It is notable that of the 10 top words in the list in the original article, 8 of them are obviously portmanteaux or frankenwords:
1. ginormous is from gigantic and enormous
2. confuzzled from confused and puzzled.
4. chillax from chill out and relax
6. gription from grip and friction
7. phonecrastinate from phone and procrastinate.
8. slickery from slick and slippery
9. snirt from snow and dirt
10. lingweenie from linguist and weenie
For several of these, Nos. 1, 2, 4, 6, and 8, the meaning remains close to meaning of the original component words, as these already are synonyms of each other and the resulting portmanteau word. Nos. 7, 9, and 10 produce special cases of one of the words. Furthermore, except for number 9, which may provide a useful description of an unpleasant condition, most of these would likely be considered insufferable cute-ish and contrived by anyone overexposed to them.
The "5. cognitive displaysia" seems to be derived from some kind of medical technical term, to make it seem more serious than it is. It still seems to be a nerdy way of saying that I'm bothered by forgetting which stuff I might have forgot before I leave, or something... Perhaps the more succinct term isn't such a waste after all.
This leaves me with the conclusion that the word "woot" is the best of the lot, as it is the more rare, non-portmanteau new-word, that also actually has a distinct meaning.
...claiming "it's difficult to envisage Groklaw's conjecture swaying a court case, but it provides SCO with valuable public relations ammunition."
Of course it is difficult to envisage this, since Groklaw isn't about trying to sway any court case. The lawyers for IBM and SCOX are the ones to make that kind of impact, and they have to rely on whatever are the facts. Groklaw is merely reporting the various twists and turns of this case, and in this process it is helping the lawyers and the technical people understand what the other say and how they think. As far as any conjecture being presented there, this appears mostly based on available material, such as court filings and technical documentation generally available. There may seem to be a bias against SCO, but if the realities of the case had been different, there could just as well have been an apparent bias against IBM.
As for being valuable to SCO PR, I'd beg to differ! Groklaw has been able to neutralize much of the FUD that Darl McBride would have us believe about SCO owning and controlling this and that... if anything can be compared to ammunition, it would be the torpedoes that are about to sink SCO. Even without Groklaw, SCO would eventually have foundered, it would have taken longer, and resulted in a slower growth of Linux.
Operatic... hmm, well, as it is said: "It ain't over 'till the fat lady sings." Actually, I've heard that the original saying comes from pool, where "it ain't over 'till the fat lady sinks", and the "fat lady sinking" referred to the 8-ball going into one of the pockets. Unfortunately, the website where this was noted is gone, and most places, including Google, has the fat lady singing. So this might be apocryphal.
Though not really questioning Tetschner and his travelling companion's sex or girth, it seems that the pool version with "sinks" really was most applicable here.:)
In other words, no more pulling out a drive to virus-scan it then replacing it or replacing a drive on an OEM machine - that won't allow it to boot.
Probably right about the virus-scan. Outside the machine, the drive probably will look like it is full of garbage.
However, I don't think replacement will become impossible. If the machines won't allow replacement disks, this means that a disk failure will result in a useless machine; this will probably also get in the way of people wanting to add disks -- and the people wanting to put Linux on a second-hand machine will cry foul -- so this is going to fly as well as those boat-anchors those machines would become.
And this iteration of Longhorn at least will not require these chips... you won't have to buy new motherboards just now. But, perhaps further down the line this may become a required peripheral for Longhorn, but this will not be until most motherboards have it in place.
It looks like mostly a way of keeping stuff on hard-drives secret. As such this is not so bad in view of how frequent notebook-theft is, or how big the security problems of second-hand equipment are.
This was a while ago, when I was at the University.
One day in the RF and Microwaves lab, we wanted to listen to some radio programme. Now, there were several kinds of receivers lying around, but most of them were in the higher frequency ranges.
Except for the HP140 series spectrum analyzers. These had a range going to 110 MHz, just right for FM broadcast radio.
So we made a simple antenna from a pair of wires that we stuck onto one of the windows, and connected this to the input of one of the HP-140 series spectrum analyzers. In addition to the convenient range, they had all kinds of nice filtering functions to limit bandwidth to an FM radio channel. We could even see the various broadcasting stations on the display.
These analyzers had a vertical output, to which we hooked up a linear power amplifier that originally was designed for driving a tele-coil system for the benefit of the hard of hearing. We hooked up this to a speaker that originally had been in someone's car but became surplus when they got a new stereo system.
By stopping the sweep and tuning the spectrum analyzer a little to the side of the frequency of interest, we got slope detection of the radio signal, and we got the sound of the station of interest.
Of course, the sound quality wasn't the best, obviously not stereo; and we noticed that the spectrum analyzer would slowly drift into and out of tune with the temperature variations of the day, so frequent readjustments were necessary. Fading was also noticed, but this wasn't too bad. Still, this set-up remained in use for quite a while, something around a year.
So this is how to make a set of $~10^4 equipment sound much like any old $~1 radio as found at flea markets, garage sales, or thrift stores...
From what I have understood from the GPL, and specifically that quote of it, it seems that the criterion for sufficiently close linkage for the GPL's "derived work" can be identified as "the same process". Like the parent poster says, there needs to be a legal ruling, but for now, this seems to be what is meant by the language in the GPL. (IANAL...)
