The article, which is quite short, is more about the commonality of large software project failures than about the COTS / custom software debate. The author is talking about projects worth $170M; everybody knows that in this range, the project cost comes mostly not from licenses of "COTS" software but from implementation costs anyway!
In fact, COTS doesn't really mean "off-the-shelf" when you're talking about software like SAP for which every $1 paid in license fees must be complemented by $2-$5 in implementation cost / consulting!!
The rate of failure of these big projects largely qualifies the software industry as "immature".
Sure, there are other industries for which going overbudget or late is routine. Take the construction business, for one--and that's one of the oldest trades, so it is not really a question of maturity. 20 centuries from now, chances are builders will still always be late on schedule and overbudget. But at least, in general, you have a roof and four walls when you spend $170 million on a building!
Not so with software, where such large projects are trashed one after another--or even "smaller" corporate projects in the $1-$2M range. I think that's still amazing.
I would attribute this to the computer illiteracy of many of today's executives; not wanting to look like they don't know jack about technology, they gobble in vendor bullshit and are afraid of asking even the "dumb but obvious questions"! They should, and then they should keep that money for something else!
Btw if some big-ass corporate exec out there wants to give me another couple M$ for nothing in return, feel free.
1) Isn't the greatest benefit of XML that it can be opened in a text editor, and made sense of? I think that would be confusing debugging with regular operation. Opening an XML file in a text editor and "making sense of it" is not all too common of an operation (hopefully) in production, solely used in debugging. A binary format accompanied by good (open source!) authoring & viewing tools that put the binary format back into a readable format and allow for easy edition would be very much sufficient for the "text editor" case. 2) Can't webservers and browsers [...] transparently compress XML with gzip or some other?
Sure, but you are just trading bandwidth for CPU, i.e. just moving the problem somewhere else and not solving it. Also, not practical on embedded devices.
3) Making it binary won't compress it all that much, using a proper compression algo will. 4) Doesn't something like XML, that makes use of latin characters and a few punctuation marks, compress with insane ratios even in lame compression algo's?
Same issue. 5) In a world moving ever closer to ubiquitous broadband, is a difference between a 10kb html file and a 17kb XML file all that fatal? Surely bittorrent and spam does more to suck up all available bandwidth than XML does (what little is out there).
2 things:
* in large business systems treating huge loads of documents, there is always a bottleneck at some point somewhere. A small amount of supplementary bandwidth/CPU doesn't cost much to consumer users, who have about 100x more than needed on average; but for a production system, increasing file size by X% means increasing pipe size requirements by X%, and costs follow suit
* wireless bitrates do not increase all that much especially if you fix battery consumption, spectrum size and equipment bulk (latest standards tend to consume more battery power, larger swaths of spectrum and require bulkier equipment than established standards, which makes the transition to "higher bitrates" sometimes not practical)
Yes, ASN.1 implements a lot of excellent ideas from the get go, such as
* being an "abstract" format, i.e. considering data to be independent from its actual byte-wise representation
* ability to define space-smart encodings
* supporting canonicalization from the get-go
* use of a well-defined ISO namespace
* modularity of grammar definitions
But take a close look at it and you will see that unfortunately, it is a standard that is very difficult to interpret, crippled with obsolete string formats, and in practice not very well implemented. As a result, useful implementations will even sometimes have to have the ability to break compliance to work with other broken implementations.
For example, ASN.1 is the underlying language of the X.509 certificate standard, which is in turn used by IETF's PKIX, SSL and HTTPS standards. Canonicalization is supposed to allow the decision of "object equality" in a well-defined manner and known time. However, a widespread HTTPS browser (not IE) did not implement canonicalization in some parts of their implementation. As a result, interoperation with that browser required the implementer to actually violate canonicalization rules so that the object would be properly understood by the browser (!)
Another sign that ASN.1 might be a little bit too complex is the fact that there are no fully compliant open-source implementations of ASN.1 parsers, parser generators, parsing libraries etc. Even the commercial offering itself is not that good and dearly priced.
What would be nice is a simpler, leaner version of ASN.1 keeping the main structural features and getting rid of the problematic / obsolete features.
What's wrong with just compressing the XML as it is with an open and easy-to-implement algorithm like gzip or bzip2?
I'll tell you one thing that's wrong: these compression algorithms might run fine on your desktop or server; but on an embedded system with restricted memory and CPU power, that's another matter...
Yeah, but if the guy had only gone 1/100 as far from the transmitter, 5.4 miles instead of 540 miles, he would probably have received at least 10000 times the power (+40dB). He could then have afforded to divide the transmitted power by 100 (-20dB), and then he would still have gotten a 20dB boost as compared to his original experiment, making the signal much easier to pick up. However, his "performance" would have increased 100 times as measured with this stupid measure.
Anyway, as long as you know how to code data well you can make any transmission reliable, even with the weakest SNR, by using redundancy... so even without moving this measure is completely worthless.
The one thing that can be measured, on the other hand, is the time I'm wasting on Slashdot!
... in "miles per watt" anybody can beat this "record" pretty much anytime.
