I pretty much agree with what you're saying, although I disagree with your characterization that the geometry & materials of a road bike constitute the "proper" bike. Sure, for extremely long distance riding (30 miles a stretch or more - and I know you're laughing at me for calling that extreme, but it is on the extreme end of how far people ride bikes) where absolute speed is most important, a road bike is probably the right geometry. But for just about all other uses for a bike, it's completely wrong.
It could be argued that for long-distance travel where a certain amount of cargo needed to be carried, recumbent bikes would probably be better suited. And for urban biking, or biking on less than ideal road conditions, which would be the case for any "practical" bike (one that serve as primary transportation for someone), the standard beach cruiser, or some evolution of it is probably the way to go.
The road bike geometry is hard on your arms & back (until you're "in shape", or have built up enough scar tissue to tolerate it), too fragile to handle city & country bumps, and too dangerous in traffic, especially if you're using clip-in pedals. Yeah, it's the most efficient, but only for atheletes.
Here's my spin on bloat: If the size or maintenance/learning requirements of an application makes its use inconvenient, relative to the features it provides, then it's bloated.
This definition obviously changes over time, context and individual user. 10 years ago, when people were using 14.4 modems and 8 meg of ram, a 4 megabyte browser that had be downloaded from the net could rightly be called bloated - the features it was providing probably didn't justify the cost in aquisition or utilization. Today it's no big deal.
If all you ever do with text files is modify a few lines here & there, emacs is bloated. I'd say by most people's standards, RealPlayer is probably one of the most bloated applications today, given it's memory demands, registry mangling and general intrusiveness when all it provides is the ability to play streaming media (all its other features are redundant with applications people are already using).
If something takes longer to start up than the task you need to accomplish, it's bloated.
That said, I pretty much agree with you. By today's standards, almost no software out there is bloated, and given the garbage that John Q. User downloads & installs in his system tray, the concept doesn't even exist to the general public.
As appealing as it may be to glue two pop-cultural phenomenons together, the connections to fictionalized mafia and EQ seem pretty tenuous.
For one thing, there isn't the zero-sum game of the Mafia - the power brokerage involved in EQ doesn't seem to benefit from betrayal, or "keeping your enemies closer" aspect that we see in that thing of theirs. There isn't the "money flows up, shit flows down" ethic, and you don't have to worry about entanglements with a more powerful outside authority (FBI).
Most of the examples given in the article examining the social networking could just as easily be seen as an excuse to have an adventure ("Someone's dead! Let's go rescue him."). You get to play the hero in a very specific mini-myth.
The larger & more formalized groupings in the game resemble fraternities a lot more than the mafia - a bunch of people who glom together who share a common outlook on life & a desire to party together. Piss off the alpha members of said community and you'll be shunned, not whacked. Heck, with all those "virtual weddings" you hear about, can "virtual date rape" be that far off?
The other posts answer the question, although I'd disagree that exposure to BASIC completely ruins a programmer (most of the great programmers now in their thirties go their start as kids writing BASIC programs on their Apples & C64s - the important thing is that they're ultimately exposed to the right ideas).
VB gets such a bad rap because it is simply the most powerful & simple programming language that can be learned with no prior programming experience, and more importantly, no exposure to other programmers. Because of IDE, a great help system (particularly in VB5 - VB6 & beyond are too messy with help from the other VS components), and tons of sample code available to steal, it allows people who know nothing about programming to just cut & paste & start modifying. Good programming practices never emerge when the developer flies solo, so it's no surprise that so many bad programmers out there are VB people - the inaccessiblity of other languages forces learners to seek the advice of more experienced developers.
That said, VB is great tool in the right hands. We do most of our development in C++ & Java, but VB has a place in our products as a front-end builder. Its help system & IDE are the best in the business. It's fast enough for 90% of all software written today. For someone who already has learned good practices from other languages, it allows you to produce usable, maintainable apps far faster & with less likelihood for hidden gotchas. I haven't tried the.NET version, but it sounds like they eliminated most of the annoyances that exist in the current versions -- mostly due to soft typing - stuff like Variants, implicit conversions, bad choice of operator overloading, and forcing you to use ActiveX (and all it's headaches with the Registry) for all component development.
But isn't it only an emotional component when you're playing against those with emotions? In a 2 player human-computer Texas Holdem game, if the computer never folded, the odds of winning any hand would end up being even. The only control the human player would have would be varying his stake in the game by choosing to raise or not, or to fold. So the trick is to have a betting algorithm for the computer that doesn't reveal the contents of its hand, yet still allows it to bet high on good hands & fold or drop on bad ones.
