Windows ME was a stop-gap measure, we all knew that. Windows 2000 was technically a better operating system than WindowsME, but it didn't have the bells and whistles needed to sell to consumers (business yes, joe sixpack in Best Buy, not so much), and XP was still too far out. Microsoft needed to give Win98 a face-lift until XP was ready.
Yep. If Microsoft had a nickel for every incorrect assessment of its impending death "real soon now," it would be richer than it already is. Back in the real world, Microsoft will probably outlive all its detractors with a decade to spare. That would be true if there were > 0 incorrect assessments (1 = $0.05 richer), so while you are factually correct, it lacks the impact you where trying for.
Can't blame PHP entirely here, the PHP process runs as the webserver, which should _not_ have root access. If a PHP forum was used, then there is a vulnerability in Apache and/or Linux.
I still think it's a lousy password and/or lousy password security to blame.
The article says there is no current evidence that the Linux machines were compromised because of a software vulnerability, they're speculating password cracking or social engineering got them root access. The Windows machines, however, get infected because of a software vulnerability on them.
And also the fact that quantity of vulnerabilities is a crappy metric in and of itself. I'd rather have 100 website spoofing vulnerabilities in Firefox than a single remote code execution vulnerability in IE.
The screenshot in the links shows KOrganizer running under the standard windows shell. I'm not sure if kwin/kpanel/kdesktop will be ported as shell replacements for Explorer. Until they do, you will not be getting proper virtual desktops.
Not having RTFM I'm wondering if it fixes some of the backwards shit in Windows (like the subdirectory separator) No, it's just a port of the QT and KDE4 libs, and some KDE programs that use those libs, to Windows. While KDE apps will probably be able to use the correct / when specifying a path, don't expect this to fix any native Windows apps.
No, but they at least have a database product now. And we all know that in a year's time they will be selling it under a different name with some enterprise management tools integrated into Solaris and ZFS and provide benchmarks running on Niagara showing it besting DB2 in some operation. And even if it still won't be anywhere near as good as DB2, it will be enough for Sun's salespeople to convince those who make purchase decisions that it is, and selling Niagara+Solaris+SJSAS+MySQL(renamed)+Storage package solutions will add up to $500million in no time.
Distributing an app that uses ODBC or JDBC doesn't distribute any MySQL. Most that I have seen that do this require that you provide the ODBC or JDBC driver yourself.
Actually, Sun would only have to grow Sun's revenue to $500million per year as a result of this purchase. If offering MySQL results in an increase of Solaris systems running on Niagra processors with Thumper data storage, instead of comparable stacks from IBM, HP or Dell, they could easily make that kind of increase.
I think it's more about IBM than Oracle. IBM is one of Sun's biggest competitors, they compete on hardware and software, from big iron to blades and desktops. They even compete against Sun's IDE and UI toolkit for Java.
Most important though, IBM has DB2, so you can get everything you need from IBM. Until now, Sun didn't have a competitive database offering, they would push Postgres, but they couldn't own Postgres. Buying MySQL not only gives them a well known database with a large user base, but it gives them copyright ownership of all the code. For a long time now Sun's stack has been incomplete, MySQL fills that gap.
Tomcat is just a Servlet/JSP container, where Glassfish is a J2EE server like Weblogic or Websphere. The apache equivalent is Geronimo, which I believe uses Tomcat for the Servlet/JSP portion of the J2EE spec.
I've not looked at the spec, but how is this implemented? There are lots of constraints that you want to be able to apply to an input box. Having one that accepts anything and one that accepts numbers sounds like an ugly hack. I'm guessing that the "Number" input would most likely be implemented by a spinner type widget, though the actual implementation may be up to the browser.
Adding a regex attribute to text input boxes, forcing the browser to only accept things that match the regex would be a lot neater. It would allow you to do things like post codes or telephone numbers with no JavaScript, for example, or tracking numbers with a particular grouping of numbers. As I mentioned, all input fields will have an optional "pattern" attribute, which I would assume lets you specify a regex pattern that the input must match in order to validate.
Are you a part of the group working on the HTML5 spec? If so, I would like to make a suggestion. There are several new elements that will replace css/javascript hacks and libraries, like Menu and the new input types, but I was disappointed to not have any kind of "popup" element. I'm talking about more than just a floating div, an element with a title bar and close button (these could be optional I guess), and is movable and resizable.