Now with fonts being executable code, (some kind of shared library or DLL) running in the process of the word processor or printer formatting program, wouldn't the GPL's conditions for close linkage also apply to these programs? The document is more of a carrier; though the presence of font-statements could be construed as invocations of the code in the font-libraries, and thus the document would become a kind of source code itself.
By publishing the source code of the fonts along with the document, this GPL requirement would seem to be satisfied. However, the GPL may still clash with other restrictions that the document author may want to impose, for example, if the document describes something that must be kept confidential, this will conflict with the GPL requirement of no additional restriction on distribution. Even making the fonts LGPL wouldn't get around this one.
However, the purpose of documents really is to disseminate information that in itself is completely irrelevant to the word processor or printer, not unlike the comments of source code or shell scripts. But unlike these, the execution of font code is incidential to this main purpose, and shouldn't get in the way.
Perhaps the situation could be resolved by looking at the analogy of an interpreter. The word processor would be the interpreter program, the fonts would correspond to specialty libraries, and the document would be the script to be interpreted. Let us presume the easiest case of the interpreter being closed-source, and the libraries being LGPL so there is no a priori conflict there. What will be the possible licences of the script?
"IP" does exist, but it does not denote the ideas that somebody has in their brain, for whatever value of "ideas".
"IP" is an unfortunate collective term for the distinct concepts of "copyrights", "patents", "trademarks", and "trade secrets",
all of which can be ascribed a monetary value, and thus qualify as "property" in an accountant's ledger. FUDsters love to bandy this term about, as it is nice and short and succinct, and completely inaccurate. To that extent, the parent poster is right: "IP" as a term ought not to exist from a utilitarian point of view.
Unfortunately it still does.
In the case of the GPL, the relevant concept is first of all "copyrights", on which underlying law the GPL itself relies. The whole idea of the GPL is to give a limited permission to copy and redistribute beyond what copyright law permits. The owner of that copyright does not lose it.
However, the GPL is basically not compatible with "trade secrets", and I suspect that here is what rubs Mr Schwartz the wrong way. However, even here there is a straighforward and clear direction to take: Avoid mixing GPLd code and code that is considered a "trade secret", lest the competitive advance due to this "trade secret" become lost. That is all there is to it. That of course also means not being able to exploit the advantages that using GPLd code can provide.
It comes down to a comparison between the value of these advantages vs the advantage of keeping the "trade secret" a secret.
I would think that any blocking actions caused by morality would be routed around. And presumably, the evil bit would be crucial to determine the quality of morals.
As for blockin pr0n, who cares. It would really be useful if this could block spam!
Still, problem is, who gets to sit and determine what is immoral and what isn't? And for that matter, whose morals are we talking about here, is it the FCC, the WWE, the people behind goatse, the Republican Right, or maybe some Islamic Council for the Promotion of Virtue and Prevention of Vice? And who will pay them? Again, remember the system's ability to block any obstacles including taxation; the ARPAnet people back in the day thought in terms of nuclear bombs not nuclear family value enforcements...
Then again, considering the date of issue of this RFC, it is all in the +1 Funny realm anyway, which as we know, is karma neutral.
I guess one could hook up an ADC and/or a DAC with some kind of serial interface, perhaps I2C, to some of the 5 GPIO lines. However, as the unit stands, it seems to be most suitable for a "RS232-to-Ethernet" converter application, to easily add networking to equipment with the simpler interface.
Although 5 GPIO lines and an RS232-style serial connection does allow for connecting other interesting kit, what would have been really nice here, would have been a larger number of such general digital IO lines, and perhaps some additional peripherals, such as a timer/counter, ADC, and DAC.
Barring MS from the EU might make too much trouble for each country.
One thing that might happen, however, is enforced free licensing of the MS patents that, according to the article, MS claim would be infringed otherwise. Along with forced disclosure of the interfaces, and probably some guarantees that this openness would be preserved in future revisions.
The only really good ideas I've seen here are the short extension cords for moving the wall-warts off the power strips, or alternately, power strips with spacing that allows the warts to sit on them so that they don't block other outlets.
The other DC-distribution system design exercises remain impractical for a couple other reasons that I can easily pull off the tip of my tongue.
First reason: Voltage, and power variations. Around here I have various kit with power-supplies rated for voltages from 3.7V DC, 5 V DC, 7.5V DC, 2x8V AC (16 V center-tapped), 9V AC, 12 V AC, 12 V DC, 19 V DC and 21 V DC. I would suspect most other installations have a similar variety.
OK, so the various DC voltages could be hacked using a variety of linear or switching regulators off an ex-equipment PSU, though the 19 V and 21 V might present an educational challenge. Besides the fact that that this is likely to be at least as unsightly as the handful of wall-warts it replaces, this does absolutely nothing for the AC supplies, which wall-warts are simple transformers and rather efficient anyways. (the dual 8V unit has its own power cord and isn't even a wall-wart). Implementing DC-AC conversion for these purposes while regular mains AC is right there seems to be way overkill.
The second reason is reliability, as in "everything put together sooner or later falls apart".
In recognition of this, for all their cheapness and nastyness, these wall warts are going to be one or more of UL, CSA, GS, VDE, CE, etc., approved and marked, lest they be illegal to sell to the general public. That does not mean that they will last any longer than the warranty plus one day, but it does mean that if one of these guys should catch fire or fail violently in other ways, there may be some recourse from the manufacturer, and your insurance company has one less reason to withhold damage payments. Perhaps more important is that the warranty on the connected equipment isn't voided because a nonstandard power supply was used. Sometimes this can be nice, or it may be irrelevant, YMMV.
