PCI is relatively sensible and still very useful It's also a high speed paralell bus which means it takes up a lot of pins on whatever chip drives it and it's a bitch to route.
IMO removing it from the southbridge is a sensible descision. Motherboard vendors who want to offer PCI slots can always add a bridge chip (which has the bonus of making things easier to route since the bridge can be sited next to the PCI slots).
I do think they are being rather stingy on the PCIe though. 8 lanes is better than the 4-6 of current southbridges but it still pushes motherboard vendors to make difficult compromises between onboard functionality and number of slots.
I don't know about bluetooth, but for lots of applications USB serial ports won't work because USB operates at 5V and serial is supposed to be 12V. That isn't the biggest problem. Yes it's an annoyance when cable vendors cheap-out on the level shifters but it's not a difficult problem to solve. You can buy the USB-serial devices as bare chips and connect them to whatever level shfiter you need, +-10ish is piss easy to get if you hook up a FTDI chip to a max202 (which comes with built in voltage doublers and inverters). If that isn't enough (usually it will be) then you will have to go for a more old fassioned level shift chip and arrange for an appropriate power supply (either converted up from the USB power or from a seperate source).
Note that it's not unheard for for non-usb serial ports to have voltage problems too, especially on laptops.
A much bigger problem than voltage (and you may well be incorrectly balaming voltage for problems caused by this) is latency. Depending on how exactly the device handles the serial port the latency can be a killer and there is no easy fix. This especially applies to devices that use the serial port not for serial comms but as a bit-bangable IO port.
Unlike the easilly solved voltage problem the only soloutions to this one are either 1: abandon USB altogether 2: redesign the serial device or 3: move the low-level control of the serial device into a seperate microcontroller which then uses a different protocol to communicate over USB.
But that's ok, I've got my trusty PCI serial card... oh. PCIe serial converters do exist and should work as well as PCI ones.
You can buy expansion cards that run a PCI bridge off the PCIe bus. The problem with that is the physical side of things. I've seen adaptors but if you plugged a non low-profile card into them (and IME most cards other than network cards aren't low profile and network cards probablly aren't worth plugging into an adaptor) it wouldn't
I guess you could use one of those cards together with a big case and a flexible riser to put the card beyond the end of the motherboard but still a very messy soloution IMO.
There are also external expansion soloutions but they are both expensive and IMO messy (you have to start thinking about things like power-up order when using them afaict)
If there's a large enough market for industrial PCs that have PCI slots then no doubt some board maker will produce a motherboard with the chip built in. You can get core 2 boards with bloody ISA slots! It's pretty much a certainty that motherboard vendors will support PCI for years to come, the only question is at what price (for comparision IIRC the core 2 board with ISA slots was comparable in price to a high end gamer board).
I guess it's time to go out and buy some PC gear with PCI I think you are panicking a little too soon. Intel is planning to remove PCIe from their next generation of cheap-end chipsets. It will be quite a while before current gen chipsets are completely phased out and even longer before motherboards with PCI dissapear completely (heck you can get core 2 motherboards with ISA if you know where to look)
similar story here, on this (windows) system running upstream firefox with autoupdates (currently at 3.6.3) things are fine. On my linux box with debian provided iceweasel (not sure of the version offhand) things are broken.
I sometimes wonder if/. forget to test with slightly older versions of the browsers when they change something.
Do you actually need to be in the US or do you just need a US IP and phone number? (both of which i'm pretty sure I could if I wanted purchase pretty easily)
The Theory of relativity is a theory that is not testable, I wouldnt exactly say that isnt science. It's testable in the sense it can be used to make predictions and those predictions can be compared to what actually happened and what other theories e.g. newtons "laws" say.
Hence why they are called Theories, not Proofs or Laws. No physical theory can ever be conclusively proved (hence why we don't call them "laws" these days apart from a few that are called that for historical reasons) because there is always the chance that we will find a set of circumstances under which they break or find better experimenal methods that expose an inaccuracy but that doesn't mean we can't test them and document whether or not they are right to within the bounds ot experimental error.