Random betting would hide the computer's hand but not allow it to gain financial advantage. But consider a random betting range that was based on the mathematically "ideal" bet. On any given hand, the human wouldn't be able to tell whether the hand was good or not (since the min & max range of the random function would always go from 0 to max bet, only the shape of the distribution curve would change). The computer could also keep track of how much it had won/lost on past hands and adjust the curve dynamically to try to account for a losing streak, but this may not be necessary.
The result would be that in the long run, the computer would gain the advantages of statistically perfect betting, without the disadvange of tells. Is there a problem with this scheme I don't see?
Like I said, there's nothing profoundly wrong the behavior, you're just being an exploiter. If you see the world (both economic & social) as a race to see who can exploit whom the best, than by all means go for it.
Probably the best analogy is comparing this behavior that of people who play online games with the intention of uncovering exploits. Their exploiting the system put in place by the company providing the service, and as you said, if everyone was employing the exploit, you'd be a fool not to.
As far as whether the salesman had your best interest in mind, I'd say it isn't an either/or proposition. It's not in the salesmans best interest to screw you over, since they'll lose you as a future customer & as an advocate for their services. I was very pleased with my purchase of a Canon G2, and because of that, I have convinced four others to buy one. Our company sells process control software - you better believe that our salesmen have our customer's best interests in mind when we try to get people to buy it - an unhappy customer is poison.
When you decide to exploit the investment a business makes in customer service but take your business elsewhere, you're voting with your wallet for stores with poor customer service. If that's what you want, fine, but don't turn around & cry about how the places you spend your money are staffed with minimum wage know-nothings.
Social contracts exist in commerce just like everywhere else; commerce is still a part of society. And like social contracts elsewhere, nothing compels you to follow them. Whether you follow them or not depends on your world view. I'm sure the Iraqis in Baghdad would also be fools not to loot when everyone else is doing so.
Finally, I'd like to point out that I'm not above employing these techniques - I recently saved $75 on a pair of sunglasses by getting them through my sister in law (an optometrist) after finding the one I liked at a local Sunglass Hut. So am I a weasel? In this case, yes. I don't try to rationalize that I'm somehow fighting the good fight or being an informed consumer. I'm exploiting the system.
You get what you pay for, unless you're stealing, in which case you're getting more. His FIL is taking advantage of the investment the more expensive store has placed in salespeople and floor displays. His FIL wouldn't be paying "more than he should" at the fancy store, he'd be paying for the expertise and equipment, as opposed to just the equipment. If everyone did this, the stores that offered good service would go out of business, leaving just the volume retailers.
There's nothing illegal, or even horribly unethical. You're just taking advantage of the system, maybe akin to going to a party & eating all the snacks before someone else has a chance to - after all, there's no law that says you have to leave anything for anyone else.
Although the idea may or may not be ultimately feasible for wide-scale use due to the presence of transformers, why hasn't the home/small office networking domain used household power lines for network cable? Given burgeoning popularity of wireless networks, couldn't a household powerline network acheive similar goals?
My problem with Amazon is that the organization is geared towards sales, rather than understanding. The is demonstrated by the fact that a given book shows up multiple times based on whether it was hardcover, softcover, printing 1, 2, 3, whatever and so on. Click on a J.R.R. Tolkien link and you get 237 results. Tolkien didn't write 237 books - I should see just a list of the books (preferably broken down like imdb does, based on his role, author, editor, contributor, etc.). When I click on a book, I can see all the different forms it came in, and what printings, etc. Amazon's format is just a huge pile you have to sift through.
Also, no crossreferences (granted, not as necessary with books as compared to movies that take dozens of people to produce), no sorting by original publication date, etc. It needs an organization like allmusic or imdb.
Finally, a more formalized way for users to enter data besides comments, such as imdb's trivia, goofs, a better (normalized?) rating scheme (Amazon's star-based scheme rarely has books that drop below 3.5 stars). Someone mentioned above that Amazon owns imdb, but it was originally a distributed effort like this one.
Something like CPU isn't really an acronym but in fact a "spelled" abbreviation (as opposed to a regular abbreviations like dept. or Dr. that is still pronounced as the original word). Prior to the 70's (there's that appostrophe again), spelled abbreviations were almost always delimited by periods (T.V.A., F.D.R., F.B.I.), and pronounced acronyms (NATO, radar) were not.
