I guess Handyman's speakers belong in that category and would do fine in the party cellar.
In fact, not really. I have a pair of Linn Keilidhs, which have a pretty clean, high-end sound to them -- not the kind of thing you want for your party cellar. I must admit that the efficiency of these speakers is pretty high for speakers in the (sort-of) high-end segment.
Also, when comparing speaker performance, Watts are definitely not the complete picture.
The first thing you need to find out is the efficiency of speakers. For instance, my speakers have an efficiency of 92 dB/W, which means that at a power level of 1 W, they will produce 92 dB of sound. As dB is a logarithmic scale, doubling the wattage will increase the number of dBs by 3, so a power level of 64 W will get me a 92 + 3 * 6 = 110 dB sound level. However, a speaker with an efficiency of 80 dB/W will only produce 98 dB for the same amount of power. I've seen efficiencies ranging from 70 dB/W up to the high ninety-somethings, so be careful to check these numbers.
The second thing you need to find out is the impedance of the speakers, combined with the impedance your amplifier is rated for. For instance, my amplifier is not simply rated as 50 W, but as 50 W for speakers with an impedance of 8 Ohms, and 100 W for speakers with an impedance of 4 Ohms. This can make some difference. Watch out with getting a speaker with very low impedance though: if your amplifier wasn't designed to handle that, they will probably draw too much power, causing the amplifier to get overheated. In addition, you will not be able to open up your volume knob more than a couple of millimeters -- and volume is probably something you like to have detailed control over.
And [FrontPage] always produces 100% valid XHTML, does it?
No, it doesn't. But my argument was that simplicity takes on a completely different dimention if the user can type:
Hello my name is
Fred
instead of
Hello my name is <b>Fred</b>
and have FrontPage generate a 100 kb HTML file. Then the responsibility for the validity of the XHTML then resides with FrontPage. IMO correcting crap HTML is only excusable if the HTML is written by normal users, because then it's a usability feature. If it's written by a computer program any browser should always barf if it's crap. The fact that FrontPage generates a shitload of crap HTML for Fred's simple sentence is purely due to the fact that FrontPage is optimized for IE, and that they never checked if other, more strict browsers would still render the garbage they generate.
90% of web builders started this way. If they'd had to create fully valid HTML before the browser gave them the time of day, they'd probably have fallen at the first hurdle. But they didn't, they went on to create web pages, which is why the web is as huge and popular and useful as it is today.
Maybe that used to be the case, but right now, people simply open FrontPage and write web pages in a WYSIWYMGOSB (What You See Is What You Might Get On Selected Browsers) kind of way. If you were right about this, then (quoting you):
the web would be about 0.1% of the size it is now.
Oh my - OO starts in 9-11 seconds on my rather anemic system. Heaven forbid they should have taken maybe 2 more seconds than abs necessary to get things going. In case you missed it the preceding sentence was laced with enough sarcasm that it threatens imminent collapse into a biting singularity of incredible destructive power. Slower? Yes I can see that but it's certainly not slow. Oh and BTW - OO was fast enough to boot simply to spell check this post.
On my 1 GHz P3 it takes about 14 seconds. For comparison, abiword takes about 3 seconds to start, and when I boot into Win XP, MS Word 2003 doesn't take much more than that. Why should it take fifteen seconds to start a text editor with an empty document? There may of course be some technical reasons why it's hard, but those reasons are always a result of technical priorities when they first built the product. They probably built the application in a monolithic, hard to split up way that makes it difficult to add on-demand or delayed loading of features. That's a design decision that's hard to reverse. Look at Linux: it took them years to make that kernel as modular as it is now. But really, if fast startup and low memory footprint would have been a priority for the OOo design in the first place, then they would have managed to have startup times in the 5 second range by now. It will now take them ages longer to get to that point because it simply wasn't designed in...
OpenOffice splash screen is not always-on-top anymore. Check the latest version.
Not true, unfortunately. I run 1.1.3, which is the latest version, and the splash screen is always-on-top. Yes, there's 1.1.4 RC, but as far as I'm concerned that's not the "latest version", that's the "development version".
Yes, a long delay without a splash screen is irritating as well. But there is nothing as irritating as a splash screen that is always-on-top. What were they THINKING making the OOo splash screen always-on-top? They already have a reputation for being the slowest starting office product out there, must they rub it in your face then?
The Windows Template Library is released as CPL as well. I don't know since when this has been, but the product has been in existence for quite some time now.
