The only part of this deal that I don't like is turning over the domain name.. just what does Lindows have to do with Windows?
Of course, I guess that's what this whole case was about =)
A quote from the article is "SFU is not shipped with Windows because SFU currently contains open-source software, such as the GNU C compiler, which cannot be distributed with commercial software. Zions confirmed that Microsoft is working to replace all open-source code in SFU with commercially licensed alternatives."
My money is that about the time Longhorn is released (+/- five years) we'll see another offering, Lindows, which is Microsoft's *nix clone, which also happens to have binary compatability with large swaths, but not complete coverage, of Linix internals. Just enough that the average consumer and most Judges will say "See, they're still competing and not illegally using their Monopoly to do it."
Your questions aren't really answered by the 'article', other than the general guidelines like "be polite" and such.
1. If you have call display, is it polite to answer the phone with the caller's name? I couple of years ago, this freaked people out, now it's very common. Older people tend to think of this as an invasion of privacy; but these same people consider it acceptable to have a peephole on their door and only open it to people they know.
Just becuase you have caller ID or cubicle ID, doesn't mean you should assume that the name is the caller. The ID might say it's Joe, but in reality, it's your boss who is over at Joe's cube. Do you really want to answer the phone with "Joe, will you let me do some work?" In a business environment, its usually best to answer with "This is Bob" or "Bob Speaking", unless there is some policy for answering phones. If there is a company policy, follow it, even if you know who it is.
2. What about call answer? Should you take the incoming call and how long should you be on it before returning to the original call? What happens if you consider the second call to be more important than the first?
Personally, I think that an interruption should be no more than 30 seconds. Then tell one of them that it will be a while, and ask that person to call you back, or ask if they want to wait until you are done. Just ask yourself how you would feel on the other end. Anything more than a minute on hold feels like forever. At one place I worked, after two minutes on hold, calls were automagically routed to a bunch of general operations people, so that customer's knew they wouldn't be lost in the phone system.
3. Is call screening using an answering machine polite?
Some people refuse to leave messages, no matter what. Even if your machine says "I'm screening my calls, leave a message", they still won't leave a message. If you must use a machine, make it easy and obvious as to how to get a real person. (Note: menu mazes are NOT polite. Not in corporate environments (where they do screen out a lot of junk), and not on home machines that support them with "press 1 to leave a message for Bill, 2 for Jane, 3 for...".)
4. I give telemarketers one chance to hang up before I slam the receiver down on them. Is this polite or should I listen to their pitch? Can I blow a Fox-40 whistle into the receiver?
Um, no. Causing the caller to go deaf in one ear is not a good thing. Remember -- They could be an unemployed geek who took a telemarketing job to pay the rent: It could be you in six months. I think the 'correct' way to do it is to interrupt them immediately with "Please put me on your do not call list."
5. What are cell phone rules? Is it acceptable to have a social call while in line at the supermarket? What about a heated business call?
If you're from Europe, then heated business calls are best done in public restrooms.:) If they call you and you're the 24-hour on-call techie, then it's their own fault. If you decide to take or make calls when out in public, just remember, you're still out in public. Don't shout to your cell phone when others are nearby. Don't talk on it when it's dangerous, such as when driving or operating a wood shredder. Listening in on one-sided conversations is very boring, don't do it when there are other people who you SHOULD be talking to, unless they understand that the call is important.
And finally...
NEVER EVER stand in aisle 4 and call your sibling/spouse/parent who is on aisle 5 to ask "Did you want regular or diet?"
Boeing and Airbus invest millions of dollars into software that takes years to test and develop.
I live near a major AFB and know several commercial and military pilots. I hate to break it to you, but some of my friends (both types of pilots) say that their computer systems routinely crash; I've heard some rather intense stories from the fighter-pilot types who have lost control of their aircraft and nearly crashed, due to computer glitches.
Text and BCD formats were popular because they were efficient.
That's all true, but when you want to compute an amortization table and tell the computer that the last payment will be made 70 years in the past rather than 30 years in the future, you're likely to get a very odd payment schedule.
My year 10 computing teacher told us a story about how in the 60s they used a single number to store the year, and when they got to 1970 they were like "wait a sec!"
Sounds like your year 10 computing teacher needs to take some more history lessons. Many of the groups who first dealt with computers (banks) were hit with the Y2K bug right at 1970, since they couldn't do 30-year loans, and many had already considered the problem. And since these companies counted out memory by the byte (or rented memory by the byte) they certainly wouldn't have been storing years as simple text. If they were, the programmer at the day would certainly have been considered wasteful, and possibly even fired for such practices.
1: Super geek. "Hired" pile
2: Benevolent dictator. "Possible" pile.
3: Idiot. "No way" pile
4: "Feed to goatse"
Having been able to hire my own boss twice now (it's a weird process), there should only be one pile. The "Hireable" pile, or the trash. If there is anything you aren't sure on, don't keep the candidate around. There are plenty of people right now out of work, so it shouldn't be too hard to find somebody.
As far as the/. question, would you really want to have a boss that gets after you for your present practices? If the boss doesn't understand the importance of a little bit external social interaction, like slashdot, then they probably won't understand other significant things in your culture.
(Though I wonder how the ESRB would handle rating this with the "shake" mode included....)
Simple, the same way they handle other games.
They'll put that the game is rated 'E', since video conferencing doesn't seem to meet any of the ratings elements, and keep the little warning that says "Game Experience May Change During Online Play"
I don't want to spend money on 3D acceleration on my mobile phone. I begrudge even having to pay for a colour display!
I agree. After shopping around, it seems you cannot find any phone company that sells a phone without all the extra garbage.