The idea, as I have always heard & understood it, is that merely putting an "s" on the end of an spelled abbreviation that is pronounced as a series of letters is confusing (do you spell or say the "s" - SEE PEE YOU ESS?), as you may think the "s" is part of the abbreviation. Traditionally, lettered abbreviations were delimited with periods, like M.D.; making them plural by adding a lowercase "s" didn't look right at all (M.D.s).
Soon after that you started seeing the creation of the InterCaps plural form of the spelled abbreviation (CDs). Now that InterCaps and spelled, non-periodified acronyms are relatively common, the traditional style guides are finally officially sanctioning the CPUs form of the plural, but the form CPU's is still correct, but falling out of favor. Back in the 60's it probably would be more correctly spelled C.P.U.'s - hmm.
Most style guides recognize that the ambiguity exists, and that grammar nazis will get their undies in a bundle, so their recommendation is to avoid sentence structures that force you to make a spelled abbreviation plural.
This particular query was run on 8.1.6 Enterprise (Solaris & NT), but I've been working with Oracle since 7.3.4 and we're currently doing 9i. It functions similarly on all of them. It could be that the sub-enterprise releases of Oracle behave differently, but I'm not sure - we were using Personal Oracle on our laptops for our sales staff, and we didn't see any drop in performance once the I/O & CPU were taken into account.
I'm surpised that Oracle would be advocating bitmaps for everything. We use them a lot, and they're great, but they're not useful for everything, and have particular problem when used on high transaction tables. Plus, some columns (particularly DATE columns) don't lend themselves to bitmaps.
Oracle will use the index Table(A,B) to locate all A where A=Y, then scan through those rows for B=X. It will not use the full index like you've said. If you want to use the full index, you either need to specify A=Y before B=X
This doesn't match with my experience. Order of the conditions in a WHERE clause doesn't effect the optimizer. Specifically, in this case, I create a sample million-row table:
create table foo (col1 integer, col2 integer); insert into foo values (1, 1); insert into foo select col1, col2 + 1 from foo; insert into foo select col1, col2 + 2 from foo; insert into foo select col1, col2 + 4 from foo; [...] insert into foo select col1, col2 + 512 from foo; insert into foo select col1 + 1, col2 from foo; insert into foo select col1 + 2, col2 from foo; insert into foo select col1 + 4, col2 from foo; [...] insert into foo select col1 + 512, col2 from foo; create unique index foo_i on foo (col1, col2); analyze table foo compute statistics;
(I had to fiddle with the above because/.'s lameness filters were being set off)
I then run two queries, reversing the order of the constraints in the where clause:
SQL> set autotrace on
SQL> select * from foo where col1 = 476 and col2 = 717;
COL1 COL2
476 717
Execution Plan
0 SELECT STATEMENT Optimizer=CHOOSE (Cost=2 Card=1 Bytes=6)
1 0 INDEX (UNIQUE SCAN) OF 'FOO_I' (UNIQUE) (Cost=2 Card=1 Bytes=6)
Statistics
0 recursive calls
0 db block gets
3 consistent gets
0 physical reads
0 redo size
262 bytes sent via SQL*Net to client
314 bytes received via SQL*Net from client
4 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed
SQL> select * from foo where col2 = 717 and col1 = 476;
COL1 COL2
476 717
Execution Plan
0 SELECT STATEMENT Optimizer=CHOOSE (Cost=2 Card=1 Bytes=6)
1 0 INDEX (UNIQUE SCAN) OF 'FOO_I' (UNIQUE) (Cost=2 Card=1 Bytes=6)
Statistics
0 recursive calls
0 db block gets
3 consistent gets
0 physical reads
0 redo size
262 bytes sent via SQL*Net to client
314 bytes received via SQL*Net from client
4 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed
You can see that the performance of the query is the same in both cases. Am I missing something?
Not exactly. The developer (Delphine) later made a somewhat similar game called Flashback, but it didn't really have the same charm. It didn't have the same characters or story, but it had a similar format (the screen based side non-scroller genre, I guess). It wasn't a bad game, but it was more like Prince of Persia than OOTW.
A gold NWC cartridge recently sold for $6,500! Will their value increase in the future? It seems a safe bet.
I doubt it. For technology items like this, the value of the collectable is function of the earning potential of buyer and the nostalgia value of the item. For example, classic cars slowly go up in value as the demographic that remembers them from their teenage years reaches the age of massive disposable income, then drop in value as the same group slowly dies off.