Kaspersky labs says they were misquoted. Quoting from a mail from kaspersky labs themselves (as found in a repost on the NTBugtraq mailing list):
A handful of sites are stating that Eugene Kaspersky, founder of Kaspersky Labs, believes that tomorrow will bring a massive terrorist attack on the Internet. This is being quoted in a range of ways, ranging from factual reporting to citing the story as an example of cyber hysteria.
However, Kaspersky is not predicting the end of the Internet tomorrow - or even in the near future. The story stems from brief comments made yesterday at a press conference which was dedicated to cybercrime and the problems of spam.
At this press conference, Kaspersky commented that the possibility of terrorists using the Internet as a tool to attack certain countries as a reality. As an example, he cited the fact that a number of Arabic and Hebrew language websites contained an announcement of an 'electronic jihad' against Israel, to start on 26th August 2004.
In an interview today, Kaspersky stressed that such information was not necessarily trustworthy. 'We don't know who is behind these statements.' He went on to clarify: 'It's not the first time the term 'electronic jihad' has been used. We've seen this before, with the focus being on sending racist emails, and defacing and hacking Israeli web sites. But it is the first time I have seen sites encouraging the use of Internet attacks against one country as a form of terrorism.'
'As we've already stated many times in the past, it would be easy enough to use a network of infected computers to launch such an attack. We saw the impact that Sasser, Mydoom and Slammer had, on the Internet, businesses and organisations. Just imagine if such an attack was directed at one country or one critical point in the infrastructure of the Internet. Computers are a tool - and just like any tool, they can be used or misused.'
Kaspersky emphasised that the likelihood of a massive attack directed against Israeli institutions tomorrow is low. However, he believes that Pandora's box has now been opened. Hackers and virus writers can be motivated by a range of factors: money, curiosity, or political conviction. But whatever their motivation, the insecure nature of the Internet and weak security precautions offer a wealth of opportunities. 'Maybe it won't be tomorrow, or the day after tomorrow - but sooner or later, terrorists will be using the Internet as another weapon in their arsenal.'
Most people don't jump up to enter their username and password as soon as the prompt appears. They turn the computer on, walk to somewhere else and such, and when they get back, all services are finished loading and it's ready and waiting for them.
In addition, processing X windows password GUI is not an extremely intensive task, and selecting the username and typing the password will take at least enough time for some of the background work to be finished. The only thing that is really overlapped is the finishing of the initialization of services and the user's interaction with the display manager -- a very processing-intensive task with a very user-intensive task. Smart. Much smarter than overlapping multiple processing-intensive tasks, because the resources involved are different -- user time versus machine time.
Oh, please. All Microsoft did was make the desktop appear sooner, when there's still a ton of shit to load.
That is not true. This page indicates that "Windows XP initializes device drivers in parallel to improve boot time." And this page indicates that not only does Windows XP do prefetching of data required at boot time, but it also optimizes the data so that it is placed sequentially on the outer edge of the disk.
Note that you assume the situation where the system boots directly into a desktop, i.e. a configuration with automatic login. This may cause the background loading to be unfinished when the desktop is already loaded. But only on systems that run too many services. Those machines are apparently servers, and you'll NEVER want to configure those with automatic logins.
I've got a home-built Athlon XP 2600 system that is running XP Pro, has an assload of RAM, and is kept well-tuned and malware-free, and that fucker takes a good 30-45 seconds after the desktop appears before it actually pays attention to any of my attempts to do anything.
Then you must be one lousy system tuner. No wonder you posted this as AC.:) Seriously, my girlfriend's Athlon 750 machine runs XP like a charm, and after booting and automatically logging in, the desktop is immediately very responsive. I don't know what you've done to your system to make it so slow. Are you running one of those virus scanners (note to Linux users: those are mandatory on windows machines to keep them running for more than a day) that scans everything that loads and that slows your machine down by >50%?
For the "feel" of a machine, latency or "response time" is the most important factor. When the user requests an action, it is the time between the request and the machine's response that counts. For instance, the almost 2x speedup of booting XP means a 2x decrease in a very annoying latency, and it makes the system feel much faster even if nothing else changes. The numerous *small* latencies in a system also count -- don't you hate it when you click a menu and you have to wait a full two seconds before it pops up? The improvements measured in the benchmarks done by tweakers.net don't do justice to the importance of latency. The user doesn't care whether some background process (e.g. eMule) is fast -- he cares whether if he clicks a button, the result will show up with or without a noticeable delay. So what they should really be measuring is the time between certain checkpoints in a trace, e.g., the time between the point where the user did something (action) and the time when all the necessary data to respond to that action has been read (response).
Note that the 2x speedup can be easily explained. Windows XP optimizes the boot process by automatically generating traces of disk accesses done at boot, and by reordering the accessed blocks on disk so that they can be read in sequentially in the next boot. And striping over two disks theoretically improves sequential read throughput by... yes, a factor 2.