As many people have mentioned, the unit chosen is stupid. It is not about nitpicking, the unit has no meaning: this guy could easily beat his own record, not by going into space but simply by sitting half as far, or (even better) right next to the transmitting station. Indeed, all things being equal (hardware used, the ham's ear), the chosen measure increases linearly with the inverse of the distance, and more if the fall-off is higher (usually 3 or 4 in "real life").
As many people mentioned, power fall-off is inverse-square of distance in free space. That means that if the guy had stood half as far, all other things remaining equal, he would have received four times as much power. He could have then halved the transmitted power, he would still have received twice the power as in the reported experiment (making things easier for him to hear). But because the power halved and the distance halved in the process too, the "measure" of distance per power didn't change.
Is this a joke by radio amateurs trying to get the media or Slashdot to publish stupid claims?
I haven't figured out why so many people use G.711 - voice doesn't need this much bandwidth, and we all know this from years of working with mp3.
Simple-- there is just so much overhead that dividing your codec bandwidth does not increase the capacity much.
In fact, if you check out this good technical presentation by Spectralink, slide 13, you will see for example that a G.711 call (64 kbps both ways, i.e. at most 128 kbps) actually utilizes 4.5% of the bandwidth in "11 Mbps" mode (i.e. in the best radio conditions). That represents 500 kbps of nominal bandwidth, to carry a 128 kbps signal.
What happens when you use the G.729 (GSM) codec at 8 kbps instead, is that the quality goes down for for sure, but the capacity does not increase that much: one call utilizes 3.5% of the bandwidth in "11 Mbps" mode, which still represents 400 kbps of nominal bandwidth.
So while you divided your codec rate by 8, sacrificing quality, the capacity has only been increased by 30%, not by 700% as one would expect in a perfectly designed system.
Why bother with lousy codecs when the underlying layer adds so much overhead in any case?
Crystal clear calls... until there are 6 of you making calls at the hotspot! There are a lot of scalability problems with VoWiFi (concurrent users competing for bandwidth "collide" all the time, huge overhead for small packets, no call admission control...)
But vendors are clever: since the overhead is huge for small packets, they might as well use a high-quality 32 kbps codecs in there; so when you're trying out their solution with just one or two phones, you think "wow, the quality is better than any phone I ever tried! I can't believe there's so much crap said about VoWiFi..."
But just wait until more of these phones make calls at the same location, along with PCs downloading data, and no call admission control to "queue" users. You'll be happy to have your cellular phone then....
This news is just a rehash of an AOL press release, which is its top-10 list of spam topics of the year. It has a footnote:
This list is unscientific, and is not in any specific order. The cited email subject headers are not ranked by volume within the Top 10 list.
So that's really just about making "AOL" appear in the media one more time; so far so good since CNN and then Slashdot picks it up.
As to "Vioxx," I would think that probably the AOL PR intern who put together this list on the basis of his/her own spambox mixed it up with Viagra.
Uncertainty principle, as in Quantum Mechanics?! That's a really funny one! The only thing fundamentally limiting signals transmission is noise, which in electromagnetic transmissions comes in its most unavoidable form from thermal noise. There are other sources of "noise" or "interference": depending on the environment, interference can come from other communicating devices, a car's engine ignition sparks, thunder, a static discharge, a microwave oven... And worsening it all, media also always attenuate the signal, more or less, therefore weakening the Signal-to-Noise ratio which along with the available bandwidth determines the theoretical capacity of the transmission system. The uncertainty principle has nothing to do with it.
Any good introductory course in Digital Communications will cover this, and introduce the notion of QAM and constellations in passing. May I suggest this Digital Communications OpenCourseWare module (viva MIT) or, for a gentler introduction that might not be as complete, this set of slides (viva University of Cambridge).
And once you've read all about it, be sure to write up an entry about it in WikiPedia;-) -- or to update this one.
Radio systems utilizing multiple antennae (MIMO), may "trump" this to some extent by playing on the fact that spectrum is a three-dimensional medium and not a one-dimensional medium (contrary to twisted pairs or coaxial cables). On that, you may want to read this intro.
It's not oversimplified to say that, in fact it's a common practice. In QAM modulation systems (Quadrature and Amplitude Modulation), "complex amplitude" (i.e. amplitude + phase) modulates a waveform with not just two values (+1 and -1) but four values (1, i, -1, -i) or more...
The set of amplitude values is represented in the complex plane as a "constellation". At the receiving end, you have to "recover" the amplitude and the phase of the emitted signal, which is the process of synchronisation.
In cable (coaxial) networks, where attenuation is lower than in free-range transmissions, 64-QAM or even 256-QAM (a modulations with 256 complex values, an 16x16 square in the complex plane) is commonly used.
Power and range do limit the effectiveness of modulations. In GPRS/EDGE packet radio, for example, QAM schemes are varied as a function of radio conditions to avoid using too ambitious of a modulation for what the channel can support.
You can't compare two technologies solely on bitrate, you are forgetting power, range, spectrum, equipment size and equipment cost as factors in your comparison.