I know PDF has embedded fonts, but that shouldn't take much room, should it? Embedded fonts can get pretty big if the software doesn't subset them or a lot of glyphs are used. DeJavu sans for example is over half a megabyte! Some fonts are much bigger (pan-unicode fonts and CJK fonts for example)
What are they doing that converts something that would be a 10K ASCII file into a 500K PDF monstrosity? PDFs will always be a bit bigger than plain text because they control the positioning of stuff exactly and that takes information. It shouldn't be a factor of 50 though unless images are involved.
Once images are involved the sky's the limit, a single large image can make a pdf huge (and remember images can be inserted at any resoloution so a huge image can display small!)
One of the things about pdfs is always embeds images and usually embeds fonts. This is a mixed blessing, on the one hand it makes the file far more portable than something like html but on the other hand it means you re-download stuff like logos with every pdf you grab.
Can't LaTeX handle it? LaTeX has it's place but afiact it was never designed to be a distrubution format. A typical LaTeX document involves a load of files that become figures in the document and many use LaTeX add-on packages that may or may not be installed.
About the only thing worse than PDFs are raster scans of documents, and those typically aren't served, they're used as an intermediate step towards porting to a more useful format. That has not been my experiance with large digitisation projects i've seen the output of (e.g. http://ethos.bl.ac.uk/ ). In my experiance they do OCR for searchability but the accuracy isn't good enough to do a full conversion so they produce pdfs with the image visible but OCR text for copy/paste/search.
It's done because it's a lot easier for computers to search text documents. Afaict this is the main reason for doing OCR at least in large digitisation projects.
And it saves lots of space. It does if you throw the originals away. But only an idiot would do that without careful proofreading of the OCRed text and careful proofreading costs a LOT more than storing the original images does.
The rendering of a html page depends on all sorts of factors. For starters there is window size, on a html page widths generally determine the flowing of the content unless a moronic web designer forces them to do otherwise. Then there is fonts. Heck there are web browsers that run on text terminals. Fonts are likely to be substituted depening on the platform and the particular install which will also affect the sizing of stuff.
Re:Does it have a monitor and full-size keyboard?
on
Flight of the Desktops
·
· Score: 1
I agree about the numeric keypad. If you do a lot of work on that, you can buy one of those pretty cheap It depends on WHY you want a numeric keypad.
Many laptops have some trickery that turns part of the keyboard into a numeric keypad whenever num-lock is on. This poses a problem for a designer of an external numeric keypad.
When I bought an external laptop numeric keypad (i've only tried one but have no reason to suspect others would be different) it got arround this problem by having it's own "num-lock" and generating scancodes that mapped to the keys on the main keyboard rather generating numeric keyboard scancodes.
This is a problem if you have software that treats the numeric keypad as a seperate entity rather than as simply duplicates of the corresponding main keyboard keys.
Re:Does it have a monitor and full-size keyboard?
on
Flight of the Desktops
·
· Score: 1
Any smaller (and I speak from experience with 10.4" machines) the keys do feel cramped. A lot of this depends on the particular machines. Some machines get a much better keyboard for a given size than others.
I for example have an (older) macbook and a HP mini 5101 (with the HD screen option). The keyboards feel very very similar (I just offered them up against each other and the keyboard width difference is about one key's worth). despite the fact the HP is a 10.something and the macbook is a a 13.something. Why? because the HP mini's keyboard goes right to the edge.
Both the macbook and the 5101 use the "chiklet" style which I personally consider a good thing but I know it's a bit of a love-hate subject.
The 5101s screen is more tiring on the eyes though due to the smaller pixels (total pixel counts are similar).
Re:Does it have a monitor and full-size keyboard?
on
Flight of the Desktops
·
· Score: 2, Interesting
You do realize laptops can have all of those things, right? Lets see
"8GB of memory" You can get that now (you can't get any more though) but it's only appeared as an option fairly recently and it's expensive. Meanwhile in a desktop form factor you get support for 16GB on fairly low end stuff (as long as it's recent) and 24GB on the high end desktop stuff.
"two 22" monitors" With most laptops you can have the internal monitor and one external monitor. If you want a matched pair of monitors this poses a problem as laptop monitors tend to cram the pixels in far tighter than desktop ones.
There are devices like the matrox dualhead2go and there are USB monitor adaptors but everything i've seen says they are shit compared to doing things the conventional way with a dual head graphics card.