Beyond all this, the fact remains that if this common power-supply fails, all of the devices powered from it will stop working, temporarily or permanently depending on the nature of the failure... but with individual power warts, only the device feeding off the failed unit will be in jeopardy. Then there is the matter of moving things around and having to lay fat low-voltage cables following, and having to worry about power quality issues and fire hazards (a PSU delivering 12 V and 25 A DC can sustain some impressive arcs!) All of this instead of just plugging it into the everpresent mains and just have it work?
I hardly ever need to take 3rds, 4ths or 6th any more often than 5ths or 7ths or 13ths... or any other such accurate fractions of any numbers at all. As far as I am concerned, if we absolutely must change numeric base at all, base 16 would be better. We will still have 10 fingers no matter what, and dividing a pie has always been easiest by subsequent halving.
For the purpose of head reckoning, I find it vastly more useful to remember the decimal periods of the simple fractions for denominators up to about 16, and then further approximations for such things as a year being pi * 10^7 seconds, pi^2 is approximately 10, yet 100/pi = 32; the square root of 2 in terms of the fractions 1/7, 1/14 and their expansions, and so forth.
Similarly, the metric system carries with it the useful and expandable realizations such as a ton of snow will eventually melt into a cubic meter of water, or that a 1mm rainfall means that there is one liter of water per square meter of land. These prove useful and easy not just to remember, but to derive and expand on.
Did you read TFA at all? The interfacing is through GPIB to the converter units, two to read the temperatures and one to control the power to the heaters. He evidently has got his GPIB card to work with Turbo C under DOS, and a similar device driver can be made or obtained for Linux, I'm quite certain.
However, this whole thing is a proof-of-concept as it stands. Professional grade equipment, which is what is being used here, even under Linux, would be way too expensive for the average coffee-lover.
I think he's done an excellent hack, and from how I understand TFA, I guess the next step would be refining and integrating the various units.
I find it odd that a company can charge thousands of dollars for commercial software licenses and not be accused of coercion, but people will talk about the GPL "forcing" developers to do things.
The requirement to pay a specific amount can be easily quantified and accounted for as "operating expenses" or whatever. Businesses are used to being on either side of the table for this kind of bargain, it is like payment for services rendered.
The requirement that instead of a payment, additional code must be revealed, is almost impossible to quantify. It can be anything, from inconsequential to a matter of either giving competitors the keys to the kingdom or not have any saleable product anymore.
This perceived risk is what scares a number of existing "closed-source" people. Though at some point it may make sense to open the code and live off of support, but until then, GPL'd libraries are bad news.
And the GPL FAQ clearly spells out what is meant by "linking" that causes this latter, difficult, requirement. This does include attaching to and calling functions in a shared library, and that is where the requirement may come up. Evidently, the GPL and non-GPL code has to run as separate processes for them to be sufficiently decoupled according to this.
While this may make sense for interoperativity between different applications, this is not universally useful for libraries that by design are meant to be linked as part of a larger system.
Does your firewall block UDP ports at all? Many of the "ready-to-run" ones only care about TCP ports, which is what your port 80 for a webserver would be. Blocking TCP ports does nothing for other protocols such as GRE (used with some VPN's) or UDP (which, like TCP, also has a set of numbered ports).
VoIP uses UDP, usually port numbers in the vicinity of 5060, and some units may have a way of moving away from these to other UDP ports that are not blocked.
The scary parts are the error messages from that file (drivers/net/sunhme.c) rather than the comments. Unlike the comments, they can be revealed to the word at large, and, since they are error messages, they will appear in a context of "my bloody network card has stopped and I'm only getting daft jokes about it!", which is probably nowhere near as funny as what was might have been intended. Of course, I have no idea about how likely this is to happen; it could be that this is a rare or obsolete unit whose inclusion might have been mandated by some PHB...
Some of the funny printk() calls can at least be called informative:
printk(KERN_ERR "happy meal: Aieee, transceiver MIF read bolixed\n");
printk(KERN_ERR "happy meal: Aieee, transceiver MIF write bolixed\n");
printk(KERN_ERR "%s: Aieee, link timer is asleep but we got one anyways!\n");
Then there are the rather less informative:
printk(KERN_ERR "happy meal: Transceiver BigMac ATTACK!");
printk(KERN_ERR "happy meal: Receiver BigMac ATTACK!");
printk(KERN_ERR "happy meal: Fry guys.");
printk(KERN_ERR "happy meal: Transceiver and a coke please.");
printk(KERN_ERR "happy meal: Eieee, rx config register gets greasy fries.\n");
printk(KERN_ERR "happymeal: Would you like that for here or to go?\n");
The author's opinion of the Sun hardware aside, it stops just short of telling off the end-user for using such hardware.
What has died here over the years. Most of this has been "natural death" rather than provoked.
1987- XT Power supply, 2 years old, its replacement still works.
1992- a 4 years old ST4144 120M HDD started producing bad sectors run in RLL, the same unit is still operating properly using MFM, yielding 80MB.
1994- a 1 year old Conner hard disk started losing data to bad sectors.
1996- Video card, hercules-compatible from the late 80s sometime stopped working
1997- Hard Disk, (ST-225) 10 years old; replacement still works, though it is was purchased used and is probably about as old.
2000- 10 years old hard disk in a portable computer gave out.
2001- A 7 year old multi-serial card quit. Later this had been found to be caused by a bad power supply.
