiling arms among those who resisted the oppressive government. It's an interesting case for gun rights as a safeguard against foreign tyranny, but it doesn't bear any resemblance to the situation in the U.S.
(Naturally, fixing the problems from within the existing system is preferable to violent revolution, since even under Bush, it's really not that bad yet, compared to, say, Russia.)
If the extra oxygen (or whatever) is there to keep gamblers more alert and energetic, that's fine if it affects the dealers too, as long as it's not a health hazard.
The last time we saw something at all like what you're describing was in Beijing in 1989. The Chinese government was clearly wrong, other communist countries around the world were throwing the bums out, and so the students, intellectuals, and other members of the tiny minority of people who are ever paying attention to politics enough to try to change things started a revolt. It was so successful that much of the rest of the city joined in, and even the local military mostly turned to the people's side.
So, they brought in military from outside the city. Those guys didn't care; it wasn't their city. Most people don't care enough about the constitution to put their own lives on the line for it, even when the negative effects of its erosion are obvious. It's the prisoner's dilemma -- if the government is killing anyone who goes against the party line, I'll live if I stay quiet, and I'll die if I fight... it doesn't matter that the fight's feasible if the entire civilian population joins in, because individually we're each strongly motivated to stay out of the fight.
If some thinkers in Texas or California decided to make an assault on Washington, D.C., they'd be tagged out by the national guard before they reached the Mississippi River. The Bush Administration would have international support in suppressing this local insurgency, in stamping out a potential civil war.
The U.S. can't hold Iraq at the moment, but Saddam could -- why? The U.S. isn't trying to maintain totalitarian rule in Iraq; they'd rather be rid of it. On home turf, the rules are different, and the results would be the same as we saw in every other totalitarian country -- that is, nothing changes until it's in the ruling party's economic interest to loosen up.
PJ at Groklaw ranted about this grudge already. Presumably on account of the release of GPLv3, InformationWeek is now on a rampage of FUD and misinformation against open-source software and particularly Linux. Last week they declared Open Source / Linux "Dead". The author of that one has since written a correction, so it appears everyone at InformationWeek had a few synapses fuse this month and we're still sifting through the fallout of that.
This troll didn't need to make it to the front page.
I'm not going to try to continue with that analogy, but the theoretical goal here is to eventually discover and fix every bug in a program -- not just reduce the number of bugs being found. For a contrast, see qmail or any of the many other programs whose creators offer a bounty for any exploits found. Security bugs in popular products are worth much more than most bounties, though, and Microsoft in particular doesn't seem to offer any bounties at all.
Your analogy would be relevant if there were an effectively unlimited number of bugs in the software in question, and the goal was not to find the bugs and fix them, but to avoid ever learning about the bugs at all. So, by failing to pay up for security exploits, Microsoft is essentially stating that:
(1) There is an inexhaustible supply of security holes in their software.
(2) They're only concerned about addressing vulnerabilities that have already been publicized, not exploits that are quietly making the rounds on the black market.
For Internet Explorer, both of these are obviously true. IE has to withstand all the evils of the Internet, and the product doesn't even make money for Microsoft -- security is hard, and the thing had to be hacked together quickly enough to squash Netscape, ergo it's infinitely bug-ridden; and making IE safer won't make it any more profitable that $0/year, so the best tactic for the company is to spend just enough resources on it to avoid embarrassment, and no more.
For any product that has to stand on its own, that feedback loop is a good thing, putting steadily more hacker pressure on a product as it gains popularity, eventually leading to a "tried-and-true" program with a solid track record. Clearly Microsoft has enough productive developers on staff -- evidenced by the insane number of developer tools and new application frameworks they churn out every year -- but security in mainstream products like IE is not part of the business plan, and they clearly have an additional mental block at the management level that prevents them from shifting resources away from "fire and motion" and towards security, except in the very most critical products.
I think that was the point.
Anyway, at least the "wrong way" takes more effort in Python. Except for the missing colon -- but colons and whitespace vs "end" vs curly braces is an argument nobody wins.
The biggest flaw in Python is that Pythonistas get too defensive when talking about the lesser flaws. Python's sweet, but it's no silver bullet.
A Singaporean told me that although Americans eat amazingly large portions at a meal, we also eat less often. I suppose three meals a day is theoretically the standard, but in practice Americans will normally skip breakfast, then have one large meal and one comparatively small meal (lunch or dinner, varying), and graze on snacks and drinks like soda or coffee throughout the day. The portion sizes served at restaurants are not what we eat at every meal; they're a matter of marketing and presentation. If an American is still hungry after finishing his meal at a restaurant, he won't come back again -- but leftovers for the next day's lunch are always appreciated.
I'm calling out the constant grazing for a large portion of the obesity problem, in addition to the content of the food (probably the biggest issue), and it's also likely that irregular meal sizes tend to confuse the body's metabolism as well.
Minor clarification: The fully assembled skyscraper would also be harmful if you inhaled it.
Dunno, I thought the panting hordes of turgid bloggers were more responsible for this superhype than Apple was. As far as I can tell, all Apple did was let Steve Jobs deliver a killer keynote address in January, feed table scraps to journalists and bloggers for six months, and then release phone on time alongside some comparatively subdued-looking TV commercials with plinky music. Correct me if I'm mistaken, but I think Apple fans and spectators just took this basic idea and ran a marathon.