"It's also got 4TB of disk space" You won't get that inside a laptop unless it's some massive monster and i'm not sure you would even then (even in the thicker size of laptop drive they only go up to 1TB afaict). You could hang it off usb I guess but that's slow and messy. There is esata but I have my doubts as to whether most laptops would get along with an eSATA port multiplier (afaict only the higher end controllers tend to support them) and most laptops only have one port.
Plus my experience has been that most laptops do not have adequate cooling to run at a reasonable noise level under heavy loads. They also often don't like running with the lid closed because the keyboard acts as a vent.
To me monster laptops just don't seem worth it unless you really need a lot of power on the go. Otherwise you are much better off with a lower spec more portable laptop and a fast expandable desktop to do the heavy lifting.
It depends on exactly what he means by the power supply, does he mean the power brick (which is usually fairly easy to replace though some laptops have weird connectors or voltage requirements which may mean a bit trickier) or does he mean the circuit inside that takes the 20V or so from the power brick and converts the power from mains input to computer, mains input to battery charging, and battery output to computer (which is likely to be somewhere between much harder and impossible to replace).
Of course even desktops these days have a significant part of the power supply circuitry on the motherboard even since logic voltages dropped below 3.3V
Pointing a detector (CCD, CMOS sensor, film or whatever) straight at your target doesn't work. Light from all sorts of angles will hit every pixel and what you get will not be an camera but merely a sensor for ambient light levels.
To get an image of an object you need an optical system that will create a mapping between the angle at which the light enters the system and the position the light ends up in on the sensor. There are many ways to do this (the simplest being a pinhole) but mirrors tend to work best for the large aperture (need lots of light to resolve distant) and high magnification (distant objects are generally small) required in a telescope.
I can think of a couple of advantages MSN had in the market at least here in the UK.
1: MSN messenger was both bundled with windows (in a renamed and slightly tweaked form) and pushed through the hotmail service which was one of the most popular webmail providers at the time. I remember huge posters for hotmail in cyber-cafes. 2: MSN messenger used your email address. If you had a hotmail address (see comment above) it just worked with your hotmail login details without needing any extra setup. It could also be used with other email addresses through the "msn passport" system. ICQ used numerical identifiers which have reached a length comparable with (possiblly even greater than by now) a narional phone number but with none of the structure that phone numbers have.
unintended intersection of the flight path and the ground, in most ungraceful ways LOL
Kinda similar to a story I heard from my phd supervisor (dunno if it's true or not and my memory of the exact term may be hazy) that someone told him not to use the "M word" and instead to call the things they were talking about single use unmanned air vehicles (i'll leave the reader to figure out what the "M word" was).
If so, what's the attraction? The attraction isn't for us geeks and hardcore online gamers who will most likely have a wired network in place already.
Afaict most normal people don't have a wired home network and many people won't have thier modem/router sitting next to the TV the TV they plan to game on. For this reason many users are likely te be on wireless. May as well make it as good wireless as possible and avoid an ugly extra box.
what country are you in? your statement seems plain WRONG for both the US and the UK.
UK: mac mini with screen keyboard and mouse: £1,362.00 27 inch core i7 imac £1,794.00
US: $1,696.00 and $2,199.00
Granted though if you are going to buy a complete setup from apple the imac looks a better bet (a 27 inch C2D imac costs about the same as the mac mini+screen+keyboard and is certainly a better system). Mainly because apple screens are so f*cking expensive.
OTOH if you want to get the cheapest possible legit mac setup your best bet it to buy the mac mini from apple and shop elsewhere for your keyboard monitor and mouse.
Yeah, in particular since most of the oil from the Middle East goes to Europe, Japan, and Asia. So what?
What matters are
1: how many resources are openly sold at open market prices (resource consumers want that to be large because it keeps prices low). 2: for resources that aren't sold at market rates who gets them and at what discount. 2: where the money from the difference between extraction costs and market value ends up.
There are a few reasons that invisible is nastier than visible.
Firstly while the bulk of the light does go in one direction dust in the air, imperfect surfaces on optical components and so on will generally make you aware that a powerful visible laser is on.
Secondly there is the blink reflex, this limits the length of an exposure to hazardous levels of visible light. With invisible light that won't happen.
Slashdot Mirror
PCI is relatively sensible and still very useful
It's also a high speed paralell bus which means it takes up a lot of pins on whatever chip drives it and it's a bitch to route.