Cellular technology *can* be made to operate faster than your WLAN. In fact, some WiMAX equipment should operate faser than many WLANs, and it's not mobile yet but it's cellular. But then again, the client-side equipment will have to work first with a roof-mounted antenna and the base station should cost about 100x the price of your average WiFi access point. Engineering hardly works miracles, just good trade-offs.
I never quite realized the telecom illiteracy of/. readers reached such levels, but a +4 informative for this one is too much (I hope that by now it has fallen).
Hz (bandwidth) is not bps (bitrate)!
With a strong enough Signal-to-Noise ratio, you can fit the largest of bitrates in the narrowest of bandwidths. See the Hartley-Shannon law for this...
This is a multiple-antenna technology, called MIMO for Multiple-Input-Multiple-Output. Instead of modeling the "medium" (electromagnetic spectrum) as a "box" taking a single signal as input and outputting a single signal, the medium is modelled as a matrix taking multiple signals on input and outputting multiple signals, on the same frequencies and at the same time.
The capacity increase is *theoretically* limited by the max number of antennae in input or output: say there are N transmitters and N receivers, if you assign each transmitter to a receiver and place a shielded coax cable between the pair instead of relying on em free-range transmission, there should not be much interference for sure. Thus you have N times the throughput, and that's how MIMO throughput numbers are computed *for PR purposes*...
In practice, when you use the air medium, there is a good deal of interference between the antennae, but research can derive some increased capacity (by "diagonalizing" the MIMO channel, i.e. identifying linearly independent sub-channels in this matrix of channels).
However, the practical increase is *in no way* linear as a function of the size of the MIMO array as some would obviously like other people (press, stock analysts, investors...) to believe. Additionally, MIMO works best in certain environments; independent sub-channels are best created by channel irregularities that are at a distance comparable to the scale of the antenna array. It means, for instance, that MIMO does not generally improve much the capacity of line-of-sight transmission but provides better improvements indoors!
The FCC is indeed opening more unlicensed bands. It is getting to the point that operators can have pretty decent "elbow room" there. For a good idea of what's going to happen with WiMAX in such bands, check out this operator in Chicago, Boston, NYC, LA and some other places...
The FCC does in fact sell licences *only for a certain period of time*, typically 10 years. Therefore, a provider who would continue to provide services would probably have to repay for the rights to continue using the spectrum later on. So in a way, this is a lease.
The spectrum licences does not grant a monopoly to the highest bidder, far from it. In every spectrum band and every region, there are several licensees (at the very least 2, more commonly 5) and the FCC enforces restrictions on the total spectrum owned by a single entity.
This model is certainly better than what the FCC used to do before (choose licensees based on an ad-hoc examination process, or worse yet attribute them via lottery!!). It does try to enforce "opportunity costs". There are also "substantial service obligations" that are enforced to guard against people trying to organise the scarcity of spectrum by hogging it without using it.
One problem is, though, what is the value of spectrum now when the FCC might open up more spectrum in the future? In 1996, you didn't know that the FCC was going to auction twice the amount of spectrum again or that it was going to create more unlicensed spectrum bands...
Unfortunately, it has become a requirement, in the software industry, to try and patent as much as possible out of the innovation contained your products, be it only for defensive purposes.
As much as you hate it, you cannot just pretend that patents in your field don't exist, or else you might later unknowingly "infringe" on someone else's IP and be sued for "counterfeiting", even if you independently came up with the same invention.
Thus however "wrong" it might be to patent certain inventions, and I wouldn't necessarily disagree with you there, today you have no other business-smart choice but to apply for patents, be it only to protect your business against "IP sharks".
Which is why precisely it is good to try and change the minds of lawmakers, but please, not through this kind of vocal activism which prefers big-name bragging ("Hey we got Linus with us!!") to bringing a comprehensive proposal to the table allowing for a smooth transition out of the current mess.
You're right about the translation propping costs up.
However, the translation is usually postponed to a couple of years after the initial application while the effective protection date remains that of the initial application. In practice, this means that only patents with a demonstrated market potential get translated, which you have ample time to figure out before you actually fork out the translation costs.
Additionaly, steps have been taken to streamline EU-wide application such as the upcoming Community Patent in which only the claims need to be translated into each language (which is quite ridiculous, since nobody could probably make a lot of sense out of the claims section without the body of the application, but that is another matter...).
Litigation costs indeed much more money. My point was just that if somebody is going to criticize the patent system, at least they should have the basic facts right or will be quickly dismissed, in the general opinion, as neo-communist GNU junkies.
...is a proposition spelling how to transition from the current world to one where software patents are outlawed...
Because the problem is, companies have *already * invested in software patents in Europe. So take a large company that has applied for maybe 50 software patents over each year, worldwide, in the past 3 years.
Some companies do so because they believe that their software methods should be patented. And while it is true that some awarded software patents are outrageously stupid, some are really nontrivial.
Other companies have mostly seen a "tactical advantage" in doing this, because (1) at a certain level (read: non-technical execs, financial analysts, shareholders) the number of patents granted per headcount per year is thought to reveal the quality of a Research & Development organization, and (2) when you have a portfolio of patents of your own, people are less likely to attack you for infringement, out of fear that you will attack them in return.