There are lots of hacks out there that do this, but it seems like it could be something better done by the rendering engine directly, instead of having javascript libraries keeping z-index stack orders and capturing onmousemove events, which often cause problems like selecting portions of a page while "dragging" a popup.
Was there any discussion of adding such an element into the new spec?
Yes, let's roll out another new standard for no reason at all when most of the web still hasn't caught up to the last one. I read the spec, and most of the additions are to incorporate what we've all been doing, with CSS/Javascript hacks and plugins, into the spec itself.
Menu tag: An actual menu element provided by the render engine, no more CSS/Javascript menu libraries
Combo-box input: An actual combo-box form input field, no more CSS/Javascript hacks
Datetime input: No more javascript libraries to popup a date chooser, it comes with the browser
Number input: No more javascript intervention to stop you from typing "bob" into a field expecting a number.
Email input: Not only can it validate email format, but it could let you pull from your email client's addressbook
URL input: Again, not only validation, but could let you pull from your browser's favorites, history, etc
Input validation: No more onSubmit javascript checks to make sure your required fields have data in them.
All Inputs have min/max and pattern (presumably regex) restraints
(BTW, all of the above should work even when Javascript is disabled, so you NoScript users get security _AND_ functionality)
Datagrid: An actual data grid widget, no more hacking Tables to provide the same functionality
Audio/Video tags: Just like <img/> tags, only for other media, what a concept!
While I won't say it is faster, I can't say I notice it being much slower Running on the same hardware? Or are you saying that a Core 2 Duo with Vista isn't any slower than a P3 with XP? If it's the latter, It's not a compliment.
There are a lot of good articles, like how to put those usb thumb drives to good use to improve load time of frequently used programs. Are those articles titled "How to reduce the life of your flash memory because Vista needs more RAM than your new computer came with"?
Memory management is much better, and if you load it up with RAM(at least 2 gigs), you'll notice the difference. If it takes 2Gb of RAM to notice the difference, I'd argue that the memory management isn't much better.
For example, you can profile a C/C++ program running with real data and then recompile using the generated profile to guide optimisation. But again you would be limited to optimizing for one situation and not another. Going to my code example in an above post, if your profile tells you to optimize for many calls to myfunc() with a small value for myLimit, it won't do you much good if the next day you're running a few calls to myfunc() with a large value for myLimit.
generally, the way I see it done is pretty much a "human debugger" -- that is, someone at least running through each line, explaining what it does, at least once, to get you to think that way. I think that's kind of the point, to make you internalize the logic process. Once you learn to think the same way your computer does you become a better programmer, because the things that won't make sense to your computer won't make sense to you. You start thinking about efficient code not by how many lines it takes to write it, but my how much memory is will consume, how many objects it is creating, how many context switches it will have to perform, etc.
Just to illustrate in code:
static int x; public void myfunc(int myLimit) { for (i = 0; i < myLimit; ++i) ++x; } What is the best way to optimize that code? Well that depends on the size of myLimit, which the C/C++ won't know. But in Java, once the process is running, and the JVM sees that myfunc is being called several times, always with a small value for myLimit, it can optimize the execution for that. If the same binary is running in another process where myLimit is large, it can optimize for that.
Java can be optimized at runtime, so the JVM can watch the running code, see what gets executed most often, and optimize the _running_ code for that situation. C/C++ can only optimize at compile time according to what it _thinks_ will be executed most often. Sometimes one it better than the other, but profiled runtime optimization on a long-running server process is one of the situations where Java can perform better than statically compiled code.
Slashdot Mirror
Windows ME was a stop-gap measure, we all knew that. Windows 2000 was technically a better operating system than WindowsME, but it didn't have the bells and whistles needed to sell to consumers (business yes, joe sixpack in Best Buy, not so much), and XP was still too far out. Microsoft needed to give Win98 a face-lift until XP was ready.
Can't blame PHP entirely here, the PHP process runs as the webserver, which should _not_ have root access. If a PHP forum was used, then there is a vulnerability in Apache and/or Linux.
I still think it's a lousy password and/or lousy password security to blame.