2002- A 13 years old MVME320 controller/Miniscribe 70 MB assembly has stopped working. Could be either one, replacing the disk with a known good one didn't help, and I haven't been able to tell if the disk itself is bad.
2003- An 8 year old HP disk failed utterly.
2003- A 6 year old Astec power supply failed in one machine.
2004- An 8 year old Astec power supply failed
2004- A 5 year old Compaq power supply failed
2004- A 1 year old Maxtor disk seized. Still under warranty, I have it off to Maxtor for possible replacement.
2004- The hard disk in a 2-year old DELL portable started failing when heated. Replaced; the disk still works OK when plugged into another machine. There are heating problems in that machine...
2005- A 17 years old HP 300MB disk has started to produce bad sectors, in a swap area, so the disk is still readable, but probably going. Another disk of the same age in the same enclosure is still working fine.
2005- A 2 year old Maxtor disk stopped, and it let out its magic smoke...
As can be seen from the above, the main failures are disks and power supplies. Not really surprising, and some of this is fairly old equipment.
PCs are non-standard since all RS232 equipment should have female connectors.
No. According to that standard, the Data Terminal Equipment (computers and terminals) should have male connectors and the Data Communications Equipment, which basically means all kinds of modems, should have female
connectors. However, many other manufacturers put female connectors on their equipment, for whatever reasons. Perhaps they considered them DCEs since these cables usually would have dumb terminal DTEs (with male connectors) at their far ends, and thus all cables could be made male-female straight-through, making them as simple and versatile as mains extension cords. Or maybe because the female connectors were slightly less fragile. Some old PC cards even had jumpers to change their configuration from the default DTE to DCE, just to confuse the matter further.
I think PCs went male to avoid confussion with the paralell port.
Very likely. The paralell port has TTL-voltage levels, 0V to 5V, whereas the serial port uses signals that can extend as far as +15 and -15V. The TTL chips sitting behind the paralell port would not stand these larger voltages, and everyone would have had a lot more hardware failures on their parallell port cards from plugging in serial-cables into them. Not everyone used more rugged chips than the 74LS240-series of the originals.
BTW DTE/DCE makes sense when you look at what it was designed for - connecting a dumb terminal to a dumb modem.
And connecting terminals to modems was the original scope of this standard. The naming of the serial-port devices on various OSes as COMX or/dev/ttyX or/dev/cuaX also reflects this old standard usage. There were teletypes, sometimes behind a pair of modems (tty), call-out connections (cua) or just some unspecified communications kit (COM) expected to be found at the other ends of those cables.
It just so happened that the serial protocol with its relatively few wires, also seemed a good choice for connecting other kinds of equipment, notably printers and plotters. After all, an old-fashioned teletype was really a kind of printer. The problem was that not all the handshake and status lines for a modem would be relevant for the printer, and everyone came up with their own idea of which ones were to be used and which ones weren't.
It isn't so much the standard, as its reappropriation outside its scope. It would have been a lot easier if there had been more guidance amongst the choices and options (male/female connectors, all the status and handshake lines: most of the 25 pins in the connector were actually assigned to some more or less obscure function)
True enough, there are plenty of http-based systems that allow only approved customers to download data or code.
However, that would also imply some kind of record-keeping on behalf of the code's owner, that they would need to record who did download what, and when, in order to issue correct bills for that service, or at least have some kind of idea as to how popular their software is.
To me, an empty dialog-box which lets anyone and everyone past doesn't seem to be much different than anonymous FTP where you have to type in anything as either username or password.
And arguing that this is a "bug" that IBM presumably knew about -- well if that is what it was, anyone else curious coming by would soon know about this "bug" as well. Including SCO themselves, and the onus would be on them to fix it if they perceived it as bad enough. Or leave it in place, but then they cannot use it as grounds for complaints afterwards!
I'll be looking forward to IBMs reply on this, that should be an interesting read.
Now, the purpose of setting up a http server is to distribute some kind of information to the world at large. And maybe accept some information, like Slashdot and a lot of other sites do.
Similarly, if someone sets up an anonymous ftp server they would also be perceived as doing this in order to distribute and maybe also receive information, to and from the world at large. Same thing really.
Now since SCO did just that, how can they then expect to be able to come afterwards and say that IBM shouldn't have looked at their site and downloaded the stuff they had to offer?
Makes no sense to me. One would expect a minimum of "due diligence", such as maybe using a locked-down ftp server with access to only authorized users, if their information was not to be made public and available to world+dog..
But what SCO is on about looks to me like posting a notice with tear-off tabs on a wall somewhere public, where everyone and anyone go by, and then claim some kind of infringement ("unclean hands") from certain people reading this posted text and tearing off a tab.
In fact, that passive OR gate in the first figure is rather similar in function, though not in construction, to these LEGO gates of the article, since they are also passive, and that the OR function is the simplest one. And since they are passive there is a limit to how many stages can be connected before the signal is degraded and lost.
Otherwise, I thought it more worthwhile to talk about the kinds of gates actually being useful in arbitrary large systems. As for whether NAND or NOR is the preferred gate, it depends on the logic family more than anything else, since everything can be made out of either all NAND (as for TTL) or all NOR (as for ECL) gates.
For these LEGO gates, it seems that the OR is the simplest function, though internal inversions only require adding an extra spur gear, so NOR, NAND and AND become only slightly more expensive, and all of these become less expensive than the standalone inverter.