You can agree or not with these reasons, but the reality is that they have pushed many companies to invest millions in software patenting. So, as long as activists out there don't propose a way for these companies to "land smoothly" in no-software-patent land, actions like these are very unrealistic.
It doesn't help that the group is using bogus figures, such as claiming an average cost of EUR 30,000 for patenting something. Application fees have been made very small (in the hundreds of EUR depending on the country). Patent attorney fees, from my own experience, are more likely to range in the EUR 2,500 range for single-country application, and twice or three times that for worldwide application. That is not EUR 30,000 at any rate, unless you count in the inventor's own time writing down his/her invention.
Except that pretty much anybody with high-school writing skills and a functional brain can produce a decent blog, whereas TV or movies does not really allow amateurism.
It's like community-access TV channels. Ever watch those? Man, they're so pathetic... That's what video blogs promises to bring you: community-access TV crap ON DEMAND
One reason why blogging (or reading in general, for that matter) is popular, is that you can access the content at your own pace.
Watching a video requires the willingness and ability to follow the pace of the videomaker--which restricts audience. While you can skim through a bad writer's rantings and see very quickly if there is anything of value in a couple of pages of text, doing so on video is impractical.
Additionally, a good-paced video is actually hard to edit, and not something that most of us have been trained for in school, contrary to writing.
When WiMAX equipment makers talk about a "range" of 30 miles, you can't compare that directly to the "range" of a WiFi connection.
For the purpose of computing these "30 miles", they usually assume a *roof-mounted antenna* in direct line of sight of the base station, with 1W of effective isotropic radiating power directed precisely towards the base station. The base station itself emits 1W in every direction and we're talking base stations worth several k$ each.
802.16 equipment is going to be much more expensive that 802.11b/g for a while, and that's not just because of volume effect. To enable high bandwidth, 802.16 to very large band modulations requiring highly dynamic power amplifiers, i.e. amplifiers that can work with the same power in a wide range of frequencies. Those cost $$$.
To compare that with your regular laptop-embedded WiFi card and your latest cheapo $40 access point, even with all the Pringles boxes you want, is just wrong.
802.16 is "wireless", yes, but not "mobile" at this point. And even when it will be, it won't be with the same range (antenna bulk and power constraints). When so-and-so promises you WiMAX is beating WLAN to its grave in terms of range and throughput, they are comparing apples and oranges really.
In the neighborhood I used to live in, all those people with expensive alarm systems got robbed by people who knew better than the legitimate occupants how the systems operated. Additionally, some folks' alarm systems keep ringing and ringing for no reason (high number of false positives), which (1) they had to suffer from and (2) had the effect that if anybody was really stealing from them, pretty much nobody in the neighborhood would care about it.
We on the other hand, without any electonic alarm system, never had one single break in. How? Small stickers on the windows conspicuously advertising alarm systems (who weren't there), and a discipline of locking every door and window pane that there was at night. When we went away for longer, steel wire was used to secure the window pane closings so that a simple crowbar wouldn't pry the window pane too easily.
Don't trust technology when it comes to home security. Just make sure you don't make yourself the weakest link in the neighborhood. Between two doors equally easy to open, and equally not armed with alarm systems, you can bet that the thief will try first the one that doesn't bear a "warning security system" sticker.
Now you just need to get convincing stickers without actually buying the alarm equipment, but you can figure that out...
I always remind myself that before the Internet stepped in, I did use my computer for creativity, especially music composition (on various "trackers" for Amiga, if you must know).
Come to think of it, it was pretty amazing given the poor technology of the times (a mere 2 MB RAM, endless floppy-swapping -- later, a "huge" 20 MB HD). The creativity of the programmers was itself amazing. They did their sound mixing routines alright, MIDI + sample synchro, and the user interface--the user interface!!--was the best thing ever.
And yet today, maybe 100 times the number of Windows PCs is out there, with 100 times the CPU power each, but I still can't find an honest tracker for my Windows machine-- when I say honest, I mean that won't crash my PC or will ask that I buy a damn compatible soundcard. I also mean "free," I mean come on, who's going to spend C-notes worth of professional sequencer software for just dabbling around!!
Dudes in the 90's, up there in Finland & other places, were swamping Europe with their trackers at a time when "electronic distribution" was a euphemism for a network of enthusiasts swapping floppies through the post and holding "copy parties" in each other's place.
Now we got the Internet for distribtion, we got fairly less fragmentation in the OS space, and you'd have thought it'd all have made it much easier?? Think again!
Sure, back then we weren't able to download Britney Spears MP3 for free... Hell, if we had, we wouldn't even have had the CPU power to decode it!! But what's the new thing there? I mean, you just listen to the same music as in the store, except cheaper...
To conclude: quit consuming pr0n and mp3's, start coding mind-opening stuff for masses to discover their own talents!
(and stop reading / posting on Slashdot too, I might add)
Slashdot Mirror
The article, which is quite short, is more about the commonality of large software project failures than about the COTS / custom software debate. The author is talking about projects worth $170M; everybody knows that in this range, the project cost comes mostly not from licenses of "COTS" software but from implementation costs anyway!