IMO removing it from the southbridge is a sensible descision. Motherboard vendors who want to offer PCI slots can always add a bridge chip (which has the bonus of making things easier to route since the bridge can be sited next to the PCI slots).
I do think they are being rather stingy on the PCIe though. 8 lanes is better than the 4-6 of current southbridges but it still pushes motherboard vendors to make difficult compromises between onboard functionality and number of slots.
It would but I dread to think how the typical lusers would get on with a machine with a serial port that looked like the ethernet port.
I don't know about bluetooth, but for lots of applications USB serial ports won't work because USB operates at 5V and serial is supposed to be 12V.
That isn't the biggest problem. Yes it's an annoyance when cable vendors cheap-out on the level shifters but it's not a difficult problem to solve. You can buy the USB-serial devices as bare chips and connect them to whatever level shfiter you need, +-10ish is piss easy to get if you hook up a FTDI chip to a max202 (which comes with built in voltage doublers and inverters). If that isn't enough (usually it will be) then you will have to go for a more old fassioned level shift chip and arrange for an appropriate power supply (either converted up from the USB power or from a seperate source).
Note that it's not unheard for for non-usb serial ports to have voltage problems too, especially on laptops.
A much bigger problem than voltage (and you may well be incorrectly balaming voltage for problems caused by this) is latency. Depending on how exactly the device handles the serial port the latency can be a killer and there is no easy fix. This especially applies to devices that use the serial port not for serial comms but as a bit-bangable IO port.
Unlike the easilly solved voltage problem the only soloutions to this one are either 1: abandon USB altogether 2: redesign the serial device or 3: move the low-level control of the serial device into a seperate microcontroller which then uses a different protocol to communicate over USB.
But that's ok, I've got my trusty PCI serial card... oh.
PCIe serial converters do exist and should work as well as PCI ones.
You can buy expansion cards that run a PCI bridge off the PCIe bus.
The problem with that is the physical side of things. I've seen adaptors but if you plugged a non low-profile card into them (and IME most cards other than network cards aren't low profile and network cards probablly aren't worth plugging into an adaptor) it wouldn't
I guess you could use one of those cards together with a big case and a flexible riser to put the card beyond the end of the motherboard but still a very messy soloution IMO.
There are also external expansion soloutions but they are both expensive and IMO messy (you have to start thinking about things like power-up order when using them afaict)
If there's a large enough market for industrial PCs that have PCI slots then no doubt some board maker will produce a motherboard with the chip built in.
You can get core 2 boards with bloody ISA slots! It's pretty much a certainty that motherboard vendors will support PCI for years to come, the only question is at what price (for comparision IIRC the core 2 board with ISA slots was comparable in price to a high end gamer board).
I guess it's time to go out and buy some PC gear with PCI
I think you are panicking a little too soon. Intel is planning to remove PCIe from their next generation of cheap-end chipsets. It will be quite a while before current gen chipsets are completely phased out and even longer before motherboards with PCI dissapear completely (heck you can get core 2 motherboards with ISA if you know where to look)
similar story here, on this (windows) system running upstream firefox with autoupdates (currently at 3.6.3) things are fine. On my linux box with debian provided iceweasel (not sure of the version offhand) things are broken.
I sometimes wonder if /. forget to test with slightly older versions of the browsers when they change something.
Do you actually need to be in the US or do you just need a US IP and phone number? (both of which i'm pretty sure I could if I wanted purchase pretty easily)
The Theory of relativity is a theory that is not testable, I wouldnt exactly say that isnt science.
It's testable in the sense it can be used to make predictions and those predictions can be compared to what actually happened and what other theories e.g. newtons "laws" say.
Hence why they are called Theories, not Proofs or Laws.
No physical theory can ever be conclusively proved (hence why we don't call them "laws" these days apart from a few that are called that for historical reasons) because there is always the chance that we will find a set of circumstances under which they break or find better experimenal methods that expose an inaccuracy but that doesn't mean we can't test them and document whether or not they are right to within the bounds ot experimental error.
And likely there will be other changes in that time to cloud the issue.
True but this is offset by the fact you often have to wait to get in touch with the human in the first place.
I know PDF has embedded fonts, but that shouldn't take much room, should it?