I'd be comfortable with the really-old style number LEDs that were common when cell phones (usually called 'car phones') were first becoming popular. I want to have a phone to talk to people. Seriously, just look at some of the stuff built into
These phones. I dont want to have the 'built in' cost of:
bluetooth
Infrared
big screen
color screen
touch screen
3D graphics processor
Address book
web browser
GPS
Digital Music
Java processing
Camera, for streamed video and audio
RealOne, MP3, AAC, Midi, Wav, and a dozen other formats supported
A bunch of games (Tony Hawk's Pro Scater, Tomb Raider, Sonic, MLB Slam!, Super Monkey Ball, more...)
stereo audio output with headphones
All this and more, for the bargin price of under $200 after a $200 rebate with your 2-year Service Contract.
When I searched for free phones at several places, the free (after rebate or with contract) phones all had at a minimum:
big screen
web browsing
voice dialing
Phone & address book
Several games
vibrating alert
GPS (for E911)
TTY/TTD
I'd still be paying for all these features, regardless of whether I want them, even if they payment is every month through the contract. I don't want them. I just want one of the big grey bricks that I had a decade ago, that you typed in the numbers and it called people. Other people could call you.
And that's it.
But the major companies WON'T do that, since that means they can't find convenient ways to charge you $50+ monthly. Well, there's always prepaid wireless...
Your life is worth infintely more than any amount of money. People who do this must be crazy. Do you really want to be beheaded because you want a new Mercedes?
For some people, the thrill and adventure and feelings of "I'm helping people" are more important that the money. There are some people who aspire to military or civilian service in war areas. Just because the path isn't one YOU wouldn't choose, there are a lot of people who do choose that.
Look at all the Red Cross / Red Cresent people who jumped in to the danger just because that's what they want to do, and they see the value of helping rebuild a country or helping others as greater than the risks. Other people jump in and will risk everything for money, or just for the thrill. Obviously YOU wouldn't but some people find that kind of life appealing.
The question was asking for an account of real dangers based on people with experience, not for beliefs on the value of life.
Or maybe it's because you are using them for useless things. There are plenty of uses where PDA's are *much* better than laptops.
The company I just left was using them as a 'portable screen' for hardware devices, and that's getting to be a more common thing.
When the engieer or repair person has to go out in the field, laptops are a real pain. When you have to be up a ladder or on a cherry-picker hoist, it becomes a chore to boot up a laptop, balance it on your knee or ladder or box or whatever (hoping it doesn't fall, which DOES happen), configure the hardware, turn off and disconnect the laptop, etc. It is MUCH easier to use a PDA (in our case it was Palm devices) connect the wires on both ends, hit power (instant on!) check off a half-dozen settings, and pull the cords out.
They cost a LOT less to replace, and they take a lot more abuse than most laptops. If you drop them 20 feet onto the grass, they usually survive. Their batteries aren't drained after 3 hours in the field, and they're a lot easier to carry, especially if you have a belt holster. You can usually use your finger rather than a stylus, where you needed two hands to type on the laptop leaving none to hold it. Thats only a few of the benefits.
My guess is that you'll start seeing this type of use increase dramatically -- it saved us quite a bit of both time and money.
[Go actually do something and prove yourself.
Until then, reel in the ego before you get pounded.
I know it's cruel but each year around this time I find great perverse pleasure in smashing the egos of poor unsuspecting recent grads. Nothing too harsh, just when they come in with some ego-dripping report like "This is perfect" or "It works perfectly" or (my personal favorite) "I dare you to find a flaw in that!"
One of my favorites was the first peer review of a young intern, only about a hundred lines of source code but probably the biggest thing he'd ever written. I could see something either wrong (buffer overflows, magic numbers, etc.) or missing (no error checking, no comments, no assertions,...) on nearly everything he did.
I know it's cruel, but to watch their over-exuberant pimply faces go from "Joy! I'm doing Important Things!" fall to "All the things I've been gloating over mean absolutely nothing" is oddly fun.
Of course, I follow the rule "reprove in private, praise in public" (with another team member present), I try to do it early, before they have much emotional investment in their work (usually after a few days of work), and always do ego rebuilding and team building afterword. Usually something like "But it's a good start, considering you just graduated. Let's sit down together and fix a few of those holes."
Who are you, Richard Stallman? Make up your own damn word instead of appropriating one and then "correcting" everyone who uses it normally.
I think it's a pretty common thing. Just look through the Fortune's company list: Abbey, Boots, Continental,... Kingfisher, Lend Lease, Metro... United, Visa, Wendy's... All of them international corporations, and a few, like Visa, attempt to rewrite the dictionary to remove the common word.
So why shouldn't the scientoligists or whatever group the person rewriting 'bright' is, join in the fray?;-)
Personally I think we should all take the words "slash" and "dot", and try to... oh, wait.
why are you running KDE on a K6? XP would bring that box to it's knees too.
No it wouldn't. My father ran a test XP system back out when it was still Whistler on a 400MHz or so Pentium II system with 256MB of RAM. It ran absolutely fine.
I have a small network which I use for testing of my programs. My two test systems are a 486-100 running Win98, and a multi-boot P2-300 running WinXP (usually), 2000, Win98, and linux. Both of these run with reasonable peformance on their own. I like to think they are reasonably well used and installed upon. One of my kids uses the 486 to play RollerCoaster Tycoon and surf the web on. I use the P2 for smaller tasks; it is usually booted into XP and has copies of WordXP Developer (showing design specs with embedded images), DevStudio, MSDN, and Internet Exploder running all the time, in addition to whatever other programs I am testing.