In fact, COTS doesn't really mean "off-the-shelf" when you're talking about software like SAP for which every $1 paid in license fees must be complemented by $2-$5 in implementation cost / consulting!!
The rate of failure of these big projects largely qualifies the software industry as "immature".
Sure, there are other industries for which going overbudget or late is routine. Take the construction business, for one--and that's one of the oldest trades, so it is not really a question of maturity. 20 centuries from now, chances are builders will still always be late on schedule and overbudget. But at least, in general, you have a roof and four walls when you spend $170 million on a building!
Not so with software, where such large projects are trashed one after another--or even "smaller" corporate projects in the $1-$2M range. I think that's still amazing.
I would attribute this to the computer illiteracy of many of today's executives; not wanting to look like they don't know jack about technology, they gobble in vendor bullshit and are afraid of asking even the "dumb but obvious questions"! They should, and then they should keep that money for something else!
Btw if some big-ass corporate exec out there wants to give me another couple M$ for nothing in return, feel free.
1) Isn't the greatest benefit of XML that it can be opened in a text editor, and made sense of?
I think that would be confusing debugging with regular operation. Opening an XML file in a text editor and "making sense of it" is not all too common of an operation (hopefully) in production, solely used in debugging.
A binary format accompanied by good (open source!) authoring & viewing tools that put the binary format back into a readable format and allow for easy edition would be very much sufficient for the "text editor" case.
2) Can't webservers and browsers [...] transparently compress XML with gzip or some other?
Sure, but you are just trading bandwidth for CPU, i.e. just moving the problem somewhere else and not solving it. Also, not practical on embedded devices.
3) Making it binary won't compress it all that much, using a proper compression algo will.
4) Doesn't something like XML, that makes use of latin characters and a few punctuation marks, compress with insane ratios even in lame compression algo's?
Same issue.
5) In a world moving ever closer to ubiquitous broadband, is a difference between a 10kb html file and a 17kb XML file all that fatal? Surely bittorrent and spam does more to suck up all available bandwidth than XML does (what little is out there).
2 things:
* in large business systems treating huge loads of documents, there is always a bottleneck at some point somewhere. A small amount of supplementary bandwidth/CPU doesn't cost much to consumer users, who have about 100x more than needed on average; but for a production system, increasing file size by X% means increasing pipe size requirements by X%, and costs follow suit
* wireless bitrates do not increase all that much especially if you fix battery consumption, spectrum size and equipment bulk (latest standards tend to consume more battery power, larger swaths of spectrum and require bulkier equipment than established standards, which makes the transition to "higher bitrates" sometimes not practical)
Yes, ASN.1 implements a lot of excellent ideas from the get go, such as
* being an "abstract" format, i.e. considering data to be independent from its actual byte-wise representation
* ability to define space-smart encodings
* supporting canonicalization from the get-go
* use of a well-defined ISO namespace
* modularity of grammar definitions
But take a close look at it and you will see that unfortunately, it is a standard that is very difficult to interpret, crippled with obsolete string formats, and in practice not very well implemented. As a result, useful implementations will even sometimes have to have the ability to break compliance to work with other broken implementations.
For example, ASN.1 is the underlying language of the X.509 certificate standard, which is in turn used by IETF's PKIX, SSL and HTTPS standards. Canonicalization is supposed to allow the decision of "object equality" in a well-defined manner and known time. However, a widespread HTTPS browser (not IE) did not implement canonicalization in some parts of their implementation. As a result, interoperation with that browser required the implementer to actually violate canonicalization rules so that the object would be properly understood by the browser (!)
Another sign that ASN.1 might be a little bit too complex is the fact that there are no fully compliant open-source implementations of ASN.1 parsers, parser generators, parsing libraries etc. Even the commercial offering itself is not that good and dearly priced.
What would be nice is a simpler, leaner version of ASN.1 keeping the main structural features and getting rid of the problematic / obsolete features.
For resources on ASN.1 and XML, including an XMLSchema-to-ASN.1 converter
What's wrong with just compressing the XML as it is with an open and easy-to-implement algorithm like gzip or bzip2? I'll tell you one thing that's wrong: these compression algorithms might run fine on your desktop or server; but on an embedded system with restricted memory and CPU power, that's another matter...
Yeah, but if the guy had only gone 1/100 as far from the transmitter, 5.4 miles instead of 540 miles, he would probably have received at least 10000 times the power (+40dB). He could then have afforded to divide the transmitted power by 100 (-20dB), and then he would still have gotten a 20dB boost as compared to his original experiment, making the signal much easier to pick up. However, his "performance" would have increased 100 times as measured with this stupid measure.
Anyway, as long as you know how to code data well you can make any transmission reliable, even with the weakest SNR, by using redundancy... so even without moving this measure is completely worthless.
The one thing that can be measured, on the other hand, is the time I'm wasting on Slashdot!
... in "miles per watt" anybody can beat this "record" pretty much anytime.