With geek items like this, the half-life is even shorter. Magic The Gathering cards are already past their prime in terms of collectable value; once the people who played the NES in their youth are past the age of buying this stuff, watch the prices plummet.
Robert Park of the American Physical Society is one of the more outspoken physicists against the ISS. His argument is that it is ostensibly for scientific research, but is really a pork-barrel project and offers little benefit to researchers in terms of cost/scientific output. If its purpose were as a stepping stone to further space exploration, it would be a different matter, but as a research facility it is expensive and ineffectual (even once it comes fully on-line).
Raid 0 is exactly as reliable as your least reliable hard drive
Actually, it's less reliable than your least reliable drive (assuming you're not using array of 1 disk). The chance of failure of the array is 1 - (product of all disks' chance of not failing). An array of 2 identical disks will fail twice as often as an individual disk.
Let's assume that a, b and c are on three separate tables. One way this could be true would be for there to be a bug or weakness in the interpreter. If there were only an index on b, then a=b and b=c might resolve efficiently, but if the optimizer was given the three choices and chose the less optimal pair (a=b and a=c, for example), it might result in more work than a=b,c=b.
A non-flaw based explanation: Assuming that there are indexes on all tables, there's still the task of chosing which ones. The standard way that an RDBMS might resolve this query is to iterate over each table in succession, but there are six permutations of such a strategy (a->b->c, a->c->b, b->a->c, b->c->a, c->a->b, c->b->a). Only one may be chosen, but how it decides to do so will depend on how the system's optimizer is designed.
Some may strictly use "semantic" analysis - just look at the query & pick an option. Most will use statistics that are gathered by the system pertaining to the distribution of data & keys in the tables and attempt to estimate how much work each strategy will take based on row count & key cardinality. Depending on how accurately & frequently the system maintains these statistics, you can get different results, so even the same RDBMS might give multiple results depending on whether its statistics are up to date.
Optimizers generally don't infer equivalence relationships in order to allow the query writer to nudge the engine in one direction or another. With a=b,b=c, there are only two routes the most systems would take (a->b->c, c->b->a). By limiting the choices, you can coax the system to avoid a plan that may look attractive based on statistics but proves to be less practical for whatever reason.
Even if all the statistics are equivalent, two different systems may still produce different results. One plan for data retrieval might result in operations overall, but the may result in a long delay before you receive the first results (low throughput cost, high first row latency); conversely, another plan may do the opposite. Depending on your use of the data, you may prefer one over the other - for example, doing some summary financial report on a company's transactions would benefit from the low-throughput cost, since you need to look at all the data, but a Google type search where the user will scan the data in order and stop looking long before he or she reaches the end of the result set will benefit from a low first-row latency). One system may default to the first behavior, others to the second.
There are plenty of other reasons, but the main reason is that since optimizers have many ways of carrying out a request, and different performance characteristics may be desired from the same statement, they don't do 'simplification' in order to give the developer some control over the process (and also because simplification is not always simple to carry out & may unnecessarily consume CPU cycles).
> > Actually, it was made from "The Castle of Cagliostro" > Right. That's what I meant. What was that game called?
The game was Cliff Hanger (The guy in the game was referred to as "Cliff"). The great thing about it compared to Dragon's Lair was that you had a "hand action" button and a "foot action" button, and that whenever you died, some text appeared at the bottom that said "You should have gone left" or "You should have pressed hand". It was much longer than DL and you didn't have to spend so much money on it to finish.
I only ever saw it at one arcade, which incidentally had another laserdisk game I never saw anywhere else Star Rider- a racing game with sprite racecars superimposed over laserdisk background. It was ingenious - the laserdisk simply played at a rate according to the speed of your racer, and the tracks you raced on were surreal and impressive. They had names like Crystalopolis & Cubitania.
Contract authentication. In a manner similar to the old "wax & signet ring", any document could be projected through the crystal and it's speckle pattern saved with the document. Although you couldn't verify that someone was who they said they were, you would be able to verify after the fact that this document did or did not originate from that person.
This could be extremely useful for e-commerce. When you're paying for your purchase, the system could send you an image of their choice (the invoice for example) and you could run it through your crystal & send it back. If there were a billing discrepancy later on, the company could produce the original image and you could demonstrate that it did or didn't match your card.
So, it's main use would be in keeping consumers from ripping off credit card companies with fake claims of "it wasn't me". Conversely, you could use it do prove that you didn't make a particular purchase.
There is no better writer when it comes to explaining science. Although it covers a lot more than just physics, the broad coverage is useful when you want to follow up on information that leads you away from physics proper.