My P2-300 is next to a P4-2Ghz. While there is an obvious speed difference, it isn't all that big. I use them both for development, and don't bother replacing it because it is 'good enough'. Usually I click the power buttons at the same time, log in to the P4 box, while I wait it to load all the startup garbage, I log in on the P2-300. Then I start my few apps on the P4, then on the P2. When I do remote debugging on the P2 (nothing really CPU intensive, mostly RAM- and disk-intensive database work) it is just as functional speed-wise as the P4. Sure, it helps that I have an ATA-66 disk and 192MB of 133 ram inside, but that shouldn't be a big thing in that age of system. The speed difference is very noticable when I do CPU intensive things, like video conversion and games, but I have only have done video conversion a few times and only play a few games like networked Warcraft 2, which also plays fine on it. Sure, it takes an extra 20 seconds when I compile or an extra few minutes for a full rebuild, but that's okay for me.
When I do remote debugging in the win98 486 box (I wouldn't dream of a full DevStudio install there), the big problem is coping with the relatively small amount of memory (64M) and finding ways to reduce the thrashing to swap space. The good thing about that is that I know my programs run well on fast machines and reasonable on slow/old ones.
Just to put it in comparison, the P2 is a duel boot into linux, since several linux-based games are fun. When I decided it might be nice to try the 2.6 kernel in the debian 'sarge' installation a few weeks ago (mid-March), I only used it for two days before re-formatting and going back to a debain 'woody' installation without most of the server components. I use KDE, and it feels slower than what XP feels like. I don't care to use another window manager because of the time to install & configure, I'll just take what comes out of the box, thank you.
For fun I just booted it into linux, started OpenOffice, Mozilla, and Anjuta. The system feels slow, and the HD sounds like it's going to blow up when they were starting up. Surfing to a few web pages (userfriendly,/., sourceforge, openoffice.org) makes me worry about the disk, it's just clattering away. Loading a design spec word document from the common windows partition just took about 45 seconds, and it stalls when I scroll through it. Eeek. It is awful compared to what I do on XP. Oh well, back to Windows. Well, a few games of Frozen Bubbles first.:-)
You are using a book, which has sentances which have much more structure than that data was. Yours in the form {1 CAP}{~50 lower case}.
In that particular data set, It was around 0-10 caps with 0-20 non-caps, for each entry. They were using an inconsistant entry rule, some all upper, some all lower, some mixed, mostly abbrviated words to fit on receipts.
It makes a difference. Even in your less relavent example, caps made a 3% difference.
C'mon, it's my own post, and I caught at least three typos in it. If your going to correct my spelling, please at least take the time to find ALL of them; Finding only some of them makes you look dumb. One word has two different spellings, only a few words away from each other.
Caps are more easily readable on envelopes, some database and spreadsheet entries, titles of chapters, and many more writing applications.
I work with several different databases. All of our database data is stored in caps. Why? In decreasing order of significance:
Postal Service Discount Requirements - Unless you like spending lots of money on mailings, use one of the discount mail formats, such as automatable mail. Automatable mail means CASS cerification, which means using all caps. Otherwise, the best bet is using an all-caps, 8-18 point sans serif, uniform stroke width font, 3/4 to 3 point space between characters. We found that 10 or 12 point ariel fits the requirements. If you use bulk mailing and you don't fit the requirements, the mail can be thrown out, not returned, so you wouldn't know that your mixed-case address had problems.
Compression size - Your compressor should notice that almost all the data are of about 30 values {0-9, the common A-Z values, space}, and a few other symbols for the database system. That tiny alphabet, and considering the rarity of certain alphabet letters, will make dynamic huffman VERY happy. Mixed case would at least double the size of the compressed alphabet, meaning larger backups. If you do backups (you do, don't you?) it helps. If you have to distribute the database to different sites, it really helps. In one example I encountered recently, the client had about 6 GB of compressed data and wanted to send it to all of their stores nightly. Each store is equipped with a dedicated high-speed line, but 6GB daily (over 9 hours on dedicated T1) is too much for a nightly transfer. After looking at their data, simply converting everything to caps let it compress to just over 1 GB of data (under 2 hours over a T1 network).
Uniform - It's easier to search for names that some people like to have different capitalization rules on (Shrudili/ShruDili/ShruDiLi, MacKay/Mackay, etc) Yes, you can do a case insensitive match (you insensitive clod) but it takes more CPU work; That's bad when a batch run already takes a long time for simple case sensitive matches.
Readability - It's much easier to read names and addresses on the screen when they aren't mixed case.
The first one will save you lots of money in mailing. If your company is in to saving money, then your mailing lists already do that. Compressability of data is *very* important for large-scale systems, and can have significant improvements with an ALL CAPS DATABASE. The last two just help reduce system load and operator error; both are good things to reduce.
I've never understood the 'need' people have to capitalize SQL. It's not case sensitive. If one has a decent syntax highlighting editor, then the capital's don't help at all!
It's a style thing. It's also the common style used almost everywhere.
You say it doesn't help if you have a syntax highlighting editor; you can't count on having a syntax highlighted editor all the time.
While I think there are benefits to it, many interns that I've worked with seem paralyzed if they don't have syntax highlighting. A few are paralized if they don't have auto-complete. If the code REQUIRES syntax highlighting to be understandable, clearly you need to adopt a different style.
If you don't have SQL syntax highlighting, the more complex the SQL statement is the more it benefits from the all-cap keywords. That in itself is enough to justify (for me) it's use. Since I compose my SQL statements in my C++ interface code, I don't have SQL syntax highlighting; the C++ editor highlights them all as strings. Using caps really helps readability, especially when the SQL statements are complex and require several lines to compose.
