was trying to exchange a HDD I'd bought that didn't damn well work.
But no; they insisted on testing the damn thing first on this testbed hardware they had, and then claimed it worked fine for them.
Trying to explain that yes, I knew what I was doing, I was a professional UNIX sysadmin and did this kind of thing for a living, didn't help one bit.
I think half of the problem was that the particular Fry's employee was the kind of woman employee who is the biggest pain in the ass: so paranoid about men condescending to her 'cause she's a woman that she wouldn't even listen.
Eventually I asked for a manager and explained to him that there were two possibilities. Either:
The drive was fine, and I was full of it. In which case, swap me the drives to make me happy and send me on my way, and resell the damn thing.
The drive was faulty, in which case I deserved a replacement.
It's not possible to develop any closed source commercial software for KDE without reimplementing everything. Therefore all the closed source applications will be Gnome.
Will this make a difference? I think so; many of us especially in commercial environments have to use at least one binary-only tool (even if we wish we could get rid of it)
Being a (twenty-six year old) Goth, I'd disagree with that. I personally have never dressed to piss anyone off. Neither has my wife (also Gothic) nor most others I know. Some, I'm sure, do, but even then you'll generally find that pissing off someone was just an added bonus of a look they would have liked even so.
There are, as in any group of people, some real losers in the Goth scene, and again as normal, they're often the most noticeable. Think of the way outsiders view us slashdotters, for example.
IMHO, qmail is pretty much as complicated as sendmail -- just the complexity is expressed differently. After all, the things people want to be able to do with their email are complex.
And yes, D.J. Bernstein writes code that is... not like anybody else's on the planet. Fundamentally, he doesn't *trust* anybody else's code, so he reinvents everything. While a lot of it is documented, DJB suffers from an extreme case of 'intellectual arrogance' -- qmail is not designed to be easy to understand (internally) if you're not as smart as DJB. Plus, too many things are 'obvious' to him that aren't necessarily to other people, so he doesn't explain them.
An example: his instructions say to use 'csh' to run qmail, to call the startup script by doing 'csh -cf/var/qmail/rc'. No explanation is given. However, if you *don't* take his advice, you're in trouble. qmail doesn't ignore SIGHUP. Children of csh are protected from SIGHUP by default; children of other shells aren't. You're supposed to realise this without needing an explanation.
Qmail might seem to spawn a new process for everything, but it doesn't spawn as many as you think. Instead, it uses what some call 'Bernstein chaining' after its creator; each program does its thing and then exec()s its successor, reincarnating itself in a succession of roles in a single process.
There's still a lot of processes forked (one per smtp connection, etc) but forking tiny processes isn't as expensive as you might think it is. Forking big processes is the killer. It's not a multiple-processes-per-email deal, though -- more like one process per email. That's not really any more than sendmail does.
You're completely wrong on qmail's performance, however. It kicks the ass of sendmail performance wise, just like most of the other sendmail replacements. It's sendmail that doesn't scale, not qmail. Some of the biggest email servers in the world use qmail; Hotmail uses qmail for mail sending, for example (though not receiving, AFAIR).
However, the RIAA is convinced radio stations don't pay enough. Anyone remember the rules for online 'radio' they are pushing through? Way more restrictive on what you can play, and costs more too. RIAA people are quoted as saying, effectively, 'We let radio get too sweet a deal, and we're not going to repeat that mistake'. Radio's argument was always 'We shouldn't pay that much, after all we do you a service by advertising your product'.
Medical & biotech fields don't have so much patent nonsense for three reasons.
One, that you have to be a big player in order to discover anything new anyway; the barrier to entry is high.
Two, that being SCIENTIFIC fields, *everything* is published. After the patent application has been filed, of course.
Three, patent examiners know how to research medical claims.
Contrast this with software. All you need to come up with 'new' inventions is a brain. Therefore there are LOTS of players, and not necessarily tied to large corporations.
Secondly, very little computing innovation is published. Sure, CS professors publish all the time, but commercial developers, or open source developers? Rarely. People only tend to publish in academic journals etc. if they're in academia or a *major* research lab and the discovery is on the main line of their research interests -- stuff created on the side doesn't get published. So it's very hard to research prior art.
Thirdly, patent examiners seem to be woefully under-skilled at evaluating software claims. They are, in effect, passing off the evaluation requirement onto the courts.
Patriot was NOT part of SDI. It was designed and always was designed to be an anti-aircraft system. It was never designed to be part of a strategic missile defense.
Yes, a missile destroyed in mid-air is going to safely contain the plutonium. These things do not arm until they are quite a ways into their flight. They are designed to be VERY tough and hard to damage. After all, if they weren't the chances of a dangerous accident would be very high. The designers of these things are paranoid, and with good reason.
SDI had problems, many problems, but not these.