For an appropriate image, picture Steve Jobs rubbing honey all over himself, then running through a field while thousands of people throw wads of loose cash at him and watch it stick. That's called marketing.
When Mark Shuttleworth won the corporate version of the lottery, he did exactly that for about 5 years. Then he started Canonical, that rare kind of company where the business model isn't a relevant question, hired a few dozen developers on his own dime, and started releasing Ubuntu in apparently a sustained effort to eventually kill Microsoft. I don't know what the moral of that story is, but I imagine it means that a future of perfectly satisfied humans means I'll someday be able to file my taxes electronically without Windows.
No, not that kind of wealth. Debtors (especially students and recent grads) may have a negative net worth, but they still have most of the luxuries of modern living -- often a small but reliable income, housing, heat and air conditioning, cheap food, computer access, e-mail accounts, cell phones, etc. Debt is really what makes that possible: without debt to support you, if you run out of resources, you starve.
The economy doesn't particularly have emotions or values (part of what the book seems to be about). Banks support the culture of debt, and overall, they profit from it, incidentally crushing some imprudent consumers in the process. Some societies make this economic issue into a moral one as well, and that's where death/prison/slavery come in. But here, technology has made widespread debt possible by making credit as readily available as it is today. Without credit cards (for example), you'd have to bargain with an individual each time you borrowed money, and it'd be difficult for a lender to set up a safe situation with interest, monthly billing, and almost entirely automated accounting.
As the grandparent insightfully described, modern life -- even with debt -- is full of luxuries where 500 years ago the equivalent comfort would have only been available to a small number of aristocrats. If you're reading Slashdot, you're probably not a starving subsistence farmer. If you're railing against the mechanization of modern life, especially without having taken the time to read and cite Pirsig's "Zen and the Art of Motorcycle Maintenance," then it's clear you've reached such a level of comfort in your life that you can contemplate actually giving up some of these modern luxuries in exchange for a closer look at an unnameable Quality threading through humanity.
OK, that's marketing. Marketing talking points are designed almost completely independently of actual company motivations in most companies, especially larger ones. Their pitch is that Apple products are better than Microsoft products in all ways, ergo, Safari must be better than any Windows browser, even on Windows.
Clearly, we should be observing that this browser was released on Windows just as Apple is ramping up to release the iPhone, and 3rd-party apps on the iPhone will run through Safari. As a side effect, giving web developers in general another reason to test their sites with Safari could result in a better web-browsing experience for Mac users. But it's mostly for the iPhone -- there's a reason Apple dropped "Computer" from their name earlier this year.
That's why Safari completely flouts Windows UI guidelines as well -- the idea isn't to make Safari popular on Windows, it's to let Windows developers preview their work exactly as it would render on an iPhone (or Mac, incidentally).
Interesting. This actually makes the higher-ups look much more inept than the author of this report. I'm noticing the following:
For a people which is nearly universal in its opposition to being occupied, attacking the occupier is a natural reaction and is widely accepted, even by those who are friendly to us. "It is nothing personal," one businessman told me, "I like you and believe you could be bringing us a better future, but I still sympathize with those who attack the coalition because it is not right for Iraq to be occupied by foreign military forces."Wildly in conflict with the Bush Administration's talking points, especially in 2004. Iraq isn't the colony of bloodthirsty lunatics that right-wing news makes them out to be, and it's not a matter of a few opium-mad terrorists ruining it for everyone else. They're all fighting a foreign occupation, and everyone on the ground there, U.S. soldiers included, appears to know it.
A number of individuals have expressed satisfaction at the announcement of the new political calendar [...] What they have gotten wrong is the idea that the military will be leaving Iraq in June, which one individual said he was sure was a major factor in the diminishing attacks.So, the attacks may have dropped because the Iraqis thought they won. Interesting. So, the key to peace is a scheduled withdrawal from the southern areas of Iraq? This may be true, but it's easy to see why it would be redacted once it passed through the Pentagon's hands.
There is also the sense among many Sunnis that we are clearly a temporary presence and other Iraqis are the real enemy so in the long term it might be better to begin to gear up and position themselves for the fight to come. This would not in the end make the reduction in attacks a good news story. In short there is plenty of fight left in the Sunnis here and plenty of weapons to carry on that fight.I.e., it's acknowledged in the Sunni areas that the will be immediate civil war as soon as the U.S. leaves; they're in a holding pattern until then. This is an addendum to the previous point, which effectively reverses it.
The only positive remark I see in this memo is Project Money Flowing: The occupation may be causing as much violence and madness as it's preventing, but public works projects still work, getting people back to work rebuilding their country and pulling an income. So, what solution does this suggest? Dump as much money into Iraq public works and infrastructure as the U.S. did during the Great Depression; split the country between Kurd/Shi'a and Sunni (Turkey will never stand for a 3-way split) to curb the impending civil war and possible genocide; and schedule a withdrawal. Is any of this politically feasible for the Bush administration? In 2007, is anything?
But yes, it's obvious why this was mostly redacted, and I think we can all agree that the military should be using plain text editors for important information.