So basically you're sayin that if there were people in your neighbourhood bullying, racketing, killing, raping and suddenly a few powerfull "families" decide to "regulate" the neighbourhood and also the money you owe them on a regular basis and also who can do what business around, you would consider that a big step?
It sounds like a form of government or policing to me, depending on what the powerful groups were/are doing.
If the ones doing the bullying, killing, and raping are the powerful then you have something like Russia under Ivan the Terrible.
If the ones doing the regulating are the ones needing the protection and are seeking revenge, then you know the problem will go away (although new ones will appear)
If the ones doing the regulating are the ones needing the protection and are seeking to install an equitable justice, then you have a good situation.
The question is, which group do they fall under: The Terrible, The Vengeful, or The Just? I would guess either the first or last, and since the companies probably aren't willing to take too much risk, they'll at least tend to be more akin to 'The Just' rather than 'The Terrible'.
... if you put it on the Internet, you're a moron...
We can start speculate about the impact of some Windows HPCs infected with some worm on the Internet. And then speculate further and compare that impact on what we alredy experienced with those hacked Solaris and Linux clusters.
Okay. Let's assume the university supercomputer gets a nasty worm. It sends it out on it's gigabit connection to the university switches. At that point, it is competing with a few thousand other windows-worms (what? You think they'd be installing a Windows-based supercomputer in an all-Linux school?)
The infected university supercomputer is really not that different from the infected campus, which admins have been dealing with for years.
Finally, many university supercomputers are already on the Internet. I've worked with 7 supercomputers at three schools, all had their head (node zero) accessable through the Internet. That was where we logged in, compiled and tested our jobs, and queued them up for running on the rest of the network. Two of the clusters had each node internet accessable, and (interestingly for this topic) one of the clusters, a bunch of multi-processor P3s, was running Windows NT on each box.
[some talk about speed tests]
Bytes=13 or 14*Bits... A standard TCP header is 120bits (15 Bytes) IIRC. [bunch of messed up math]
How your post ended up as +5 informative I'll leave up to the imagination.
I would suggest that you look up either the actual RFCs or an authoritative reference like 'TCP/IP Illustrated.' The grandparent post was wrong, and your correction is also wrong. Let's look at the facts.
First, the application will probably encode the real data within their own packet. FTP can use compression, and can (but doesn't often) use higher encodings and structures. It's data is then wrapped into a TCP packet (assuming TCP).
A TCP header is between 20-60 bytes, depending on what flags are set. (Normally it's 20 bytes). Also, there is a minimum data size which is replaced by padding if your packet is really small.
The TCP packet is wrapped inside an IP packet, which is also a 20-60 byte header (normally 20).
The now 46-1500 byte packet is then wrapped and unwrapped as it travels across each network. Presumably within your own network, this is a series of ethernet packets with a 14-byte header and 4 byte trailer (18 byte total).
Once you get into ethernet collisions, TCP retransmittals, packet fragmenting, header compression, ICMP messages, etc., the number of bits/bytes actually on the wire doesn't fit into some formula for finding the actual bytes of application-level data being recieved. On a noisy line, it is entirely possible to transmit a gigabyte of packets thorugh TCP/IP and still not receive even a signle byte of application-usable data.
A blanket statement of assuming 8 bits of application data for every 13 or 14 bits of wire data is pretty bad. That's saying 5/13 or 5/14 (about 37%) of your data is headers, control packets, retransmission, etc. On a clean transmission, it could be 40/1500 (about 2%) or potentially even lower if there is header compression or bigger MTU. Or if you are transmitting one byte of data at a time, it could be worse than 45/46 (98%) spent in overhead. Or you could have the amount you stated, or you could have 100% of it as controlling and lost, and get no useable data out.
Even on an exclusive and clean line, you have a range of below 2% to over 98% overhead, which is a huge spread. A little noise on the line can also have have a huge impact. Even though the 'speed test' programs transmit a large file, something that presumably uses large packets, the amount of usable data coming over the line can and does vary based on many factors. Simply assuming 10-15 transmitted bytes for each useful byte MIGHT be useful for you as a very rough guess, but the only way to tell for certain would be to monitor on your own system the usable data passing through each datagram, and compare it to what the hardware says was actually transmitted.
These speed tests are only useful at diagnosing that you are getting roughly the speed that you paid for on a fairly noise-free connection, and for ego-trips. Nothing more. And certainly not for evaluating in a blanket statement that there are x bits of overhead for y bits of data.
On my drive home from work I pass a farm selling "96% fat free milk".... I think 96% fat-free should have 4% of the fat of 'normal' full fat, not be full fat milk.
I can just see the following dialog:
customer:What is this solid block of lard doing in the dairy section?
clerk:That's not lard, it's non-fat milk fat.
customer:What is the solid block of not-really milk fat doing in the dairy section?
clerk:It's the new Atkin's Brand 100% Milk for Lactose Intolerant. It's popular among teenage girls and dieters.
It's been my experience that the idiots are the ones making the purchasing decisions, hence the nature of the advertising.
It's not just the idiots. If you didn't know anything else about the product, which would you buy?
Product A -- Claims to be 73% good.
Product B -- Claims to be 96% good.
Product C -- Claims to be 99.999% good.
Product D -- Claims to be 100% good.
Being skeptical, you would probably pick product A has having truthful ads. Product B, you might think, has really good real-world performance. Product C is just marketing hype, and product D is impossible in the real world.
But if you see a big brand name (Microsoft, Cisco, Intel, etc.) on product C, you might say "Well, it isn't 100%, and they are a good company. Maybe it's the truth. Of course, claiming to be Product C happens, and that's where the trap is.