The article says there is no current evidence that the Linux machines were compromised because of a software vulnerability, they're speculating password cracking or social engineering got them root access. The Windows machines, however, get infected because of a software vulnerability on them.
http://finance.yahoo.com/q/bc?t=5y&s=MSFT&l=on&z=m&q=l&c=NOVL
It turns out, that may not be bad financial advice.
And also the fact that quantity of vulnerabilities is a crappy metric in and of itself. I'd rather have 100 website spoofing vulnerabilities in Firefox than a single remote code execution vulnerability in IE.
Short answer: No.
Long Answer: No, but I think kwin has (or had) compositing capability, so it could potentially provide Vista-like features on Windows.
The screenshot in the links shows KOrganizer running under the standard windows shell. I'm not sure if kwin/kpanel/kdesktop will be ported as shell replacements for Explorer. Until they do, you will not be getting proper virtual desktops.
No, but they at least have a database product now. And we all know that in a year's time they will be selling it under a different name with some enterprise management tools integrated into Solaris and ZFS and provide benchmarks running on Niagara showing it besting DB2 in some operation. And even if it still won't be anywhere near as good as DB2, it will be enough for Sun's salespeople to convince those who make purchase decisions that it is, and selling Niagara+Solaris+SJSAS+MySQL(renamed)+Storage package solutions will add up to $500million in no time.
Yes, but that goes against the definition of distribution in the GPL, which in the end is the only definition that matters here.
Distributing an app that uses ODBC or JDBC doesn't distribute any MySQL. Most that I have seen that do this require that you provide the ODBC or JDBC driver yourself.
Actually, Sun would only have to grow Sun's revenue to $500million per year as a result of this purchase. If offering MySQL results in an increase of Solaris systems running on Niagra processors with Thumper data storage, instead of comparable stacks from IBM, HP or Dell, they could easily make that kind of increase.
I think it's more about IBM than Oracle. IBM is one of Sun's biggest competitors, they compete on hardware and software, from big iron to blades and desktops. They even compete against Sun's IDE and UI toolkit for Java.
Most important though, IBM has DB2, so you can get everything you need from IBM. Until now, Sun didn't have a competitive database offering, they would push Postgres, but they couldn't own Postgres. Buying MySQL not only gives them a well known database with a large user base, but it gives them copyright ownership of all the code. For a long time now Sun's stack has been incomplete, MySQL fills that gap.
Tomcat is just a Servlet/JSP container, where Glassfish is a J2EE server like Weblogic or Websphere. The apache equivalent is Geronimo, which I believe uses Tomcat for the Servlet/JSP portion of the J2EE spec.
Are you a part of the group working on the HTML5 spec? If so, I would like to make a suggestion. There are several new elements that will replace css/javascript hacks and libraries, like Menu and the new input types, but I was disappointed to not have any kind of "popup" element. I'm talking about more than just a floating div, an element with a title bar and close button (these could be optional I guess), and is movable and resizable.
There are lots of hacks out there that do this, but it seems like it could be something better done by the rendering engine directly, instead of having javascript libraries keeping z-index stack orders and capturing onmousemove events, which often cause problems like selecting portions of a page while "dragging" a popup.
Was there any discussion of adding such an element into the new spec?
(BTW, all of the above should work even when Javascript is disabled, so you NoScript users get security _AND_ functionality)
That's how you know you did it right, when half the programmers in the world say it's too easy, and the other half say it's too hard.
Either that, or it means you did it horribly horribly wrong.
public void myfunc(int myLimit) { for (i = 0; i < myLimit; ++i) ++x; } What is the best way to optimize that code? Well that depends on the size of myLimit, which the C/C++ won't know. But in Java, once the process is running, and the JVM sees that myfunc is being called several times, always with a small value for myLimit, it can optimize the execution for that. If the same binary is running in another process where myLimit is large, it can optimize for that.
Java can be optimized at runtime, so the JVM can watch the running code, see what gets executed most often, and optimize the _running_ code for that situation. C/C++ can only optimize at compile time according to what it _thinks_ will be executed most often. Sometimes one it better than the other, but profiled runtime optimization on a long-running server process is one of the situations where Java can perform better than statically compiled code.