About compiling in Visual Studio: that can be fixed. A colleague of mine wrote a small tool that monitored the system for processes named CL.EXE and LINK.EXE, and that immediately set them to idle priority whenever they appeared. Problem solved.
The real problem is of course in the Windows scheduler. The Windows scheduler simply gives the _foreground task_ a fixed boost for being interactive, and if you're working in Visual Studio, then Visual Studio is the foreground process. The compiler/linker processes are children of the foreground process, so it wouldn't surprise me if those processes inherited that "foreground boost", making it impossible to do anything else on the machine...
The Linux 2.6 scheduler is WAY better at determining which tasks require an "interactivity boost". Also, the interactivity boost given by Linux ultimately influences only how soon a process will be scheduled (latency), not how much processing time it will get (throughput). This is caused by the metric for interactivity, which boils down to "how often does a process give up the processor". This works because interactive processes tend to do short bursts of processing and then stop again (e.g., repaint the screen, process a keystroke, etcetera), whereas noninteractive processes tend to continue processing until they're completely done.
In the good old days (this would be the early 90s), I bought an MFM controller and an accompanying 20 MB hard drive. Little did I know that an MFM controller and an IDE controller wouldn't work together in the same box. I was disappointed that it didn't work, and, not knowing much about hardware then, I figured that the controller card might need some power as well. So I plugged a small power connector (that is normally used to supply power to the 3.5" disk drive) onto the four pins that came out of the MFM controller. When I turned on the computer, the hard drive spun up a lot slower than normal. Then black smoke came from the connection between the power and the MFM controller. I quickly turned the computer off again, and after removing the MFM controller the computer still worked. Phew! When I checked what the connector actually said, it turned out that I had connected the power to the MFM controller's hard disk led connection.:)
I'm a programmer and I'm listening to Pink Floyd while I write this, yet I'm closer to 20 years old:).
OK, I can see why you wrote this anonymously. Don't want to upset the electro crowd at work.:)
(On a sidenote: I'm a programmer as well, doing mostly prog rock. I was considering to write this anonymously because *grrrr* if someone mods you down as overrated after you've been modded up as funny, you *lose* karma in this wicked system. If anyone is listening: the karma increase should be something like regular_positive_mods - max(0, overrated_mods - funny_mods). The current system not only discourages the semi-funny running-gag posts (e.g. 1. You're in Japan, 2. You move to soviet Russia, 3. Profit!!!) but also the seriously originally funny posts, because if you get modded up to +5 Funny, then +4, Overrated, you lose karma, even if you _were_ really worth a +4 Funny!)
I guess the differences in Rock styles can be explained by the age profile of people going to IT courses:
Developer profile: 25-35 years old, teenager when Iron Maiden and Megadeth were all that.
Project manager profile: 40-50 years old, teenager when Pink Floyd was hot.
Security profile: same age or slightly older than a project manager, given up hopes of ever becoming a project manager, not young enough to be a top-of-the-line developer anymore. Gone into security (and taking courses on that) because the "experience of old age" does give an edge in (a) making young developers listen to you when you give them security advice, and (b) not having enough dreams for the future anymore to let features go before security (no enthusiasm to cloud judgment), etcetera. Just the kind of person to have grown up in the days when Grateful Dead / The Doors / Jimi Hendrix were cool.
Sun must be a major sponsor of theregister.co.uk or something...
I must agree that the article is pretty positive about the whole thing. But a positive review is not necessarily a bought review. In fact, the article does contain an angle that is not likely to be sponsored by Sun: open sourcing it, and open sourcing solaris. They explicitly make fun of Sun's president, who is evading their questions about open sourcing DTrace. From the article:
So there you have it. DTrace may or may not end up in the public domain. Glad that's settled.
At my university we have some 30-40 X terminals running on one Solaris server (a quad processor server with 4G RAM, aptly named "beast"). The thin clients are Sun Rays. I don't know how much they cost, but I know the X terminal principle is very flexible compared to the 4-terminals-on-one-PC option: you can use a single server to serve terminals that can be spread out over a wide area, instead of next to the PC in question. If there would just be a cheap X terminal solution available, I think it would be a perfect solution for public libraries and internet cafes.
Regarding (2), taskbar grouping != tray minimizing. Tray minimizing minimizes to the icon list in the lower right part of the screen (which is usually called the "system tray"), so that the only space taken is an icon's worth. If he has so many open windows that he can only see the icons anyway, why not minimize them so that they take *exactly* the real estate of one icon each? OK, there's a downside -- can't Alt+Tab through those windows anymore.