In some cases it really is justified. If I were a chemist, maybe having iridium wire that is only 99.9% pure might cause problems, and those extra 9's might be significant. But that usually isn't the case for most marketing.
But I don't think it's just a PHB issue, it's a problem of 'I really want the best, and I only want to spend 5 minutes to find out which one that is'.
Just now I'm working on an econometric model... in real life applications, playing with a Targa file is not the same as service critical, 300 users, number crunching, data handling systems, where a small performance improvement must be multiplied by the number of users/uses, by many many hours of operation and by years in service to understand its true impact.... I'm playing with over 300 variables and 100 parameters to simulate dozens of different scenarios can make any server beg for more cycles, and any user beg for a crystal ball.
I don't think that fits into the description the article was talking about.
The point of this article is not targeted to you. I've seen interns as recent as last year complain about the same things mentioned in the article: division is slow, floating point is slow, missed branch prediction is slow, use MMX whenever more than one float is used, etc.
The point I get out of the article is not to bother with what is wasteful at a low level, but be concerned about the high levels. A common one I've seen lately is young programmers trying to pack all their floats into SSE2. Since that computation was not slow to begin with, they wonder why all their 'improvements' didn't speed up the code. Even the fact that they are doing a few hundred unneccessary matrix ops (each taking a few hundred CPU cycles) didn't show up on profiling. Their basic algorithm in a few cases I'm thinking about are either very wasteful, or could have been improved by a few minor adjustments.
The article mentions some basic techniques: choosing a different algorithm, pruning data, caching a few previously computed results, finding commonalities in data to improve the altorithm. Those are timeless techniques, which you probably have already learned since you work on such a big system. Writing your code so that you can find and easily implement high-level changes; that's generally more important than rewriting some specific block of code to run in the fewest CPU cycles.
A very specific example. At the last place I worked, there was one eager asm coder who write template specializations on most of the classes in the STL for intrinsic types in pure asm. His code was high quality, and had very few bugs. He re-wrote memory management so there were almost no calls to the OS for memory. When we used his libraries, it DID result in some speed gains, and it was enough to notice on wall-clock time.
However... Unlike his spending hundreds of hours on this low-return fruit, I could spend a day with a profiler, find one of the slower-running functions or pieces of functionality, figure out what made it slow, and make some small improvements. Usually, a little work on 'low-hanging fruit', stuff that gives a lot of result for a little bit of work, is the best place to look. For repeatedly computed values, I would sometimes cache a few results. Other times, I might see if there is some few system functions that can be made to do the same work. On math-heavy functions, there were times when I'd look for a better solution or 'accurate enough but much faster' solution using calculus. I'd never spend more than two days optimizing a bit of functionality, and I'd get better results than our 'optimize it in asm' guru.
Yes, I would spend a little time thinking about data locality (stuff in the CPU cache vs. ram) but typically that doesn't give me the biggest bang for the buck. But I'm not inherently wasteful, either. I still invert and multiply rather than divide (it's a habit), but I know that we have automatic vectorizers and both high-level and low-level optimizers in our compilers, and an out-of-order core with AT LEAST two floating point, two integer, and one memory interface unit.
And did I mention, I write software with realtime constraints; I'm constantly fighting my co-workers over CPU quotas. I read and refer to the intel and AMD processor documentation, but usually only to see which high-level functionality best lends itself to the hardware. I am tempted to 'go straight to the metal' occasionally, or to count the CPU cycles of everything, but I know that I can get bigger gains elsewhere. That's what the point of the article is, I believe.
The terms sold were "AXA" and "Direct Assurance".... it appears to be the assertion of AXA (the company) that their trademarks were sold to AXA's direct competitors.
This is the thing that bothers me about the whole thing.
The company claims that the use of their words infringes on its trademarks, patents and copyrights, and wants to stop Google from selling them.
[Begin rant]
Google is not selling people's trademarks. The company still owns them.
Google is not selling people's copyrights. They are making fair use of published material, as an indexing service.
Google is not infringing on people's patents by linking to their pages, unless that patant was on linking to someone's pages through a search engine.
Google is not supporting the competitors, other than through paid advertizing.
Google is not making any statements about which product to use.
Google is not making any statements about which company to support.
Google is not transferring any legal rights or ownership of the rights.
Google is not stating that you have a criminal record, or that you are doing something illegal, not making libel claims about you, and not violating your privacy, even though they might offer that information if it is searched for.
What ARE they doing?
Google is saying "You just asked us about one or more word. We have been paid to show you these ads when you typed that word, and these web sites use that word or have a high correlation to that word."
There are billions of web sites. These sites are publically available. Since Google is doing nothing more than indexing publically available content using fair-use excerpts, THESE LAWSUITS SHOULD ALL BE THROWN OUT.
If somebody is to be sued, it is the company paying for the ads, and the people publishing the source sites, making the information available to the public. Not Google.
Slashdot Mirror
A quote from the article is "SFU is not shipped with Windows because SFU currently contains open-source software, such as the GNU C compiler, which cannot be distributed with commercial software. Zions confirmed that Microsoft is working to replace all open-source code in SFU with commercially licensed alternatives."
My money is that about the time Longhorn is released (+/- five years) we'll see another offering, Lindows, which is Microsoft's *nix clone, which also happens to have binary compatability with large swaths, but not complete coverage, of Linix internals. Just enough that the average consumer and most Judges will say "See, they're still competing and not illegally using their Monopoly to do it."
frob
And finally...
NEVER EVER stand in aisle 4 and call your sibling/spouse/parent who is on aisle 5 to ask "Did you want regular or diet?"
frob
As far as the /. question, would you really want to have a boss that gets after you for your present practices? If the boss doesn't understand the importance of a little bit external social interaction, like slashdot, then they probably won't understand other significant things in your culture.