This suggests using a mix of the inverted and non-inverted inputs of the basic OR gate as required so as to save on standalone inverters.
All it would need in addition for arbitrary large systems would be some means of reconstituting the signals, that is, some kind of active device.
I'd say that depends on the particular kind of logic circuit. DTL and original TTL were natural for NAND-functions, just add another emitter in the input stage for TTL, or equivalently, another diode for DTL to add more inputs. A one-input gate would be the NOT gate. NOR-functions would be achieved by parallelling transistors in the second
stage of the gate, hence adding another input then meant adding an another entire input stage also. This would be more space-consuming an thus more expensive.
For ECL, the situation is the opposite, since here everything is mostly transistor stages connected in parallel, as well as generally available true and complementary outputs, so the NOR or OR functions are the most common. Add another transistor in parallel for each additional input, and outputs can be tied together in some cases, forming more OR-functionality. With inversions, the necessary AND and NAND functions may be generated as per DeMorgans theorem.
The situation for NMOS, PMOS and RTL are similar to the one for ECL: transistors in parallel for the basic NOR function are generally preferred to transistors in series for the NAND function.
In CMOS circuits, NAND and NOR are about the same in complexity, it is a matter of parallel-connecting the P-channel transistors and series-connecting the N-channel transistors for a NAND function, and vice versa for the NOR function.
You'd be amazed how soon this could catch on. It is already commonplace with trains, where signalling units in the track are sending messages to the controllers in the locomotive about such things as speed limits, distance to signals, states of signals and so on.
And the railway people tend to care a lot about the safety of life and limb. Of course their advantage is that they have closed courses and professional trained drivers. I'd imagine something like this could begin to be used in dedicated bus lanes and suchlike.
When the one-year old is approaching two, he or she will no longer be limited in altitude. Not as long as there are any moveable chairs or anything else around that they can attempt to climb up on. And what they cannot climb up to reach they will, sometimes successfully, attempt to pull down instead.
Only thing I can think of that might work is something along the lines of a floor-level locked cabinet containing the PC, the keys to which are in your pocket.
Slashdot Mirror
1. ginormous is from gigantic and enormous
2. confuzzled from confused and puzzled.
4. chillax from chill out and relax
6. gription from grip and friction
7. phonecrastinate from phone and procrastinate.
8. slickery from slick and slippery
9. snirt from snow and dirt
10. lingweenie from linguist and weenie
For several of these, Nos. 1, 2, 4, 6, and 8, the meaning remains close to meaning of the original component words, as these already are synonyms of each other and the resulting portmanteau word. Nos. 7, 9, and 10 produce special cases of one of the words. Furthermore, except for number 9, which may provide a useful description of an unpleasant condition, most of these would likely be considered insufferable cute-ish and contrived by anyone overexposed to them.
The "5. cognitive displaysia" seems to be derived from some kind of medical technical term, to make it seem more serious than it is. It still seems to be a nerdy way of saying that I'm bothered by forgetting which stuff I might have forgot before I leave, or something... Perhaps the more succinct term isn't such a waste after all.
This leaves me with the conclusion that the word "woot" is the best of the lot, as it is the more rare, non-portmanteau new-word, that also actually has a distinct meaning.
Of course it is difficult to envisage this, since Groklaw isn't about trying to sway any court case. The lawyers for IBM and SCOX are the ones to make that kind of impact, and they have to rely on whatever are the facts. Groklaw is merely reporting the various twists and turns of this case, and in this process it is helping the lawyers and the technical people understand what the other say and how they think. As far as any conjecture being presented there, this appears mostly based on available material, such as court filings and technical documentation generally available. There may seem to be a bias against SCO, but if the realities of the case had been different, there could just as well have been an apparent bias against IBM.
As for being valuable to SCO PR, I'd beg to differ! Groklaw has been able to neutralize much of the FUD that Darl McBride would have us believe about SCO owning and controlling this and that... if anything can be compared to ammunition, it would be the torpedoes that are about to sink SCO. Even without Groklaw, SCO would eventually have foundered, it would have taken longer, and resulted in a slower growth of Linux.
Though not really questioning Tetschner and his travelling companion's sex or girth, it seems that the pool version with "sinks" really was most applicable here. :)
Probably right about the virus-scan. Outside the machine, the drive probably will look like it is full of garbage.
However, I don't think replacement will become impossible. If the machines won't allow replacement disks, this means that a disk failure will result in a useless machine; this will probably also get in the way of people wanting to add disks -- and the people wanting to put Linux on a second-hand machine will cry foul -- so this is going to fly as well as those boat-anchors those machines would become.
And this iteration of Longhorn at least will not require these chips... you won't have to buy new motherboards just now. But, perhaps further down the line this may become a required peripheral for Longhorn, but this will not be until most motherboards have it in place.
It looks like mostly a way of keeping stuff on hard-drives secret. As such this is not so bad in view of how frequent notebook-theft is, or how big the security problems of second-hand equipment are.
One day in the RF and Microwaves lab, we wanted to listen to some radio programme. Now, there were several kinds of receivers lying around, but most of them were in the higher frequency ranges.
Except for the HP140 series spectrum analyzers. These had a range going to 110 MHz, just right for FM broadcast radio.
So we made a simple antenna from a pair of wires that we stuck onto one of the windows, and connected this to the input of one of the HP-140 series spectrum analyzers. In addition to the convenient range, they had all kinds of nice filtering functions to limit bandwidth to an FM radio channel. We could even see the various broadcasting stations on the display.