The only disadvantage is that it is no longer in print, but thanks to the wonder of the internet, that's no longer a problem.
Bottom line: expanding that 2-digit packed field, from one byte to two bytes, to fix the future problem, would have cost several hundreds of thousands of 1970's dollars just for the hardware cost of the storage, never mind the professional labor cost to program it and do the conversion. It was decided for what, in hindsight, were good reasons: it will be much cheaper to fix it in the future, if in fact it is still a problem when that time comes.
I've heard this argument before, but I'm unclear about something: Why the decision to represent years in a BCD format (1 digit per nibble)? Why not represent the year as an 8-bit integer, offset from 1900? This would have allowed the byte to hold dates from 1900 - 2155? Take it one step farther, and have the other byte hold the Julian date portion (1-365), and you could have fit the whole date into 2 bytes (instead of the 3 bytes required for the 6-digit BCD form), and had simpler date arithmetic to boot? Surely the storage space cost savings would have justified the 33% decrease in size?
Slashdot Mirror
I pretty much agree with what you're saying, although I disagree with your characterization that the geometry & materials of a road bike constitute the "proper" bike. Sure, for extremely long distance riding (30 miles a stretch or more - and I know you're laughing at me for calling that extreme, but it is on the extreme end of how far people ride bikes) where absolute speed is most important, a road bike is probably the right geometry. But for just about all other uses for a bike, it's completely wrong.
It could be argued that for long-distance travel where a certain amount of cargo needed to be carried, recumbent bikes would probably be better suited. And for urban biking, or biking on less than ideal road conditions, which would be the case for any "practical" bike (one that serve as primary transportation for someone), the standard beach cruiser, or some evolution of it is probably the way to go.
The road bike geometry is hard on your arms & back (until you're "in shape", or have built up enough scar tissue to tolerate it), too fragile to handle city & country bumps, and too dangerous in traffic, especially if you're using clip-in pedals. Yeah, it's the most efficient, but only for atheletes.
-BbT
Fixies still need brakes. It's probably not a Fixie, but a bike with hub brakes.
I actually saw this in stores.
-BbT
If lack of time travel is proof that we're in a simulation, then the inviolability of the speed of light is proof as well:
It's obviously hard-coded.
-BbT
Here's my spin on bloat: If the size or maintenance/learning requirements of an application makes its use inconvenient, relative to the features it provides, then it's bloated.
This definition obviously changes over time, context and individual user. 10 years ago, when people were using 14.4 modems and 8 meg of ram, a 4 megabyte browser that had be downloaded from the net could rightly be called bloated - the features it was providing probably didn't justify the cost in aquisition or utilization. Today it's no big deal.
If all you ever do with text files is modify a few lines here & there, emacs is bloated. I'd say by most people's standards, RealPlayer is probably one of the most bloated applications today, given it's memory demands, registry mangling and general intrusiveness when all it provides is the ability to play streaming media (all its other features are redundant with applications people are already using).
If something takes longer to start up than the task you need to accomplish, it's bloated.
That said, I pretty much agree with you. By today's standards, almost no software out there is bloated, and given the garbage that John Q. User downloads & installs in his system tray, the concept doesn't even exist to the general public.
-BbT
As appealing as it may be to glue two pop-cultural phenomenons together, the connections to fictionalized mafia and EQ seem pretty tenuous.
For one thing, there isn't the zero-sum game of the Mafia - the power brokerage involved in EQ doesn't seem to benefit from betrayal, or "keeping your enemies closer" aspect that we see in that thing of theirs. There isn't the "money flows up, shit flows down" ethic, and you don't have to worry about entanglements with a more powerful outside authority (FBI).
Most of the examples given in the article examining the social networking could just as easily be seen as an excuse to have an adventure ("Someone's dead! Let's go rescue him."). You get to play the hero in a very specific mini-myth.
The larger & more formalized groupings in the game resemble fraternities a lot more than the mafia - a bunch of people who glom together who share a common outlook on life & a desire to party together. Piss off the alpha members of said community and you'll be shunned, not whacked. Heck, with all those "virtual weddings" you hear about, can "virtual date rape" be that far off?
-BbT
The other posts answer the question, although I'd disagree that exposure to BASIC completely ruins a programmer (most of the great programmers now in their thirties go their start as kids writing BASIC programs on their Apples & C64s - the important thing is that they're ultimately exposed to the right ideas).