They'll put that the game is rated 'E', since video conferencing doesn't seem to meet any of the ratings elements, and keep the little warning that says "Game Experience May Change During Online Play"
I'd be comfortable with the really-old style number LEDs that were common when cell phones (usually called 'car phones') were first becoming popular. I want to have a phone to talk to people. Seriously, just look at some of the stuff built into These phones. I dont want to have the 'built in' cost of:
- bluetooth
- Infrared
- big screen
- color screen
- touch screen
- 3D graphics processor
- Address book
- web browser
- GPS
- Digital Music
- Java processing
- Camera, for streamed video and audio
- RealOne, MP3, AAC, Midi, Wav, and a dozen other formats supported
- A bunch of games (Tony Hawk's Pro Scater, Tomb Raider, Sonic, MLB Slam!, Super Monkey Ball, more...)
- stereo audio output with headphones
All this and more, for the bargin price of under $200 after a $200 rebate with your 2-year Service Contract. When I searched for free phones at several places, the free (after rebate or with contract) phones all had at a minimum:- big screen
- web browsing
- voice dialing
- Phone & address book
- Several games
- vibrating alert
- GPS (for E911)
- TTY/TTD
I'd still be paying for all these features, regardless of whether I want them, even if they payment is every month through the contract. I don't want them. I just want one of the big grey bricks that I had a decade ago, that you typed in the numbers and it called people. Other people could call you.And that's it.
But the major companies WON'T do that, since that means they can't find convenient ways to charge you $50+ monthly. Well, there's always prepaid wireless...
Look at all the Red Cross / Red Cresent people who jumped in to the danger just because that's what they want to do, and they see the value of helping rebuild a country or helping others as greater than the risks. Other people jump in and will risk everything for money, or just for the thrill. Obviously YOU wouldn't but some people find that kind of life appealing.
The question was asking for an account of real dangers based on people with experience, not for beliefs on the value of life.
Or maybe it's because you are using them for useless things. There are plenty of uses where PDA's are *much* better than laptops.
The company I just left was using them as a 'portable screen' for hardware devices, and that's getting to be a more common thing.
When the engieer or repair person has to go out in the field, laptops are a real pain. When you have to be up a ladder or on a cherry-picker hoist, it becomes a chore to boot up a laptop, balance it on your knee or ladder or box or whatever (hoping it doesn't fall, which DOES happen), configure the hardware, turn off and disconnect the laptop, etc. It is MUCH easier to use a PDA (in our case it was Palm devices) connect the wires on both ends, hit power (instant on!) check off a half-dozen settings, and pull the cords out.
They cost a LOT less to replace, and they take a lot more abuse than most laptops. If you drop them 20 feet onto the grass, they usually survive. Their batteries aren't drained after 3 hours in the field, and they're a lot easier to carry, especially if you have a belt holster. You can usually use your finger rather than a stylus, where you needed two hands to type on the laptop leaving none to hold it. Thats only a few of the benefits.
My guess is that you'll start seeing this type of use increase dramatically -- it saved us quite a bit of both time and money.
One of my favorites was the first peer review of a young intern, only about a hundred lines of source code but probably the biggest thing he'd ever written. I could see something either wrong (buffer overflows, magic numbers, etc.) or missing (no error checking, no comments, no assertions, ...) on nearly everything he did.
I know it's cruel, but to watch their over-exuberant pimply faces go from "Joy! I'm doing Important Things!" fall to "All the things I've been gloating over mean absolutely nothing" is oddly fun.
Of course, I follow the rule "reprove in private, praise in public" (with another team member present), I try to do it early, before they have much emotional investment in their work (usually after a few days of work), and always do ego rebuilding and team building afterword. Usually something like "But it's a good start, considering you just graduated. Let's sit down together and fix a few of those holes."
Am I alone in this perverse pleasure?
I think it's a pretty common thing. Just look through the Fortune's company list: Abbey, Boots, Continental, ... Kingfisher, Lend Lease, Metro ... United, Visa, Wendy's ... All of them international corporations, and a few, like Visa, attempt to rewrite the dictionary to remove the common word.
So why shouldn't the scientoligists or whatever group the person rewriting 'bright' is, join in the fray? ;-)
Personally I think we should all take the words "slash" and "dot", and try to... oh, wait.
My P2-300 is next to a P4-2Ghz. While there is an obvious speed difference, it isn't all that big. I use them both for development, and don't bother replacing it because it is 'good enough'. Usually I click the power buttons at the same time, log in to the P4 box, while I wait it to load all the startup garbage, I log in on the P2-300. Then I start my few apps on the P4, then on the P2. When I do remote debugging on the P2 (nothing really CPU intensive, mostly RAM- and disk-intensive database work) it is just as functional speed-wise as the P4. Sure, it helps that I have an ATA-66 disk and 192MB of 133 ram inside, but that shouldn't be a big thing in that age of system. The speed difference is very noticable when I do CPU intensive things, like video conversion and games, but I have only have done video conversion a few times and only play a few games like networked Warcraft 2, which also plays fine on it. Sure, it takes an extra 20 seconds when I compile or an extra few minutes for a full rebuild, but that's okay for me.
When I do remote debugging in the win98 486 box (I wouldn't dream of a full DevStudio install there), the big problem is coping with the relatively small amount of memory (64M) and finding ways to reduce the thrashing to swap space. The good thing about that is that I know my programs run well on fast machines and reasonable on slow/old ones.