These analyzers had a vertical output, to which we hooked up a linear power amplifier that originally was designed for driving a tele-coil system for the benefit of the hard of hearing. We hooked up this to a speaker that originally had been in someone's car but became surplus when they got a new stereo system.
By stopping the sweep and tuning the spectrum analyzer a little to the side of the frequency of interest, we got slope detection of the radio signal, and we got the sound of the station of interest.
Of course, the sound quality wasn't the best, obviously not stereo; and we noticed that the spectrum analyzer would slowly drift into and out of tune with the temperature variations of the day, so frequent readjustments were necessary. Fading was also noticed, but this wasn't too bad. Still, this set-up remained in use for quite a while, something around a year.
So this is how to make a set of $~10^4 equipment sound much like any old $~1 radio as found at flea markets, garage sales, or thrift stores...
Same sound for a 40 dB increase in price...
Now with fonts being executable code, (some kind of shared library or DLL) running in the process of the word processor or printer formatting program, wouldn't the GPL's conditions for close linkage also apply to these programs? The document is more of a carrier; though the presence of font-statements could be construed as invocations of the code in the font-libraries, and thus the document would become a kind of source code itself.
By publishing the source code of the fonts along with the document, this GPL requirement would seem to be satisfied. However, the GPL may still clash with other restrictions that the document author may want to impose, for example, if the document describes something that must be kept confidential, this will conflict with the GPL requirement of no additional restriction on distribution. Even making the fonts LGPL wouldn't get around this one.
However, the purpose of documents really is to disseminate information that in itself is completely irrelevant to the word processor or printer, not unlike the comments of source code or shell scripts. But unlike these, the execution of font code is incidential to this main purpose, and shouldn't get in the way.
Perhaps the situation could be resolved by looking at the analogy of an interpreter. The word processor would be the interpreter program, the fonts would correspond to specialty libraries, and the document would be the script to be interpreted. Let us presume the easiest case of the interpreter being closed-source, and the libraries being LGPL so there is no a priori conflict there. What will be the possible licences of the script?
"IP" is an unfortunate collective term for the distinct concepts of "copyrights", "patents", "trademarks", and "trade secrets", all of which can be ascribed a monetary value, and thus qualify as "property" in an accountant's ledger. FUDsters love to bandy this term about, as it is nice and short and succinct, and completely inaccurate. To that extent, the parent poster is right: "IP" as a term ought not to exist from a utilitarian point of view.
Unfortunately it still does.
In the case of the GPL, the relevant concept is first of all "copyrights", on which underlying law the GPL itself relies. The whole idea of the GPL is to give a limited permission to copy and redistribute beyond what copyright law permits. The owner of that copyright does not lose it.
However, the GPL is basically not compatible with "trade secrets", and I suspect that here is what rubs Mr Schwartz the wrong way. However, even here there is a straighforward and clear direction to take: Avoid mixing GPLd code and code that is considered a "trade secret", lest the competitive advance due to this "trade secret" become lost. That is all there is to it. That of course also means not being able to exploit the advantages that using GPLd code can provide.
It comes down to a comparison between the value of these advantages vs the advantage of keeping the "trade secret" a secret.
It seems a fair deal.
As for blockin pr0n, who cares. It would really be useful if this could block spam!
Still, problem is, who gets to sit and determine what is immoral and what isn't? And for that matter, whose morals are we talking about here, is it the FCC, the WWE, the people behind goatse, the Republican Right, or maybe some Islamic Council for the Promotion of Virtue and Prevention of Vice? And who will pay them? Again, remember the system's ability to block any obstacles including taxation; the ARPAnet people back in the day thought in terms of nuclear bombs not nuclear family value enforcements...
Then again, considering the date of issue of this RFC, it is all in the +1 Funny realm anyway, which as we know, is karma neutral.
I rest my case.
Although 5 GPIO lines and an RS232-style serial connection does allow for connecting other interesting kit, what would have been really nice here, would have been a larger number of such general digital IO lines, and perhaps some additional peripherals, such as a timer/counter, ADC, and DAC.
One thing that might happen, however, is enforced free licensing of the MS patents that, according to the article, MS claim would be infringed otherwise. Along with forced disclosure of the interfaces, and probably some guarantees that this openness would be preserved in future revisions.
The other DC-distribution system design exercises remain impractical for a couple other reasons that I can easily pull off the tip of my tongue.
First reason: Voltage, and power variations. Around here I have various kit with power-supplies rated for voltages from 3.7V DC, 5 V DC, 7.5V DC, 2x8V AC (16 V center-tapped), 9V AC, 12 V AC, 12 V DC, 19 V DC and 21 V DC. I would suspect most other installations have a similar variety.
OK, so the various DC voltages could be hacked using a variety of linear or switching regulators off an ex-equipment PSU, though the 19 V and 21 V might present an educational challenge. Besides the fact that that this is likely to be at least as unsightly as the handful of wall-warts it replaces, this does absolutely nothing for the AC supplies, which wall-warts are simple transformers and rather efficient anyways. (the dual 8V unit has its own power cord and isn't even a wall-wart). Implementing DC-AC conversion for these purposes while regular mains AC is right there seems to be way overkill.
The second reason is reliability, as in "everything put together sooner or later falls apart".