As many people have mentioned, the unit chosen is stupid. It is not about nitpicking, the unit has no meaning: this guy could easily beat his own record, not by going into space but simply by sitting half as far, or (even better) right next to the transmitting station. Indeed, all things being equal (hardware used, the ham's ear), the chosen measure increases linearly with the inverse of the distance, and more if the fall-off is higher (usually 3 or 4 in "real life").
As many people mentioned, power fall-off is inverse-square of distance in free space. That means that if the guy had stood half as far, all other things remaining equal, he would have received four times as much power. He could have then halved the transmitted power, he would still have received twice the power as in the reported experiment (making things easier for him to hear). But because the power halved and the distance halved in the process too, the "measure" of distance per power didn't change.
Is this a joke by radio amateurs trying to get the media or Slashdot to publish stupid claims?
I haven't figured out why so many people use G.711 - voice doesn't need this much bandwidth, and we all know this from years of working with mp3.
Simple-- there is just so much overhead that dividing your codec bandwidth does not increase the capacity much.
In fact, if you check out this good technical presentation by Spectralink, slide 13, you will see for example that a G.711 call (64 kbps both ways, i.e. at most 128 kbps) actually utilizes 4.5% of the bandwidth in "11 Mbps" mode (i.e. in the best radio conditions). That represents 500 kbps of nominal bandwidth, to carry a 128 kbps signal.
What happens when you use the G.729 (GSM) codec at 8 kbps instead, is that the quality goes down for for sure, but the capacity does not increase that much: one call utilizes 3.5% of the bandwidth in "11 Mbps" mode, which still represents 400 kbps of nominal bandwidth.
So while you divided your codec rate by 8, sacrificing quality, the capacity has only been increased by 30%, not by 700% as one would expect in a perfectly designed system.
Why bother with lousy codecs when the underlying layer adds so much overhead in any case?
Like the console that fits in a joystick or these oldish-looking cell phones... what's next in this trend?
Crystal clear calls... until there are 6 of you making calls at the hotspot! There are a lot of scalability problems with VoWiFi (concurrent users competing for bandwidth "collide" all the time, huge overhead for small packets, no call admission control...)
But vendors are clever: since the overhead is huge for small packets, they might as well use a high-quality 32 kbps codecs in there; so when you're trying out their solution with just one or two phones, you think "wow, the quality is better than any phone I ever tried! I can't believe there's so much crap said about VoWiFi..."
But just wait until more of these phones make calls at the same location, along with PCs downloading data, and no call admission control to "queue" users. You'll be happy to have your cellular phone then....
This news is just a rehash of an AOL press release, which is its top-10 list of spam topics of the year. It has a footnote: This list is unscientific, and is not in any specific order. The cited email subject headers are not ranked by volume within the Top 10 list. So that's really just about making "AOL" appear in the media one more time; so far so good since CNN and then Slashdot picks it up. As to "Vioxx," I would think that probably the AOL PR intern who put together this list on the basis of his/her own spambox mixed it up with Viagra.
Uncertainty principle, as in Quantum Mechanics?! That's a really funny one! The only thing fundamentally limiting signals transmission is noise, which in electromagnetic transmissions comes in its most unavoidable form from thermal noise. There are other sources of "noise" or "interference": depending on the environment, interference can come from other communicating devices, a car's engine ignition sparks, thunder, a static discharge, a microwave oven... And worsening it all, media also always attenuate the signal, more or less, therefore weakening the Signal-to-Noise ratio which along with the available bandwidth determines the theoretical capacity of the transmission system. The uncertainty principle has nothing to do with it.
;-) -- or to update this one.
Any good introductory course in Digital Communications will cover this, and introduce the notion of QAM and constellations in passing. May I suggest this Digital Communications OpenCourseWare module (viva MIT) or, for a gentler introduction that might not be as complete, this set of slides (viva University of Cambridge).
And once you've read all about it, be sure to write up an entry about it in WikiPedia
Radio systems utilizing multiple antennae (MIMO), may "trump" this to some extent by playing on the fact that spectrum is a three-dimensional medium and not a one-dimensional medium (contrary to twisted pairs or coaxial cables). On that, you may want to read this intro.
It's not oversimplified to say that, in fact it's a common practice. In QAM modulation systems (Quadrature and Amplitude Modulation), "complex amplitude" (i.e. amplitude + phase) modulates a waveform with not just two values (+1 and -1) but four values (1, i, -1, -i) or more...
The set of amplitude values is represented in the complex plane as a "constellation". At the receiving end, you have to "recover" the amplitude and the phase of the emitted signal, which is the process of synchronisation.
In cable (coaxial) networks, where attenuation is lower than in free-range transmissions, 64-QAM or even 256-QAM (a modulations with 256 complex values, an 16x16 square in the complex plane) is commonly used.
Power and range do limit the effectiveness of modulations. In GPRS/EDGE packet radio, for example, QAM schemes are varied as a function of radio conditions to avoid using too ambitious of a modulation for what the channel can support.
Bitrate is just one of the features of networks.
You can't compare two technologies solely on bitrate, you are forgetting power, range, spectrum, equipment size and equipment cost as factors in your comparison.