Embedded fonts can get pretty big if the software doesn't subset them or a lot of glyphs are used. DeJavu sans for example is over half a megabyte! Some fonts are much bigger (pan-unicode fonts and CJK fonts for example)
What are they doing that converts something that would be a 10K ASCII file into a 500K PDF monstrosity?
PDFs will always be a bit bigger than plain text because they control the positioning of stuff exactly and that takes information. It shouldn't be a factor of 50 though unless images are involved.
Once images are involved the sky's the limit, a single large image can make a pdf huge (and remember images can be inserted at any resoloution so a huge image can display small!)
One of the things about pdfs is always embeds images and usually embeds fonts. This is a mixed blessing, on the one hand it makes the file far more portable than something like html but on the other hand it means you re-download stuff like logos with every pdf you grab.
Can't LaTeX handle it?
LaTeX has it's place but afiact it was never designed to be a distrubution format. A typical LaTeX document involves a load of files that become figures in the document and many use LaTeX add-on packages that may or may not be installed.
About the only thing worse than PDFs are raster scans of documents, and those typically aren't served, they're used as an intermediate step towards porting to a more useful format.
That has not been my experiance with large digitisation projects i've seen the output of (e.g. http://ethos.bl.ac.uk/ ). In my experiance they do OCR for searchability but the accuracy isn't good enough to do a full conversion so they produce pdfs with the image visible but OCR text for copy/paste/search.
It's done because it's a lot easier for computers to search text documents.
Afaict this is the main reason for doing OCR at least in large digitisation projects.
And it saves lots of space.
It does if you throw the originals away. But only an idiot would do that without careful proofreading of the OCRed text and careful proofreading costs a LOT more than storing the original images does.
No
The rendering of a html page depends on all sorts of factors. For starters there is window size, on a html page widths generally determine the flowing of the content unless a moronic web designer forces them to do otherwise. Then there is fonts. Heck there are web browsers that run on text terminals. Fonts are likely to be substituted depening on the platform and the particular install which will also affect the sizing of stuff.
I agree about the numeric keypad. If you do a lot of work on that, you can buy one of those pretty cheap
It depends on WHY you want a numeric keypad.
Many laptops have some trickery that turns part of the keyboard into a numeric keypad whenever num-lock is on. This poses a problem for a designer of an external numeric keypad.
When I bought an external laptop numeric keypad (i've only tried one but have no reason to suspect others would be different) it got arround this problem by having it's own "num-lock" and generating scancodes that mapped to the keys on the main keyboard rather generating numeric keyboard scancodes.
This is a problem if you have software that treats the numeric keypad as a seperate entity rather than as simply duplicates of the corresponding main keyboard keys.
Any smaller (and I speak from experience with 10.4" machines) the keys do feel cramped.
A lot of this depends on the particular machines. Some machines get a much better keyboard for a given size than others.
I for example have an (older) macbook and a HP mini 5101 (with the HD screen option). The keyboards feel very very similar (I just offered them up against each other and the keyboard width difference is about one key's worth). despite the fact the HP is a 10.something and the macbook is a a 13.something. Why? because the HP mini's keyboard goes right to the edge.
Both the macbook and the 5101 use the "chiklet" style which I personally consider a good thing but I know it's a bit of a love-hate subject.
The 5101s screen is more tiring on the eyes though due to the smaller pixels (total pixel counts are similar).
You do realize laptops can have all of those things, right?
Lets see
"8GB of memory"
You can get that now (you can't get any more though) but it's only appeared as an option fairly recently and it's expensive. Meanwhile in a desktop form factor you get support for 16GB on fairly low end stuff (as long as it's recent) and 24GB on the high end desktop stuff.
"two 22" monitors"
With most laptops you can have the internal monitor and one external monitor. If you want a matched pair of monitors this poses a problem as laptop monitors tend to cram the pixels in far tighter than desktop ones.
There are devices like the matrox dualhead2go and there are USB monitor adaptors but everything i've seen says they are shit compared to doing things the conventional way with a dual head graphics card.
"It's also got 4TB of disk space"
You won't get that inside a laptop unless it's some massive monster and i'm not sure you would even then (even in the thicker size of laptop drive they only go up to 1TB afaict). You could hang it off usb I guess but that's slow and messy. There is esata but I have my doubts as to whether most laptops would get along with an eSATA port multiplier (afaict only the higher end controllers tend to support them) and most laptops only have one port.