I looked as his taskbar the other day, and it was full of 25-30 little Firefox icons (with a few others mixed in) followed by one, maybe two letters of the page name.
Poor guy. Some hints for him:
1. Use Firefox's tabbed browsing, preferably with the Tabbrowser extension. You can still keep a couple of browser windows open, but you can open related pages in tabs within the same browser window.
2. If you use Windows, you should consider installing a tray minimizer to get rid of those programs that you recognize by the icon anyway but that take you hours to find in the task bar if they're stuck between 30 browser windows. At work I use a tray minimizer for Outlook (for work mail), the Novell Application Launcher, Mozilla Mail (for private mail) and PuTTY (the SSH client). The added advantage of putting them there is that they are gone from my Alt+Tab list, which means I can tab between the windows I'm actually *working* with. This is the one crucial feature that I still haven't found a replacement for in Linux -- applications still have to support minimizing to the tray themselves, so it's not always possible to minimize the applications to the tray that *I* want to put there.
3. If that doesn't help enough, extend the size of your task bar by a row, so that you can better see what's there.
4. Put the taskbar on the left/right of your desktop. This takes a bit more screen real estate, but at least you can control precisely how many characters of each window's name you can read.
From my personal experience I need to make the following comment about classes. Classes are highly overrated, at least by everybody I know. People go to classes for one of the following reasons:
1. To avoid having to read the book. They think everything that is covered in the classes is important, everything that isn't, isn't. In effect, going to classes is the way to learn the least possible amount that will still get you to pass.
2. It's easy. No effort required, someone is reading the book to you.
3. Habits left over from the days they were still in school.
4. Pressure from people (parents, classmates, etc.) who think going to classes is the thing that you do when you study.
5. Social purposes -- meet classmates, etc.
6. To meet the teachers. No, that doesn't count as social purposes. It may eventually help you to get a degree.
Basically, classes are good for anything but learning what you want to learn. They can really be very demotivating because of the less intelligent people who are constantly slowing down the pace. The book is usually better written than the sheets used in class, and as an added benefit it doesn't require you to wait until next week's class to hear how the story continues.
Since about a year after I started studying at university, I simply stopped attending classes, and it hasn't hurt me in the least: my grades went up from 6ish (in a 1-to-10 system, 10 is best) to 9.5ish, it all cost me less time, and I got to read the whole book for every course, so I actually learned more. And I had time to do other stuff that I liked. As an added bonus, the courses didn't seem like a long haul anymore, because I was only working on the courses in a small number of relatively short, intense stretches. My perfect studying schedule for a course is as follows:
1. Read the book the first time (one week)
2. Let it rest for a couple of weeks. Then read it a second time and make notes of the most important things, a week before the exam (four days -- should yield about 10-20 pages of written notes)
3. Read through the notes and write down in separate notes what I didn't already know by heart by now (one day, should yield about 2-5 pages of notes).
4. Look at those last notes a couple more times until I really know those things. (half a day at most)
5. The evening before the exam, I read the full notes from (3) and (4) through once or twice. I usually notice that I actually know all of that by heart right now, so it won't take longer than one or two hours to do this.
This way of studying requires amazingly little discipline. Reading the book the first time usually comes easily, as it's only a short stretch that can be easily covered before you lose interest. The rest is done while everybody else is also cramming for their exams, which is a very short and manageable stretch as well. The only discipline required is to actually pick up the book again a week before the exam.
Copy: Ctrl-Insert Paste: Shift-Insert
(I can't remember what Cut is, I never use it.. probably ctrl-delete)
It is shift-delete, actually. The combinations make sense: for both copy+paste and cut+paste you have to only switch one key between keystrokes, not two. For copy+paste you only need to go from ctrl to shift, and for cut+paste only from delete to insert. This, of course, is all in the assumption that you move the cursor to the destination location with the mouse.:)
True, there was a lot of GUI stuff before Win 3.1. The outrageous thing is, not only did this guy not acknowledge this earlier stuff (which might be explained by ignorance), but he didn't acknowledge earlier versions of the same stuff (i.e., Windows). If something has version number 3.1, with me that rings a bell. Apparently with this guy it doesn't.:)
1. Since when did NT stand for "Network Technology" instead of "New Technology"?
2. It calls Windows 3.1 "the second OS with a GUI" (after the Mac), as if 3.1 was the first version of windows ever.
3. I quote:
Windows 3.1 was still based on MS-DOS because it was really just a front end. All it did was pass commands to MS-DOS which then passed commands to the kernel.