Cellular technology *can* be made to operate faster than your WLAN. In fact, some WiMAX equipment should operate faser than many WLANs, and it's not mobile yet but it's cellular. But then again, the client-side equipment will have to work first with a roof-mounted antenna and the base station should cost about 100x the price of your average WiFi access point. Engineering hardly works miracles, just good trade-offs.
I never quite realized the telecom illiteracy of /. readers reached such levels, but a +4 informative for this one is too much (I hope that by now it has fallen).
Hz (bandwidth) is not bps (bitrate)!
With a strong enough Signal-to-Noise ratio, you can fit the largest of bitrates in the narrowest of bandwidths. See the Hartley-Shannon law for this...
This is a multiple-antenna technology, called MIMO for Multiple-Input-Multiple-Output. Instead of modeling the "medium" (electromagnetic spectrum) as a "box" taking a single signal as input and outputting a single signal, the medium is modelled as a matrix taking multiple signals on input and outputting multiple signals, on the same frequencies and at the same time.
The capacity increase is *theoretically* limited by the max number of antennae in input or output: say there are N transmitters and N receivers, if you assign each transmitter to a receiver and place a shielded coax cable between the pair instead of relying on em free-range transmission, there should not be much interference for sure. Thus you have N times the throughput, and that's how MIMO throughput numbers are computed *for PR purposes*...
In practice, when you use the air medium, there is a good deal of interference between the antennae, but research can derive some increased capacity (by "diagonalizing" the MIMO channel, i.e. identifying linearly independent sub-channels in this matrix of channels).
However, the practical increase is *in no way* linear as a function of the size of the MIMO array as some would obviously like other people (press, stock analysts, investors...) to believe. Additionally, MIMO works best in certain environments; independent sub-channels are best created by channel irregularities that are at a distance comparable to the scale of the antenna array. It means, for instance, that MIMO does not generally improve much the capacity of line-of-sight transmission but provides better improvements indoors!
The FCC is indeed opening more unlicensed bands. It is getting to the point that operators can have pretty decent "elbow room" there. For a good idea of what's going to happen with WiMAX in such bands, check out this operator in Chicago, Boston, NYC, LA and some other places...
The FCC does in fact sell licences *only for a certain period of time*, typically 10 years. Therefore, a provider who would continue to provide services would probably have to repay for the rights to continue using the spectrum later on. So in a way, this is a lease.
The spectrum licences does not grant a monopoly to the highest bidder, far from it. In every spectrum band and every region, there are several licensees (at the very least 2, more commonly 5) and the FCC enforces restrictions on the total spectrum owned by a single entity.
This model is certainly better than what the FCC used to do before (choose licensees based on an ad-hoc examination process, or worse yet attribute them via lottery!!). It does try to enforce "opportunity costs". There are also "substantial service obligations" that are enforced to guard against people trying to organise the scarcity of spectrum by hogging it without using it.
One problem is, though, what is the value of spectrum now when the FCC might open up more spectrum in the future? In 1996, you didn't know that the FCC was going to auction twice the amount of spectrum again or that it was going to create more unlicensed spectrum bands...
Unfortunately, it has become a requirement, in the software industry, to try and patent as much as possible out of the innovation contained your products, be it only for defensive purposes.
As much as you hate it, you cannot just pretend that patents in your field don't exist, or else you might later unknowingly "infringe" on someone else's IP and be sued for "counterfeiting", even if you independently came up with the same invention.
Thus however "wrong" it might be to patent certain inventions, and I wouldn't necessarily disagree with you there, today you have no other business-smart choice but to apply for patents, be it only to protect your business against "IP sharks".
Which is why precisely it is good to try and change the minds of lawmakers, but please, not through this kind of vocal activism which prefers big-name bragging ("Hey we got Linus with us!!") to bringing a comprehensive proposal to the table allowing for a smooth transition out of the current mess.
You're right about the translation propping costs up.
However, the translation is usually postponed to a couple of years after the initial application while the effective protection date remains that of the initial application. In practice, this means that only patents with a demonstrated market potential get translated, which you have ample time to figure out before you actually fork out the translation costs.
Additionaly, steps have been taken to streamline EU-wide application such as the upcoming Community Patent in which only the claims need to be translated into each language (which is quite ridiculous, since nobody could probably make a lot of sense out of the claims section without the body of the application, but that is another matter...).
Litigation costs indeed much more money. My point was just that if somebody is going to criticize the patent system, at least they should have the basic facts right or will be quickly dismissed, in the general opinion, as neo-communist GNU junkies.
...is a proposition spelling how to transition from the current world to one where software patents are outlawed...
Because the problem is, companies have *already * invested in software patents in Europe. So take a large company that has applied for maybe 50 software patents over each year, worldwide, in the past 3 years.
Some companies do so because they believe that their software methods should be patented. And while it is true that some awarded software patents are outrageously stupid, some are really nontrivial.
Other companies have mostly seen a "tactical advantage" in doing this, because (1) at a certain level (read: non-technical execs, financial analysts, shareholders) the number of patents granted per headcount per year is thought to reveal the quality of a Research & Development organization, and (2) when you have a portfolio of patents of your own, people are less likely to attack you for infringement, out of fear that you will attack them in return.