.NET version, but it sounds like they eliminated most of the annoyances that exist in the current versions -- mostly due to soft typing - stuff like Variants, implicit conversions, bad choice of operator overloading, and forcing you to use ActiveX (and all it's headaches with the Registry) for all component development.
VB gets such a bad rap because it is simply the most powerful & simple programming language that can be learned with no prior programming experience, and more importantly, no exposure to other programmers. Because of IDE, a great help system (particularly in VB5 - VB6 & beyond are too messy with help from the other VS components), and tons of sample code available to steal, it allows people who know nothing about programming to just cut & paste & start modifying. Good programming practices never emerge when the developer flies solo, so it's no surprise that so many bad programmers out there are VB people - the inaccessiblity of other languages forces learners to seek the advice of more experienced developers.
That said, VB is great tool in the right hands. We do most of our development in C++ & Java, but VB has a place in our products as a front-end builder. Its help system & IDE are the best in the business. It's fast enough for 90% of all software written today. For someone who already has learned good practices from other languages, it allows you to produce usable, maintainable apps far faster & with less likelihood for hidden gotchas. I haven't tried the
-BbT
Browse at -1 and you'll see who he's talking to.
-BbT
But isn't it only an emotional component when you're playing against those with emotions? In a 2 player human-computer Texas Holdem game, if the computer never folded, the odds of winning any hand would end up being even. The only control the human player would have would be varying his stake in the game by choosing to raise or not, or to fold. So the trick is to have a betting algorithm for the computer that doesn't reveal the contents of its hand, yet still allows it to bet high on good hands & fold or drop on bad ones.
Random betting would hide the computer's hand but not allow it to gain financial advantage. But consider a random betting range that was based on the mathematically "ideal" bet. On any given hand, the human wouldn't be able to tell whether the hand was good or not (since the min & max range of the random function would always go from 0 to max bet, only the shape of the distribution curve would change). The computer could also keep track of how much it had won/lost on past hands and adjust the curve dynamically to try to account for a losing streak, but this may not be necessary.
The result would be that in the long run, the computer would gain the advantages of statistically perfect betting, without the disadvange of tells. Is there a problem with this scheme I don't see?
-BbT
Like I said, there's nothing profoundly wrong the behavior, you're just being an exploiter. If you see the world (both economic & social) as a race to see who can exploit whom the best, than by all means go for it.
Probably the best analogy is comparing this behavior that of people who play online games with the intention of uncovering exploits. Their exploiting the system put in place by the company providing the service, and as you said, if everyone was employing the exploit, you'd be a fool not to.
As far as whether the salesman had your best interest in mind, I'd say it isn't an either/or proposition. It's not in the salesmans best interest to screw you over, since they'll lose you as a future customer & as an advocate for their services. I was very pleased with my purchase of a Canon G2, and because of that, I have convinced four others to buy one. Our company sells process control software - you better believe that our salesmen have our customer's best interests in mind when we try to get people to buy it - an unhappy customer is poison.
When you decide to exploit the investment a business makes in customer service but take your business elsewhere, you're voting with your wallet for stores with poor customer service. If that's what you want, fine, but don't turn around & cry about how the places you spend your money are staffed with minimum wage know-nothings.
Social contracts exist in commerce just like everywhere else; commerce is still a part of society. And like social contracts elsewhere, nothing compels you to follow them. Whether you follow them or not depends on your world view. I'm sure the Iraqis in Baghdad would also be fools not to loot when everyone else is doing so.
Finally, I'd like to point out that I'm not above employing these techniques - I recently saved $75 on a pair of sunglasses by getting them through my sister in law (an optometrist) after finding the one I liked at a local Sunglass Hut. So am I a weasel? In this case, yes. I don't try to rationalize that I'm somehow fighting the good fight or being an informed consumer. I'm exploiting the system.
-BbT
You get what you pay for, unless you're stealing, in which case you're getting more. His FIL is taking advantage of the investment the more expensive store has placed in salespeople and floor displays. His FIL wouldn't be paying "more than he should" at the fancy store, he'd be paying for the expertise and equipment, as opposed to just the equipment. If everyone did this, the stores that offered good service would go out of business, leaving just the volume retailers.
There's nothing illegal, or even horribly unethical. You're just taking advantage of the system, maybe akin to going to a party & eating all the snacks before someone else has a chance to - after all, there's no law that says you have to leave anything for anyone else.
-BbT
Although the idea may or may not be ultimately feasible for wide-scale use due to the presence of transformers, why hasn't the home/small office networking domain used household power lines for network cable? Given burgeoning popularity of wireless networks, couldn't a household powerline network acheive similar goals?