Excuse me? What is this "MS-DOS" thing that passes things to "the kernel"? The only thing I can think of is that he might mean the MS-DOS prompt. This sounds as if Windows 3.1 did everything by simulating typing on the DOS prompt (i.e., "pass commands to MS-DOS") and letting the DOS prompt pass things on to "the kernel". My take on this: the kernel is actually what "MS-DOS" really is -- the command prompt is just the equivalent of a shell. I have no clue what separation between "MS-DOS" and "the kernel" this guy had in mind.
4. Since when did Windows 98SE stand for "Special Edition" instead of "Second Edition"?
5. Since when was Windows ME a bugfix release for the Y2K problem? I quote: The Y2K (Year 2000) problem was discovered and fixed with the release of Windows ME (Millennium Edition). This is actually funny, so it might be intended as a joke, but I don't think it is.
6. If Windows NT was really based on the source of VMS, M$ would have definitely been sued. And they haven't AFAIK. Instead, M$ had just been done with the OS/2 cooperation debacle, and it's pretty probable that they took quite a bit of code from that to get them started on NT.
There's more I could say, but I think this enough.
Slashdot Mirror
I guess Handyman's speakers belong in that category and would do fine in the party cellar.
In fact, not really. I have a pair of Linn Keilidhs, which have a pretty clean, high-end sound to them -- not the kind of thing you want for your party cellar. I must admit that the efficiency of these speakers is pretty high for speakers in the (sort-of) high-end segment.
Also, when comparing speaker performance, Watts are definitely not the complete picture.
The first thing you need to find out is the efficiency of speakers. For instance, my speakers have an efficiency of 92 dB/W, which means that at a power level of 1 W, they will produce 92 dB of sound. As dB is a logarithmic scale, doubling the wattage will increase the number of dBs by 3, so a power level of 64 W will get me a 92 + 3 * 6 = 110 dB sound level. However, a speaker with an efficiency of 80 dB/W will only produce 98 dB for the same amount of power. I've seen efficiencies ranging from 70 dB/W up to the high ninety-somethings, so be careful to check these numbers.
The second thing you need to find out is the impedance of the speakers, combined with the impedance your amplifier is rated for. For instance, my amplifier is not simply rated as 50 W, but as 50 W for speakers with an impedance of 8 Ohms, and 100 W for speakers with an impedance of 4 Ohms. This can make some difference. Watch out with getting a speaker with very low impedance though: if your amplifier wasn't designed to handle that, they will probably draw too much power, causing the amplifier to get overheated. In addition, you will not be able to open up your volume knob more than a couple of millimeters -- and volume is probably something you like to have detailed control over.
Trust me. It isn't.
OpenOffice splash screen is not always-on-top anymore. Check the latest version.
Not true, unfortunately. I run 1.1.3, which is the latest version, and the splash screen is always-on-top. Yes, there's 1.1.4 RC, but as far as I'm concerned that's not the "latest version", that's the "development version".
Yes, a long delay without a splash screen is irritating as well. But there is nothing as irritating as a splash screen that is always-on-top. What were they THINKING making the OOo splash screen always-on-top? They already have a reputation for being the slowest starting office product out there, must they rub it in your face then?
When taken out of context this is a gem, worthy of inclusion in my fortune database:
You're an adult, and you can make your own choices. That is when you get engaged.
The Windows Template Library is released as CPL as well. I don't know since when this has been, but the product has been in existence for quite some time now.
Kaspersky labs says they were misquoted. Quoting from a mail from kaspersky labs themselves (as found in a repost on the NTBugtraq mailing list):
A handful of sites are stating that Eugene Kaspersky, founder of Kaspersky Labs, believes that tomorrow will bring a massive terrorist attack on the Internet. This is being quoted in a range of ways, ranging from factual reporting to citing the story as an example of cyber hysteria.
However, Kaspersky is not predicting the end of the Internet tomorrow - or even in the near future. The story stems from brief comments made yesterday at a press conference which was dedicated to cybercrime and the problems of spam.
At this press conference, Kaspersky commented that the possibility of terrorists using the Internet as a tool to attack certain countries as a reality. As an example, he cited the fact that a number of Arabic and Hebrew language websites contained an announcement of an 'electronic jihad' against Israel, to start on 26th August 2004.
In an interview today, Kaspersky stressed that such information was not necessarily trustworthy. 'We don't know who is behind these statements.' He went on to clarify: 'It's not the first time the term 'electronic jihad' has been used. We've seen this before, with the focus being on sending racist emails, and defacing and hacking Israeli web sites. But it is the first time I have seen sites encouraging the use of Internet attacks against one country as a form of terrorism.'