And of course everyone realised that terrorist nuclear bombs were a threat.
Don't make the mistake of thinking that the people behind things like this are that stupid. Their overall priorities might not agree with yours and you might think their high-level decisions suck, but they're not dumb people.
Each nation was allowed to protect ONE target. The USSR protected Moscow (probably still does, in fact, though its readiness might be suspect) and the USA protected a missile base. Yes, both nations DID have ABM technology then, though it was pretty short range, and mostly intended to prevent direct hits, not fallout damage etc. The US system was subsequently dismantled, I recall.
Thanks to Motif, or to Netscape's programmers?
on
LSB: A position paper
·
· Score: 1
Netscape manages to produce a decent interface using Motif. All this proves is that with sufficient work you CAN produce a reasonable interface with Motif; not that Motif helps.
I'm sure the Netscape folks would have some choice words to say about the quality of Motif. Certainly everyone I know of who's used Motif for serious applications hates it...
I'm not sure how much better GTK is having not used it much. Netscape's major reasoning behind going to GTK is not because it's BETTER but because it's free software and it's easier to get people to work with it.
The normal cypherpunks/cypherpunks user/passwd works on this site, like everywhere else. I've seen lots of people saying 'I couldn'd read the original article, but...' so I thought I'd post this reminder!
I've read through the description on IBM's patent server, and while I'm not a lawyer nor a patent expert, this patent seems impossibly general. It says very little despite its length; the length is more devoted to extending the generality further. Also, they're not patenting a technique, but a concept; there is very little detail in the patent as to how exactly any of this is done.
The claims are astoundingly verbose and hard to decypher even for a patent, but as far as I can tell they are:
A distributed environment comprised of multiple servers, each may have many users. To save bandwidth, information is sent only once per server rather than duplicated for each user. The communications may be multicast or point-to-point. The environment can support applications. There is a global timebase, and every event is timestamped.
Users specify what information they're interested in, and events have parameters which can be filtered on. Only information the user's interested on will be sent to them. The system can be object oriented, and objects can have types, which can be used to filter.
Events and objects in the system may have position, priority, data rate etc associated with them, and these can be used to filter on.
The priorities of multiple users on a server can be coalesced into groups for easier filtering.
Information sent to a remote server can be pre-filtered by the sender according to its knowledge of the remote server's needs, bandwidth. Also, information recieved by a server is filtered to remove unwanted info.
Servers can be grouped according to shared preferences.
The multi-user environment can be grouped into regions, where at least one server administers each region. Regions can overlap and there is the concept of 'neighboring' regions. Filtering based on this geographic knowledge. Regions may be grouped. Does this sound patentable to YOU?
A very large proportion of free software in existence was developed wholly or partly by people whose time was being paid for by taxes. A large proportion of stuff produced by people at Universities fits into this category.
The only difference is that much of this development was done as a side project, or as tools for a main project. The only difference being discussed here is making the free software the MAIN part of a grant application rather than just a subsidiary benefit.
Who employed RMS originally and paid for much of the original GNU development? MIT, a university, and therefore at least partly, government taxation.
Slashdot Mirror
was trying to exchange a HDD I'd bought that didn't damn well work.
But no; they insisted on testing the damn thing first on this testbed hardware they had, and then claimed it worked fine for them.
Trying to explain that yes, I knew what I was doing, I was a professional UNIX sysadmin and did this kind of thing for a living, didn't help one bit.
I think half of the problem was that the particular Fry's employee was the kind of woman employee who is the biggest pain in the ass: so paranoid about men condescending to her 'cause she's a woman that she wouldn't even listen.
Eventually I asked for a manager and explained to him that there were two possibilities. Either:
He saw the logic, and let me make the exchange...
No, AFAIR www.microsoft.com always was NT. Their internal mail servers, though, were UNIX (SCO Xenix?) for a LONG while.
Get the word right.
It's not possible to develop any closed source commercial software for KDE without reimplementing everything. Therefore all the closed source applications will be Gnome.
Will this make a difference? I think so; many of us especially in commercial environments have to use at least one binary-only tool (even if we wish we could get rid of it)
Being a (twenty-six year old) Goth, I'd disagree with that. I personally have never dressed to piss anyone off. Neither has my wife (also Gothic) nor most others I know. Some, I'm sure, do, but even then you'll generally find that pissing off someone was just an added bonus of a look they would have liked even so.
There are, as in any group of people, some real losers in the Goth scene, and again as normal, they're often the most noticeable. Think of the way outsiders view us slashdotters, for example.
IMHO, qmail is pretty much as complicated as sendmail -- just the complexity is expressed differently. After all, the things people want to be able to do with their email are complex.