-BbT
My problem with Amazon is that the organization is geared towards sales, rather than understanding. The is demonstrated by the fact that a given book shows up multiple times based on whether it was hardcover, softcover, printing 1, 2, 3, whatever and so on. Click on a J.R.R. Tolkien link and you get 237 results. Tolkien didn't write 237 books - I should see just a list of the books (preferably broken down like imdb does, based on his role, author, editor, contributor, etc.). When I click on a book, I can see all the different forms it came in, and what printings, etc. Amazon's format is just a huge pile you have to sift through.
Also, no crossreferences (granted, not as necessary with books as compared to movies that take dozens of people to produce), no sorting by original publication date, etc. It needs an organization like allmusic or imdb.
Finally, a more formalized way for users to enter data besides comments, such as imdb's trivia, goofs, a better (normalized?) rating scheme (Amazon's star-based scheme rarely has books that drop below 3.5 stars). Someone mentioned above that Amazon owns imdb, but it was originally a distributed effort like this one.
-BbT
Something like CPU isn't really an acronym but in fact a "spelled" abbreviation (as opposed to a regular abbreviations like dept. or Dr. that is still pronounced as the original word). Prior to the 70's (there's that appostrophe again), spelled abbreviations were almost always delimited by periods (T.V.A., F.D.R., F.B.I.), and pronounced acronyms (NATO, radar) were not.
The idea, as I have always heard & understood it, is that merely putting an "s" on the end of an spelled abbreviation that is pronounced as a series of letters is confusing (do you spell or say the "s" - SEE PEE YOU ESS?), as you may think the "s" is part of the abbreviation. Traditionally, lettered abbreviations were delimited with periods, like M.D.; making them plural by adding a lowercase "s" didn't look right at all (M.D.s).
Soon after that you started seeing the creation of the InterCaps plural form of the spelled abbreviation (CDs). Now that InterCaps and spelled, non-periodified acronyms are relatively common, the traditional style guides are finally officially sanctioning the CPUs form of the plural, but the form CPU's is still correct, but falling out of favor. Back in the 60's it probably would be more correctly spelled C.P.U.'s - hmm.
Most style guides recognize that the ambiguity exists, and that grammar nazis will get their undies in a bundle, so their recommendation is to avoid sentence structures that force you to make a spelled abbreviation plural.
-BbT
This particular query was run on 8.1.6 Enterprise (Solaris & NT), but I've been working with Oracle since 7.3.4 and we're currently doing 9i. It functions similarly on all of them. It could be that the sub-enterprise releases of Oracle behave differently, but I'm not sure - we were using Personal Oracle on our laptops for our sales staff, and we didn't see any drop in performance once the I/O & CPU were taken into account.
I'm surpised that Oracle would be advocating bitmaps for everything. We use them a lot, and they're great, but they're not useful for everything, and have particular problem when used on high transaction tables. Plus, some columns (particularly DATE columns) don't lend themselves to bitmaps.
-BbT
This doesn't match with my experience. Order of the conditions in a WHERE clause doesn't effect the optimizer. Specifically, in this case, I create a sample million-row table:(I had to fiddle with the above because
I then run two queries, reversing the order of the constraints in the where clause:You can see that the performance of the query is the same in both cases. Am I missing something?
-BbT
-BbT
With geek items like this, the half-life is even shorter. Magic The Gathering cards are already past their prime in terms of collectable value; once the people who played the NES in their youth are past the age of buying this stuff, watch the prices plummet.
-BbT
Robert Park of the American Physical Society is one of the more outspoken physicists against the ISS. His argument is that it is ostensibly for scientific research, but is really a pork-barrel project and offers little benefit to researchers in terms of cost/scientific output. If its purpose were as a stepping stone to further space exploration, it would be a different matter, but as a research facility it is expensive and ineffectual (even once it comes fully on-line).
-BbT
Raid 0 is exactly as reliable as your least reliable hard drive
Actually, it's less reliable than your least reliable drive (assuming you're not using array of 1 disk). The chance of failure of the array is 1 - (product of all disks' chance of not failing). An array of 2 identical disks will fail twice as often as an individual disk.
-BbT
Let's assume that a, b and c are on three separate tables. One way this could be true would be for there to be a bug or weakness in the interpreter. If there were only an index on b, then a=b and b=c might resolve efficiently, but if the optimizer was given the three choices and chose the less optimal pair (a=b and a=c, for example), it might result in more work than a=b,c=b.