In recognition of this, for all their cheapness and nastyness, these wall warts are going to be one or more of UL, CSA, GS, VDE, CE, etc., approved and marked, lest they be illegal to sell to the general public. That does not mean that they will last any longer than the warranty plus one day, but it does mean that if one of these guys should catch fire or fail violently in other ways, there may be some recourse from the manufacturer, and your insurance company has one less reason to withhold damage payments. Perhaps more important is that the warranty on the connected equipment isn't voided because a nonstandard power supply was used. Sometimes this can be nice, or it may be irrelevant, YMMV.
Beyond all this, the fact remains that if this common power-supply fails, all of the devices powered from it will stop working, temporarily or permanently depending on the nature of the failure... but with individual power warts, only the device feeding off the failed unit will be in jeopardy. Then there is the matter of moving things around and having to lay fat low-voltage cables following, and having to worry about power quality issues and fire hazards (a PSU delivering 12 V and 25 A DC can sustain some impressive arcs!) All of this instead of just plugging it into the everpresent mains and just have it work?
No way, I'd use time on more fun things.
For the purpose of head reckoning, I find it vastly more useful to remember the decimal periods of the simple fractions for denominators up to about 16, and then further approximations for such things as a year being pi * 10^7 seconds, pi^2 is approximately 10, yet 100/pi = 32; the square root of 2 in terms of the fractions 1/7, 1/14 and their expansions, and so forth.
Similarly, the metric system carries with it the useful and expandable realizations such as a ton of snow will eventually melt into a cubic meter of water, or that a 1mm rainfall means that there is one liter of water per square meter of land. These prove useful and easy not just to remember, but to derive and expand on.
However, this whole thing is a proof-of-concept as it stands. Professional grade equipment, which is what is being used here, even under Linux, would be way too expensive for the average coffee-lover.
I think he's done an excellent hack, and from how I understand TFA, I guess the next step would be refining and integrating the various units.
The requirement to pay a specific amount can be easily quantified and accounted for as "operating expenses" or whatever. Businesses are used to being on either side of the table for this kind of bargain, it is like payment for services rendered.
The requirement that instead of a payment, additional code must be revealed, is almost impossible to quantify. It can be anything, from inconsequential to a matter of either giving competitors the keys to the kingdom or not have any saleable product anymore.
This perceived risk is what scares a number of existing "closed-source" people. Though at some point it may make sense to open the code and live off of support, but until then, GPL'd libraries are bad news.
And the GPL FAQ clearly spells out what is meant by "linking" that causes this latter, difficult, requirement. This does include attaching to and calling functions in a shared library, and that is where the requirement may come up. Evidently, the GPL and non-GPL code has to run as separate processes for them to be sufficiently decoupled according to this.
While this may make sense for interoperativity between different applications, this is not universally useful for libraries that by design are meant to be linked as part of a larger system.
VoIP uses UDP, usually port numbers in the vicinity of 5060, and some units may have a way of moving away from these to other UDP ports that are not blocked.
Some of the funny printk() calls can at least be called informative:
printk(KERN_ERR "happy meal: Aieee, transceiver MIF read bolixed\n");
printk(KERN_ERR "happy meal: Aieee, transceiver MIF write bolixed\n");
printk(KERN_ERR "%s: Aieee, link timer is asleep but we got one anyways!\n");
Then there are the rather less informative:
printk(KERN_ERR "happy meal: Transceiver BigMac ATTACK!");
printk(KERN_ERR "happy meal: Receiver BigMac ATTACK!");
printk(KERN_ERR "happy meal: Fry guys.");
printk(KERN_ERR "happy meal: Transceiver and a coke please.");
printk(KERN_ERR "happy meal: Eieee, rx config register gets greasy fries.\n");
printk(KERN_ERR "happymeal: Would you like that for here or to go?\n");
The author's opinion of the Sun hardware aside, it stops just short of telling off the end-user for using such hardware.
1987- XT Power supply, 2 years old, its replacement still works.
1992- a 4 years old ST4144 120M HDD started producing bad sectors run in RLL, the same unit is still operating properly using MFM, yielding 80MB.
1994- a 1 year old Conner hard disk started losing data to bad sectors.
1996- Video card, hercules-compatible from the late 80s sometime stopped working
1997- Hard Disk, (ST-225) 10 years old; replacement still works, though it is was purchased used and is probably about as old.
2000- 10 years old hard disk in a portable computer gave out.
2001- A 7 year old multi-serial card quit. Later this had been found to be caused by a bad power supply.
2002- A 13 years old MVME320 controller/Miniscribe 70 MB assembly has stopped working. Could be either one, replacing the disk with a known good one didn't help, and I haven't been able to tell if the disk itself is bad. 2003- An 8 year old HP disk failed utterly.
2003- A 6 year old Astec power supply failed in one machine.
2004- An 8 year old Astec power supply failed
2004- A 5 year old Compaq power supply failed
2004- A 1 year old Maxtor disk seized. Still under warranty, I have it off to Maxtor for possible replacement.
2004- The hard disk in a 2-year old DELL portable started failing when heated. Replaced; the disk still works OK when plugged into another machine. There are heating problems in that machine...
2005- A 17 years old HP 300MB disk has started to produce bad sectors, in a swap area, so the disk is still readable, but probably going. Another disk of the same age in the same enclosure is still working fine.
2005- A 2 year old Maxtor disk stopped, and it let out its magic smoke...
As can be seen from the above, the main failures are disks and power supplies. Not really surprising, and some of this is fairly old equipment.