Just to put it in comparison, the P2 is a duel boot into linux, since several linux-based games are fun. When I decided it might be nice to try the 2.6 kernel in the debian 'sarge' installation a few weeks ago (mid-March), I only used it for two days before re-formatting and going back to a debain 'woody' installation without most of the server components. I use KDE, and it feels slower than what XP feels like. I don't care to use another window manager because of the time to install & configure, I'll just take what comes out of the box, thank you.
For fun I just booted it into linux, started OpenOffice, Mozilla, and Anjuta. The system feels slow, and the HD sounds like it's going to blow up when they were starting up. Surfing to a few web pages (userfriendly, /., sourceforge, openoffice.org) makes me worry about the disk, it's just clattering away. Loading a design spec word document from the common windows partition just took about 45 seconds, and it stalls when I scroll through it. Eeek. It is awful compared to what I do on XP. Oh well, back to Windows. Well, a few games of Frozen Bubbles first. :-)
You are using a book, which has sentances which have much more structure than that data was. Yours in the form {1 CAP}{~50 lower case}.
In that particular data set, It was around 0-10 caps with 0-20 non-caps, for each entry. They were using an inconsistant entry rule, some all upper, some all lower, some mixed, mostly abbrviated words to fit on receipts.
It makes a difference. Even in your less relavent example, caps made a 3% difference.
C'mon, it's my own post, and I caught at least three typos in it. If your going to correct my spelling, please at least take the time to find ALL of them; Finding only some of them makes you look dumb. One word has two different spellings, only a few words away from each other.
frob
The first one will save you lots of money in mailing. If your company is in to saving money, then your mailing lists already do that. Compressability of data is *very* important for large-scale systems, and can have significant improvements with an ALL CAPS DATABASE. The last two just help reduce system load and operator error; both are good things to reduce.
frob
You say it doesn't help if you have a syntax highlighting editor; you can't count on having a syntax highlighted editor all the time.
While I think there are benefits to it, many interns that I've worked with seem paralyzed if they don't have syntax highlighting. A few are paralized if they don't have auto-complete. If the code REQUIRES syntax highlighting to be understandable, clearly you need to adopt a different style.
If you don't have SQL syntax highlighting, the more complex the SQL statement is the more it benefits from the all-cap keywords. That in itself is enough to justify (for me) it's use. Since I compose my SQL statements in my C++ interface code, I don't have SQL syntax highlighting; the C++ editor highlights them all as strings. Using caps really helps readability, especially when the SQL statements are complex and require several lines to compose.
It sounds like a form of government or policing to me, depending on what the powerful groups were/are doing.
The question is, which group do they fall under: The Terrible, The Vengeful, or The Just? I would guess either the first or last, and since the companies probably aren't willing to take too much risk, they'll at least tend to be more akin to 'The Just' rather than 'The Terrible'.
The infected university supercomputer is really not that different from the infected campus, which admins have been dealing with for years.
Finally, many university supercomputers are already on the Internet. I've worked with 7 supercomputers at three schools, all had their head (node zero) accessable through the Internet. That was where we logged in, compiled and tested our jobs, and queued them up for running on the rest of the network. Two of the clusters had each node internet accessable, and (interestingly for this topic) one of the clusters, a bunch of multi-processor P3s, was running Windows NT on each box.
I would suggest that you look up either the actual RFCs or an authoritative reference like 'TCP/IP Illustrated.' The grandparent post was wrong, and your correction is also wrong. Let's look at the facts.
First, the application will probably encode the real data within their own packet. FTP can use compression, and can (but doesn't often) use higher encodings and structures. It's data is then wrapped into a TCP packet (assuming TCP).
A TCP header is between 20-60 bytes, depending on what flags are set. (Normally it's 20 bytes). Also, there is a minimum data size which is replaced by padding if your packet is really small.
The TCP packet is wrapped inside an IP packet, which is also a 20-60 byte header (normally 20).
The now 46-1500 byte packet is then wrapped and unwrapped as it travels across each network. Presumably within your own network, this is a series of ethernet packets with a 14-byte header and 4 byte trailer (18 byte total).
Once you get into ethernet collisions, TCP retransmittals, packet fragmenting, header compression, ICMP messages, etc., the number of bits/bytes actually on the wire doesn't fit into some formula for finding the actual bytes of application-level data being recieved. On a noisy line, it is entirely possible to transmit a gigabyte of packets thorugh TCP/IP and still not receive even a signle byte of application-usable data.
A blanket statement of assuming 8 bits of application data for every 13 or 14 bits of wire data is pretty bad. That's saying 5/13 or 5/14 (about 37%) of your data is headers, control packets, retransmission, etc. On a clean transmission, it could be 40/1500 (about 2%) or potentially even lower if there is header compression or bigger MTU. Or if you are transmitting one byte of data at a time, it could be worse than 45/46 (98%) spent in overhead. Or you could have the amount you stated, or you could have 100% of it as controlling and lost, and get no useable data out.
Even on an exclusive and clean line, you have a range of below 2% to over 98% overhead, which is a huge spread. A little noise on the line can also have have a huge impact. Even though the 'speed test' programs transmit a large file, something that presumably uses large packets, the amount of usable data coming over the line can and does vary based on many factors. Simply assuming 10-15 transmitted bytes for each useful byte MIGHT be useful for you as a very rough guess, but the only way to tell for certain would be to monitor on your own system the usable data passing through each datagram, and compare it to what the hardware says was actually transmitted.