Plus my experience has been that most laptops do not have adequate cooling to run at a reasonable noise level under heavy loads. They also often don't like running with the lid closed because the keyboard acts as a vent.
To me monster laptops just don't seem worth it unless you really need a lot of power on the go. Otherwise you are much better off with a lower spec more portable laptop and a fast expandable desktop to do the heavy lifting.
It depends on exactly what he means by the power supply, does he mean the power brick (which is usually fairly easy to replace though some laptops have weird connectors or voltage requirements which may mean a bit trickier) or does he mean the circuit inside that takes the 20V or so from the power brick and converts the power from mains input to computer, mains input to battery charging, and battery output to computer (which is likely to be somewhere between much harder and impossible to replace).
Of course even desktops these days have a significant part of the power supply circuitry on the motherboard even since logic voltages dropped below 3.3V
Pointing a detector (CCD, CMOS sensor, film or whatever) straight at your target doesn't work. Light from all sorts of angles will hit every pixel and what you get will not be an camera but merely a sensor for ambient light levels.
To get an image of an object you need an optical system that will create a mapping between the angle at which the light enters the system and the position the light ends up in on the sensor. There are many ways to do this (the simplest being a pinhole) but mirrors tend to work best for the large aperture (need lots of light to resolve distant) and high magnification (distant objects are generally small) required in a telescope.
I can think of a couple of advantages MSN had in the market at least here in the UK.
1: MSN messenger was both bundled with windows (in a renamed and slightly tweaked form) and pushed through the hotmail service which was one of the most popular webmail providers at the time. I remember huge posters for hotmail in cyber-cafes.
2: MSN messenger used your email address. If you had a hotmail address (see comment above) it just worked with your hotmail login details without needing any extra setup. It could also be used with other email addresses through the "msn passport" system. ICQ used numerical identifiers which have reached a length comparable with (possiblly even greater than by now) a narional phone number but with none of the structure that phone numbers have.
It's not, it's just an example of how sony has gone back on similar statements in the past.
unintended intersection of the flight path and the ground, in most ungraceful ways
LOL
Kinda similar to a story I heard from my phd supervisor (dunno if it's true or not and my memory of the exact term may be hazy) that someone told him not to use the "M word" and instead to call the things they were talking about single use unmanned air vehicles (i'll leave the reader to figure out what the "M word" was).
If so, what's the attraction?
The attraction isn't for us geeks and hardcore online gamers who will most likely have a wired network in place already.
Afaict most normal people don't have a wired home network and many people won't have thier modem/router sitting next to the TV the TV they plan to game on. For this reason many users are likely te be on wireless. May as well make it as good wireless as possible and avoid an ugly extra box.
Admin tools are a special case. Why bother with injection attacks when one click will get you a box to type in and run whatever sql you like!
Don't give someone access to phpmyadmin if you wouldn't give them direct access to the database server!
what country are you in? your statement seems plain WRONG for both the US and the UK.
UK:
mac mini with screen keyboard and mouse: £1,362.00 27 inch
core i7 imac £1,794.00
US: $1,696.00 and $2,199.00
Granted though if you are going to buy a complete setup from apple the imac looks a better bet (a 27 inch C2D imac costs about the same as the mac mini+screen+keyboard and is certainly a better system). Mainly because apple screens are so f*cking expensive.
OTOH if you want to get the cheapest possible legit mac setup your best bet it to buy the mac mini from apple and shop elsewhere for your keyboard monitor and mouse.
Yeah, in particular since most of the oil from the Middle East goes to Europe, Japan, and Asia.
So what?
What matters are
1: how many resources are openly sold at open market prices (resource consumers want that to be large because it keeps prices low).
2: for resources that aren't sold at market rates who gets them and at what discount.
2: where the money from the difference between extraction costs and market value ends up.
There are a few reasons that invisible is nastier than visible.
Firstly while the bulk of the light does go in one direction dust in the air, imperfect surfaces on optical components and so on will generally make you aware that a powerful visible laser is on.
Secondly there is the blink reflex, this limits the length of an exposure to hazardous levels of visible light. With invisible light that won't happen.