'As we've already stated many times in the past, it would be easy enough to use a network of infected computers to launch such an attack. We saw the impact that Sasser, Mydoom and Slammer had, on the Internet, businesses and organisations. Just imagine if such an attack was directed at one country or one critical point in the infrastructure of the Internet. Computers are a tool - and just like any tool, they can be used or misused.'
Kaspersky emphasised that the likelihood of a massive attack directed against Israeli institutions tomorrow is low. However, he believes that Pandora's box has now been opened. Hackers and virus writers can be motivated by a range of factors: money, curiosity, or political
conviction. But whatever their motivation, the insecure nature of the Internet and weak security precautions offer a wealth of opportunities. 'Maybe it won't be tomorrow, or the day after tomorrow - but sooner or later, terrorists will be using the Internet as another weapon in their arsenal.'
Most people don't jump up to enter their username and password as soon as the prompt appears. They turn the computer on, walk to somewhere else and such, and when they get back, all services are finished loading and it's ready and waiting for them.
In addition, processing X windows password GUI is not an extremely intensive task, and selecting the username and typing the password will take at least enough time for some of the background work to be finished. The only thing that is really overlapped is the finishing of the initialization of services and the user's interaction with the display manager -- a very processing-intensive task with a very user-intensive task. Smart. Much smarter than overlapping multiple processing-intensive tasks, because the resources involved are different -- user time versus machine time.
Oh, please. All Microsoft did was make the desktop appear sooner, when there's still a ton of shit to load.
:) Seriously, my girlfriend's Athlon 750 machine runs XP like a charm, and after booting and automatically logging in, the desktop is immediately very responsive. I don't know what you've done to your system to make it so slow. Are you running one of those virus scanners (note to Linux users: those are mandatory on windows machines to keep them running for more than a day) that scans everything that loads and that slows your machine down by >50%?
That is not true. This page indicates that "Windows XP initializes device drivers in parallel to improve boot time." And this page indicates that not only does Windows XP do prefetching of data required at boot time, but it also optimizes the data so that it is placed sequentially on the outer edge of the disk.
Note that you assume the situation where the system boots directly into a desktop, i.e. a configuration with automatic login. This may cause the background loading to be unfinished when the desktop is already loaded. But only on systems that run too many services. Those machines are apparently servers, and you'll NEVER want to configure those with automatic logins.
I've got a home-built Athlon XP 2600 system that is running XP Pro, has an assload of RAM, and is kept well-tuned and malware-free, and that fucker takes a good 30-45 seconds after the desktop appears before it actually pays attention to any of my attempts to do anything.
Then you must be one lousy system tuner. No wonder you posted this as AC.
For the "feel" of a machine, latency or "response time" is the most important factor. When the user requests an action, it is the time between the request and the machine's response that counts. For instance, the almost 2x speedup of booting XP means a 2x decrease in a very annoying latency, and it makes the system feel much faster even if nothing else changes. The numerous *small* latencies in a system also count -- don't you hate it when you click a menu and you have to wait a full two seconds before it pops up? The improvements measured in the benchmarks done by tweakers.net don't do justice to the importance of latency. The user doesn't care whether some background process (e.g. eMule) is fast -- he cares whether if he clicks a button, the result will show up with or without a noticeable delay. So what they should really be measuring is the time between certain checkpoints in a trace, e.g., the time between the point where the user did something (action) and the time when all the necessary data to respond to that action has been read (response).
Note that the 2x speedup can be easily explained. Windows XP optimizes the boot process by automatically generating traces of disk accesses done at boot, and by reordering the accessed blocks on disk so that they can be read in sequentially in the next boot. And striping over two disks theoretically improves sequential read throughput by... yes, a factor 2.
About compiling in Visual Studio: that can be fixed. A colleague of mine wrote a small tool that monitored the system for processes named CL.EXE and LINK.EXE, and that immediately set them to idle priority whenever they appeared. Problem solved.
The real problem is of course in the Windows scheduler. The Windows scheduler simply gives the _foreground task_ a fixed boost for being interactive, and if you're working in Visual Studio, then Visual Studio is the foreground process. The compiler/linker processes are children of the foreground process, so it wouldn't surprise me if those processes inherited that "foreground boost", making it impossible to do anything else on the machine...
The Linux 2.6 scheduler is WAY better at determining which tasks require an "interactivity boost". Also, the interactivity boost given by Linux ultimately influences only how soon a process will be scheduled (latency), not how much processing time it will get (throughput). This is caused by the metric for interactivity, which boils down to "how often does a process give up the processor". This works because interactive processes tend to do short bursts of processing and then stop again (e.g., repaint the screen, process a keystroke, etcetera), whereas noninteractive processes tend to continue processing until they're completely done.