These speed tests are only useful at diagnosing that you are getting roughly the speed that you paid for on a fairly noise-free connection, and for ego-trips. Nothing more. And certainly not for evaluating in a blanket statement that there are x bits of overhead for y bits of data.
frob
customer:What is this solid block of lard doing in the dairy section?
clerk:That's not lard, it's non-fat milk fat.
customer:What is the solid block of not-really milk fat doing in the dairy section?
clerk:It's the new Atkin's Brand 100% Milk for Lactose Intolerant. It's popular among teenage girls and dieters.
- Product A -- Claims to be 73% good.
- Product B -- Claims to be 96% good.
- Product C -- Claims to be 99.999% good.
- Product D -- Claims to be 100% good.
Being skeptical, you would probably pick product A has having truthful ads. Product B, you might think, has really good real-world performance. Product C is just marketing hype, and product D is impossible in the real world.But if you see a big brand name (Microsoft, Cisco, Intel, etc.) on product C, you might say "Well, it isn't 100%, and they are a good company. Maybe it's the truth. Of course, claiming to be Product C happens, and that's where the trap is.
It might be that you are looking at Microsoft statement claiming "5 nines" of 99.999% uptime (that's down for 5 minutes each year). Or Sun claiming the same 99.999%. Or Cingular Wireless claiming 99.999% reliable networks, excluding several days of downtime that they must not factor into their percentage. Maybe it's that 99.999% pure copper speaker cable you were looking for. (For the chemists, here's a site where you can buy over a dozen other '99.999% pure metal' wires.) Lots of people get caught into that.
In some cases it really is justified. If I were a chemist, maybe having iridium wire that is only 99.9% pure might cause problems, and those extra 9's might be significant. But that usually isn't the case for most marketing.
But I don't think it's just a PHB issue, it's a problem of 'I really want the best, and I only want to spend 5 minutes to find out which one that is'.
frob
I don't think that fits into the description the article was talking about.
The point of this article is not targeted to you. I've seen interns as recent as last year complain about the same things mentioned in the article: division is slow, floating point is slow, missed branch prediction is slow, use MMX whenever more than one float is used, etc.
The point I get out of the article is not to bother with what is wasteful at a low level, but be concerned about the high levels. A common one I've seen lately is young programmers trying to pack all their floats into SSE2. Since that computation was not slow to begin with, they wonder why all their 'improvements' didn't speed up the code. Even the fact that they are doing a few hundred unneccessary matrix ops (each taking a few hundred CPU cycles) didn't show up on profiling. Their basic algorithm in a few cases I'm thinking about are either very wasteful, or could have been improved by a few minor adjustments.
The article mentions some basic techniques: choosing a different algorithm, pruning data, caching a few previously computed results, finding commonalities in data to improve the altorithm. Those are timeless techniques, which you probably have already learned since you work on such a big system. Writing your code so that you can find and easily implement high-level changes; that's generally more important than rewriting some specific block of code to run in the fewest CPU cycles.
A very specific example. At the last place I worked, there was one eager asm coder who write template specializations on most of the classes in the STL for intrinsic types in pure asm. His code was high quality, and had very few bugs. He re-wrote memory management so there were almost no calls to the OS for memory. When we used his libraries, it DID result in some speed gains, and it was enough to notice on wall-clock time.
However... Unlike his spending hundreds of hours on this low-return fruit, I could spend a day with a profiler, find one of the slower-running functions or pieces of functionality, figure out what made it slow, and make some small improvements. Usually, a little work on 'low-hanging fruit', stuff that gives a lot of result for a little bit of work, is the best place to look. For repeatedly computed values, I would sometimes cache a few results. Other times, I might see if there is some few system functions that can be made to do the same work. On math-heavy functions, there were times when I'd look for a better solution or 'accurate enough but much faster' solution using calculus. I'd never spend more than two days optimizing a bit of functionality, and I'd get better results than our 'optimize it in asm' guru.
Yes, I would spend a little time thinking about data locality (stuff in the CPU cache vs. ram) but typically that doesn't give me the biggest bang for the buck. But I'm not inherently wasteful, either. I still invert and multiply rather than divide (it's a habit), but I know that we have automatic vectorizers and both high-level and low-level optimizers in our compilers, and an out-of-order core with AT LEAST two floating point, two integer, and one memory interface unit.
And did I mention, I write software with realtime constraints; I'm constantly fighting my co-workers over CPU quotas. I read and refer to the intel and AMD processor documentation, but usually only to see which high-level functionality best lends itself to the hardware. I am tempted to 'go straight to the metal' occasionally, or to count the CPU cycles of everything, but I know that I can get bigger gains elsewhere. That's what the point of the article is, I believe.
The company claims that the use of their words infringes on its trademarks, patents and copyrights, and wants to stop Google from selling them.
[Begin rant]
Google is not selling people's trademarks. The company still owns them.
Google is not selling people's copyrights. They are making fair use of published material, as an indexing service.
Google is not infringing on people's patents by linking to their pages, unless that patant was on linking to someone's pages through a search engine.
Google is not supporting the competitors, other than through paid advertizing.
Google is not making any statements about which product to use.
Google is not making any statements about which company to support.
Google is not transferring any legal rights or ownership of the rights.
Google is not stating that you have a criminal record, or that you are doing something illegal, not making libel claims about you, and not violating your privacy, even though they might offer that information if it is searched for.
What ARE they doing?
Google is saying "You just asked us about one or more word. We have been paid to show you these ads when you typed that word, and these web sites use that word or have a high correlation to that word."
I'm sick of lawsuits against them for indexing public sites.
There are billions of web sites. These sites are publically available. Since Google is doing nothing more than indexing publically available content using fair-use excerpts, THESE LAWSUITS SHOULD ALL BE THROWN OUT.
If somebody is to be sued, it is the company paying for the ads, and the people publishing the source sites, making the information available to the public. Not Google.
[End Rant]