A non-flaw based explanation: Assuming that there are indexes on all tables, there's still the task of chosing which ones. The standard way that an RDBMS might resolve this query is to iterate over each table in succession, but there are six permutations of such a strategy (a->b->c, a->c->b, b->a->c, b->c->a, c->a->b, c->b->a). Only one may be chosen, but how it decides to do so will depend on how the system's optimizer is designed.
Some may strictly use "semantic" analysis - just look at the query & pick an option. Most will use statistics that are gathered by the system pertaining to the distribution of data & keys in the tables and attempt to estimate how much work each strategy will take based on row count & key cardinality. Depending on how accurately & frequently the system maintains these statistics, you can get different results, so even the same RDBMS might give multiple results depending on whether its statistics are up to date.
Optimizers generally don't infer equivalence relationships in order to allow the query writer to nudge the engine in one direction or another. With a=b,b=c, there are only two routes the most systems would take (a->b->c, c->b->a). By limiting the choices, you can coax the system to avoid a plan that may look attractive based on statistics but proves to be less practical for whatever reason.
Even if all the statistics are equivalent, two different systems may still produce different results. One plan for data retrieval might result in operations overall, but the may result in a long delay before you receive the first results (low throughput cost, high first row latency); conversely, another plan may do the opposite. Depending on your use of the data, you may prefer one over the other - for example, doing some summary financial report on a company's transactions would benefit from the low-throughput cost, since you need to look at all the data, but a Google type search where the user will scan the data in order and stop looking long before he or she reaches the end of the result set will benefit from a low first-row latency). One system may default to the first behavior, others to the second.
There are plenty of other reasons, but the main reason is that since optimizers have many ways of carrying out a request, and different performance characteristics may be desired from the same statement, they don't do 'simplification' in order to give the developer some control over the process (and also because simplification is not always simple to carry out & may unnecessarily consume CPU cycles).
-BbT
> > Actually, it was made from "The Castle of Cagliostro"
> Right. That's what I meant. What was that game called?
The game was Cliff Hanger (The guy in the game was referred to as "Cliff"). The great thing about it compared to Dragon's Lair was that you had a "hand action" button and a "foot action" button, and that whenever you died, some text appeared at the bottom that said "You should have gone left" or "You should have pressed hand". It was much longer than DL and you didn't have to spend so much money on it to finish.
I only ever saw it at one arcade, which incidentally had another laserdisk game I never saw anywhere else Star Rider- a racing game with sprite racecars superimposed over laserdisk background. It was ingenious - the laserdisk simply played at a rate according to the speed of your racer, and the tracks you raced on were surreal and impressive. They had names like Crystalopolis & Cubitania.
-BbT
Contract authentication. In a manner similar to the old "wax & signet ring", any document could be projected through the crystal and it's speckle pattern saved with the document. Although you couldn't verify that someone was who they said they were, you would be able to verify after the fact that this document did or did not originate from that person.
This could be extremely useful for e-commerce. When you're paying for your purchase, the system could send you an image of their choice (the invoice for example) and you could run it through your crystal & send it back. If there were a billing discrepancy later on, the company could produce the original image and you could demonstrate that it did or didn't match your card.
So, it's main use would be in keeping consumers from ripping off credit card companies with fake claims of "it wasn't me". Conversely, you could use it do prove that you didn't make a particular purchase.
-BbT
There is no better writer when it comes to explaining science. Although it covers a lot more than just physics, the broad coverage is useful when you want to follow up on information that leads you away from physics proper.
The only disadvantage is that it is no longer in print, but thanks to the wonder of the internet, that's no longer a problem.
-BbT
Bottom line: expanding that 2-digit packed field, from one byte to two bytes, to fix the future problem, would have cost several hundreds of thousands of 1970's dollars just for the hardware cost of the storage, never mind the professional labor cost to program it and do the conversion. It was decided for what, in hindsight, were good reasons: it will be much cheaper to fix it in the future, if in fact it is still a problem when that time comes.
I've heard this argument before, but I'm unclear about something: Why the decision to represent years in a BCD format (1 digit per nibble)? Why not represent the year as an 8-bit integer, offset from 1900? This would have allowed the byte to hold dates from 1900 - 2155? Take it one step farther, and have the other byte hold the Julian date portion (1-365), and you could have fit the whole date into 2 bytes (instead of the 3 bytes required for the 6-digit BCD form), and had simpler date arithmetic to boot? Surely the storage space cost savings would have justified the 33% decrease in size?
-BbT