In the good old days (this would be the early 90s), I bought an MFM controller and an accompanying 20 MB hard drive. Little did I know that an MFM controller and an IDE controller wouldn't work together in the same box. I was disappointed that it didn't work, and, not knowing much about hardware then, I figured that the controller card might need some power as well. So I plugged a small power connector (that is normally used to supply power to the 3.5" disk drive) onto the four pins that came out of the MFM controller. When I turned on the computer, the hard drive spun up a lot slower than normal. Then black smoke came from the connection between the power and the MFM controller. I quickly turned the computer off again, and after removing the MFM controller the computer still worked. Phew! When I checked what the connector actually said, it turned out that I had connected the power to the MFM controller's hard disk led connection. :)
I'm a programmer and I'm listening to Pink Floyd while I write this, yet I'm closer to 20 years old :).
:)
OK, I can see why you wrote this anonymously. Don't want to upset the electro crowd at work.
(On a sidenote: I'm a programmer as well, doing mostly prog rock. I was considering to write this anonymously because *grrrr* if someone mods you down as overrated after you've been modded up as funny, you *lose* karma in this wicked system. If anyone is listening: the karma increase should be something like regular_positive_mods - max(0, overrated_mods - funny_mods). The current system not only discourages the semi-funny running-gag posts (e.g. 1. You're in Japan, 2. You move to soviet Russia, 3. Profit!!!) but also the seriously originally funny posts, because if you get modded up to +5 Funny, then +4, Overrated, you lose karma, even if you _were_ really worth a +4 Funny!)
I guess the differences in Rock styles can be explained by the age profile of people going to IT courses:
Developer profile: 25-35 years old, teenager when Iron Maiden and Megadeth were all that.
Project manager profile: 40-50 years old, teenager when Pink Floyd was hot.
Security profile: same age or slightly older than a project manager, given up hopes of ever becoming a project manager, not young enough to be a top-of-the-line developer anymore. Gone into security (and taking courses on that) because the "experience of old age" does give an edge in (a) making young developers listen to you when you give them security advice, and (b) not having enough dreams for the future anymore to let features go before security (no enthusiasm to cloud judgment), etcetera. Just the kind of person to have grown up in the days when Grateful Dead / The Doors / Jimi Hendrix were cool.
Or am I way off the mark here?
Sun must be a major sponsor of theregister.co.uk or something...
I must agree that the article is pretty positive about the whole thing. But a positive review is not necessarily a bought review. In fact, the article does contain an angle that is not likely to be sponsored by Sun: open sourcing it, and open sourcing solaris. They explicitly make fun of Sun's president, who is evading their questions about open sourcing DTrace. From the article:
So there you have it. DTrace may or may not end up in the public domain. Glad that's settled.
At my university we have some 30-40 X terminals running on one Solaris server (a quad processor server with 4G RAM, aptly named "beast"). The thin clients are Sun Rays. I don't know how much they cost, but I know the X terminal principle is very flexible compared to the 4-terminals-on-one-PC option: you can use a single server to serve terminals that can be spread out over a wide area, instead of next to the PC in question. If there would just be a cheap X terminal solution available, I think it would be a perfect solution for public libraries and internet cafes.
Regarding (1), sorry, that was a reado.
Regarding (2), taskbar grouping != tray minimizing. Tray minimizing minimizes to the icon list in the lower right part of the screen (which is usually called the "system tray"), so that the only space taken is an icon's worth. If he has so many open windows that he can only see the icons anyway, why not minimize them so that they take *exactly* the real estate of one icon each? OK, there's a downside -- can't Alt+Tab through those windows anymore.
I looked as his taskbar the other day, and it was full of 25-30 little Firefox icons (with a few others mixed in) followed by one, maybe two letters of the page name.
Poor guy. Some hints for him:
1. Use Firefox's tabbed browsing, preferably with the Tabbrowser extension. You can still keep a couple of browser windows open, but you can open related pages in tabs within the same browser window.
2. If you use Windows, you should consider installing a tray minimizer to get rid of those programs that you recognize by the icon anyway but that take you hours to find in the task bar if they're stuck between 30 browser windows. At work I use a tray minimizer for Outlook (for work mail), the Novell Application Launcher, Mozilla Mail (for private mail) and PuTTY (the SSH client). The added advantage of putting them there is that they are gone from my Alt+Tab list, which means I can tab between the windows I'm actually *working* with. This is the one crucial feature that I still haven't found a replacement for in Linux -- applications still have to support minimizing to the tray themselves, so it's not always possible to minimize the applications to the tray that *I* want to put there.