No. According to that standard, the Data Terminal Equipment (computers and terminals) should have male connectors and the Data Communications Equipment, which basically means all kinds of modems, should have female connectors. However, many other manufacturers put female connectors on their equipment, for whatever reasons. Perhaps they considered them DCEs since these cables usually would have dumb terminal DTEs (with male connectors) at their far ends, and thus all cables could be made male-female straight-through, making them as simple and versatile as mains extension cords. Or maybe because the female connectors were slightly less fragile. Some old PC cards even had jumpers to change their configuration from the default DTE to DCE, just to confuse the matter further.
I think PCs went male to avoid confussion with the paralell port.
Very likely. The paralell port has TTL-voltage levels, 0V to 5V, whereas the serial port uses signals that can extend as far as +15 and -15V. The TTL chips sitting behind the paralell port would not stand these larger voltages, and everyone would have had a lot more hardware failures on their parallell port cards from plugging in serial-cables into them. Not everyone used more rugged chips than the 74LS240-series of the originals.
BTW DTE/DCE makes sense when you look at what it was designed for - connecting a dumb terminal to a dumb modem.
And connecting terminals to modems was the original scope of this standard. The naming of the serial-port devices on various OSes as COMX or /dev/ttyX or /dev/cuaX also reflects this old standard usage. There were teletypes, sometimes behind a pair of modems (tty), call-out connections (cua) or just some unspecified communications kit (COM) expected to be found at the other ends of those cables.
It just so happened that the serial protocol with its relatively few wires, also seemed a good choice for connecting other kinds of equipment, notably printers and plotters. After all, an old-fashioned teletype was really a kind of printer. The problem was that not all the handshake and status lines for a modem would be relevant for the printer, and everyone came up with their own idea of which ones were to be used and which ones weren't.
It isn't so much the standard, as its reappropriation outside its scope. It would have been a lot easier if there had been more guidance amongst the choices and options (male/female connectors, all the status and handshake lines: most of the 25 pins in the connector were actually assigned to some more or less obscure function)
"wee" simply means "small" in Scots, and a "dram" would refer to something to drink, like a shot of whisky or similar.
However, that would also imply some kind of record-keeping on behalf of the code's owner, that they would need to record who did download what, and when, in order to issue correct bills for that service, or at least have some kind of idea as to how popular their software is.
To me, an empty dialog-box which lets anyone and everyone past doesn't seem to be much different than anonymous FTP where you have to type in anything as either username or password.
And arguing that this is a "bug" that IBM presumably knew about -- well if that is what it was, anyone else curious coming by would soon know about this "bug" as well. Including SCO themselves, and the onus would be on them to fix it if they perceived it as bad enough. Or leave it in place, but then they cannot use it as grounds for complaints afterwards!
I'll be looking forward to IBMs reply on this, that should be an interesting read.
Now, the purpose of setting up a http server is to distribute some kind of information to the world at large. And maybe accept some information, like Slashdot and a lot of other sites do.
Similarly, if someone sets up an anonymous ftp server they would also be perceived as doing this in order to distribute and maybe also receive information, to and from the world at large. Same thing really.
Now since SCO did just that, how can they then expect to be able to come afterwards and say that IBM shouldn't have looked at their site and downloaded the stuff they had to offer?
Makes no sense to me. One would expect a minimum of "due diligence", such as maybe using a locked-down ftp server with access to only authorized users, if their information was not to be made public and available to world+dog..
But what SCO is on about looks to me like posting a notice with tear-off tabs on a wall somewhere public, where everyone and anyone go by, and then claim some kind of infringement ("unclean hands") from certain people reading this posted text and tearing off a tab.
IANAL, YMMV etc...
Otherwise, I thought it more worthwhile to talk about the kinds of gates actually being useful in arbitrary large systems. As for whether NAND or NOR is the preferred gate, it depends on the logic family more than anything else, since everything can be made out of either all NAND (as for TTL) or all NOR (as for ECL) gates.
For these LEGO gates, it seems that the OR is the simplest function, though internal inversions only require adding an extra spur gear, so NOR, NAND and AND become only slightly more expensive, and all of these become less expensive than the standalone inverter.
This suggests using a mix of the inverted and non-inverted inputs of the basic OR gate as required so as to save on standalone inverters.
All it would need in addition for arbitrary large systems would be some means of reconstituting the signals, that is, some kind of active device.
For ECL, the situation is the opposite, since here everything is mostly transistor stages connected in parallel, as well as generally available true and complementary outputs, so the NOR or OR functions are the most common. Add another transistor in parallel for each additional input, and outputs can be tied together in some cases, forming more OR-functionality. With inversions, the necessary AND and NAND functions may be generated as per DeMorgans theorem.
The situation for NMOS, PMOS and RTL are similar to the one for ECL: transistors in parallel for the basic NOR function are generally preferred to transistors in series for the NAND function.
In CMOS circuits, NAND and NOR are about the same in complexity, it is a matter of parallel-connecting the P-channel transistors and series-connecting the N-channel transistors for a NAND function, and vice versa for the NOR function.
Here is some information about the internal connections of RTL, DTL, TTL, ECL and CMOS circuits.
And the railway people tend to care a lot about the safety of life and limb. Of course their advantage is that they have closed courses and professional trained drivers. I'd imagine something like this could begin to be used in dedicated bus lanes and suchlike.
Only thing I can think of that might work is something along the lines of a floor-level locked cabinet containing the PC, the keys to which are in your pocket.