You can agree or not with these reasons, but the reality is that they have pushed many companies to invest millions in software patenting. So, as long as activists out there don't propose a way for these companies to "land smoothly" in no-software-patent land, actions like these are very unrealistic.
It doesn't help that the group is using bogus figures, such as claiming an average cost of EUR 30,000 for patenting something. Application fees have been made very small (in the hundreds of EUR depending on the country). Patent attorney fees, from my own experience, are more likely to range in the EUR 2,500 range for single-country application, and twice or three times that for worldwide application. That is not EUR 30,000 at any rate, unless you count in the inventor's own time writing down his/her invention.
Yeah ;-)
Except that pretty much anybody with high-school writing skills and a functional brain can produce a decent blog, whereas TV or movies does not really allow amateurism.
It's like community-access TV channels. Ever watch those? Man, they're so pathetic... That's what video blogs promises to bring you: community-access TV crap ON DEMAND
One reason why blogging (or reading in general, for that matter) is popular, is that you can access the content at your own pace.
Watching a video requires the willingness and ability to follow the pace of the videomaker--which restricts audience. While you can skim through a bad writer's rantings and see very quickly if there is anything of value in a couple of pages of text, doing so on video is impractical.
Additionally, a good-paced video is actually hard to edit, and not something that most of us have been trained for in school, contrary to writing.
Sounds like a gimmick doomed to fail.
When WiMAX equipment makers talk about a "range" of 30 miles, you can't compare that directly to the "range" of a WiFi connection.
For the purpose of computing these "30 miles", they usually assume a *roof-mounted antenna* in direct line of sight of the base station, with 1W of effective isotropic radiating power directed precisely towards the base station. The base station itself emits 1W in every direction and we're talking base stations worth several k$ each.
802.16 equipment is going to be much more expensive that 802.11b/g for a while, and that's not just because of volume effect. To enable high bandwidth, 802.16 to very large band modulations requiring highly dynamic power amplifiers, i.e. amplifiers that can work with the same power in a wide range of frequencies. Those cost $$$.
To compare that with your regular laptop-embedded WiFi card and your latest cheapo $40 access point, even with all the Pringles boxes you want, is just wrong.
802.16 is "wireless", yes, but not "mobile" at this point. And even when it will be, it won't be with the same range (antenna bulk and power constraints). When so-and-so promises you WiMAX is beating WLAN to its grave in terms of range and throughput, they are comparing apples and oranges really.
In the neighborhood I used to live in, all those people with expensive alarm systems got robbed by people who knew better than the legitimate occupants how the systems operated. Additionally, some folks' alarm systems keep ringing and ringing for no reason (high number of false positives), which (1) they had to suffer from and (2) had the effect that if anybody was really stealing from them, pretty much nobody in the neighborhood would care about it.
We on the other hand, without any electonic alarm system, never had one single break in. How? Small stickers on the windows conspicuously advertising alarm systems (who weren't there), and a discipline of locking every door and window pane that there was at night. When we went away for longer, steel wire was used to secure the window pane closings so that a simple crowbar wouldn't pry the window pane too easily.
Don't trust technology when it comes to home security. Just make sure you don't make yourself the weakest link in the neighborhood. Between two doors equally easy to open, and equally not armed with alarm systems, you can bet that the thief will try first the one that doesn't bear a "warning security system" sticker.
Now you just need to get convincing stickers without actually buying the alarm equipment, but you can figure that out...
I always remind myself that before the Internet stepped in, I did use my computer for creativity, especially music composition (on various "trackers" for Amiga, if you must know).
Come to think of it, it was pretty amazing given the poor technology of the times (a mere 2 MB RAM, endless floppy-swapping -- later, a "huge" 20 MB HD). The creativity of the programmers was itself amazing. They did their sound mixing routines alright, MIDI + sample synchro, and the user interface--the user interface!!--was the best thing ever.
And yet today, maybe 100 times the number of Windows PCs is out there, with 100 times the CPU power each, but I still can't find an honest tracker for my Windows machine-- when I say honest, I mean that won't crash my PC or will ask that I buy a damn compatible soundcard. I also mean "free," I mean come on, who's going to spend C-notes worth of professional sequencer software for just dabbling around!!
Dudes in the 90's, up there in Finland & other places, were swamping Europe with their trackers at a time when "electronic distribution" was a euphemism for a network of enthusiasts swapping floppies through the post and holding "copy parties" in each other's place.
Now we got the Internet for distribtion, we got fairly less fragmentation in the OS space, and you'd have thought it'd all have made it much easier?? Think again!
Sure, back then we weren't able to download Britney Spears MP3 for free... Hell, if we had, we wouldn't even have had the CPU power to decode it!! But what's the new thing there? I mean, you just listen to the same music as in the store, except cheaper...
To conclude: quit consuming pr0n and mp3's, start coding mind-opening stuff for masses to discover their own talents!
(and stop reading / posting on Slashdot too, I might add)