3. If that doesn't help enough, extend the size of your task bar by a row, so that you can better see what's there.
4. Put the taskbar on the left/right of your desktop. This takes a bit more screen real estate, but at least you can control precisely how many characters of each window's name you can read.
From my personal experience I need to make the following comment about classes. Classes are highly overrated, at least by everybody I know. People go to classes for one of the following reasons:
1. To avoid having to read the book. They think everything that is covered in the classes is important, everything that isn't, isn't. In effect, going to classes is the way to learn the least possible amount that will still get you to pass.
2. It's easy. No effort required, someone is reading the book to you.
3. Habits left over from the days they were still in school.
4. Pressure from people (parents, classmates, etc.) who think going to classes is the thing that you do when you study.
5. Social purposes -- meet classmates, etc.
6. To meet the teachers. No, that doesn't count as social purposes. It may eventually help you to get a degree.
Basically, classes are good for anything but learning what you want to learn. They can really be very demotivating because of the less intelligent people who are constantly slowing down the pace. The book is usually better written than the sheets used in class, and as an added benefit it doesn't require you to wait until next week's class to hear how the story continues.
Since about a year after I started studying at university, I simply stopped attending classes, and it hasn't hurt me in the least: my grades went up from 6ish (in a 1-to-10 system, 10 is best) to 9.5ish, it all cost me less time, and I got to read the whole book for every course, so I actually learned more. And I had time to do other stuff that I liked. As an added bonus, the courses didn't seem like a long haul anymore, because I was only working on the courses in a small number of relatively short, intense stretches. My perfect studying schedule for a course is as follows:
1. Read the book the first time (one week)
2. Let it rest for a couple of weeks. Then read it a second time and make notes of the most important things, a week before the exam (four days -- should yield about 10-20 pages of written notes)
3. Read through the notes and write down in separate notes what I didn't already know by heart by now (one day, should yield about 2-5 pages of notes).
4. Look at those last notes a couple more times until I really know those things. (half a day at most)
5. The evening before the exam, I read the full notes from (3) and (4) through once or twice. I usually notice that I actually know all of that by heart right now, so it won't take longer than one or two hours to do this.
This way of studying requires amazingly little discipline. Reading the book the first time usually comes easily, as it's only a short stretch that can be easily covered before you lose interest. The rest is done while everybody else is also cramming for their exams, which is a very short and manageable stretch as well. The only discipline required is to actually pick up the book again a week before the exam.
Copy: Ctrl-Insert
:)
Paste: Shift-Insert
(I can't remember what Cut is, I never use it.. probably ctrl-delete)
It is shift-delete, actually. The combinations make sense: for both copy+paste and cut+paste you have to only switch one key between keystrokes, not two. For copy+paste you only need to go from ctrl to shift, and for cut+paste only from delete to insert. This, of course, is all in the assumption that you move the cursor to the destination location with the mouse.
True, there was a lot of GUI stuff before Win 3.1. The outrageous thing is, not only did this guy not acknowledge this earlier stuff (which might be explained by ignorance), but he didn't acknowledge earlier versions of the same stuff (i.e., Windows). If something has version number 3.1, with me that rings a bell. Apparently with this guy it doesn't. :)
Boy, that's a lousy article:
1. Since when did NT stand for "Network Technology" instead of "New Technology"?
2. It calls Windows 3.1 "the second OS with a GUI" (after the Mac), as if 3.1 was the first version of windows ever.
3. I quote:
Windows 3.1 was still based on MS-DOS because it was really just a front end. All it did was pass commands to MS-DOS which then passed commands to the kernel.
Excuse me? What is this "MS-DOS" thing that passes things to "the kernel"? The only thing I can think of is that he might mean the MS-DOS prompt. This sounds as if Windows 3.1 did everything by simulating typing on the DOS prompt (i.e., "pass commands to MS-DOS") and letting the DOS prompt pass things on to "the kernel". My take on this: the kernel is actually what "MS-DOS" really is -- the command prompt is just the equivalent of a shell. I have no clue what separation between "MS-DOS" and "the kernel" this guy had in mind.
4. Since when did Windows 98SE stand for "Special Edition" instead of "Second Edition"?
5. Since when was Windows ME a bugfix release for the Y2K problem? I quote: The Y2K (Year 2000) problem was discovered and fixed with the release of Windows ME (Millennium Edition). This is actually funny, so it might be intended as a joke, but I don't think it is.
6. If Windows NT was really based on the source of VMS, M$ would have definitely been sued. And they haven't AFAIK. Instead, M$ had just been done with the OS/2 cooperation debacle, and it's pretty probable that they took quite a bit of code from that to get them started on NT.
There's more I could say, but I think this enough.