... not like anybody else's on the planet. Fundamentally, he doesn't *trust* anybody else's code, so he reinvents everything. While a lot of it is documented, DJB suffers from an extreme case of 'intellectual arrogance' -- qmail is not designed to be easy to understand (internally) if you're not as smart as DJB. Plus, too many things are 'obvious' to him that aren't necessarily to other people, so he doesn't explain them.
/var/qmail/rc'. No explanation is given. However, if you *don't* take his advice, you're in trouble. qmail doesn't ignore SIGHUP. Children of csh are protected from SIGHUP by default; children of other shells aren't. You're supposed to realise this without needing an explanation.
And yes, D.J. Bernstein writes code that is
An example: his instructions say to use 'csh' to run qmail, to call the startup script by doing 'csh -cf
Qmail might seem to spawn a new process for everything, but it doesn't spawn as many as you think. Instead, it uses what some call 'Bernstein chaining' after its creator; each program does its thing and then exec()s its successor, reincarnating itself in a succession of roles in a single process.
There's still a lot of processes forked (one per smtp connection, etc) but forking tiny processes isn't as expensive as you might think it is. Forking big processes is the killer. It's not a multiple-processes-per-email deal, though -- more like one process per email. That's not really any more than sendmail does.
You're completely wrong on qmail's performance, however. It kicks the ass of sendmail performance wise, just like most of the other sendmail replacements. It's sendmail that doesn't scale, not qmail. Some of the biggest email servers in the world use qmail; Hotmail uses qmail for mail sending, for example (though not receiving, AFAIR).
However, the RIAA is convinced radio stations don't pay enough. Anyone remember the rules for online 'radio' they are pushing through? Way more restrictive on what you can play, and costs more too. RIAA people are quoted as saying, effectively, 'We let radio get too sweet a deal, and we're not going to repeat that mistake'. Radio's argument was always 'We shouldn't pay that much, after all we do you a service by advertising your product'.
Medical & biotech fields don't have so much patent nonsense for three reasons.
One, that you have to be a big player in order to discover anything new anyway; the barrier to entry is high.
Two, that being SCIENTIFIC fields, *everything* is published. After the patent application has been filed, of course.
Three, patent examiners know how to research medical claims.
Contrast this with software. All you need to come up with 'new' inventions is a brain. Therefore there are LOTS of players, and not necessarily tied to large corporations.
Secondly, very little computing innovation is published. Sure, CS professors publish all the time, but commercial developers, or open source developers? Rarely. People only tend to publish in academic journals etc. if they're in academia or a *major* research lab and the discovery is on the main line of their research interests -- stuff created on the side doesn't get published. So it's very hard to research prior art.
Thirdly, patent examiners seem to be woefully under-skilled at evaluating software claims. They are, in effect, passing off the evaluation requirement onto the courts.
Patriot was NOT part of SDI. It was designed and always was designed to be an anti-aircraft system. It was never designed to be part of a strategic missile defense.
Yes, a missile destroyed in mid-air is going to safely contain the plutonium. These things do not arm until they are quite a ways into their flight. They are designed to be VERY tough and hard to damage. After all, if they weren't the chances of a dangerous accident would be very high. The designers of these things are paranoid, and with good reason.
SDI had problems, many problems, but not these.
And of course everyone realised that terrorist nuclear bombs were a threat.
Don't make the mistake of thinking that the people behind things like this are that stupid. Their overall priorities might not agree with yours and you might think their high-level decisions suck, but they're not dumb people.
Each nation was allowed to protect ONE target. The USSR protected Moscow (probably still does, in fact, though its readiness might be suspect) and the USA protected a missile base. Yes, both nations DID have ABM technology then, though it was pretty short range, and mostly intended to prevent direct hits, not fallout damage etc. The US system was subsequently dismantled, I recall.
Netscape manages to produce a decent interface using Motif. All this proves is that with sufficient work you CAN produce a reasonable interface with Motif; not that Motif helps.
...
I'm sure the Netscape folks would have some choice words to say about the quality of Motif. Certainly everyone I know of who's used Motif for serious applications hates it
I'm not sure how much better GTK is having not used it much. Netscape's major reasoning behind going to GTK is not because it's BETTER but because it's free software and it's easier to get people to work with it.
Entirely possible they've been doing work for Western companies under contract, and have all the software etc. Isn't that hard to envisage.
Install it someplace else and name it 'egcs'. The linux kernel will not build with egcs right now, among other things. Read the linux kernel docs.
The normal cypherpunks/cypherpunks user/passwd works on this site, like everywhere else. I've seen lots of people saying 'I couldn'd read the original article, but ...' so I thought I'd post this reminder!
I've read through the description on IBM's patent server, and while I'm not a lawyer nor a patent expert, this patent seems impossibly general. It says very little despite its length; the length is more devoted to extending the generality further. Also, they're not patenting a technique, but a concept; there is very little detail in the patent as to how exactly any of this is done.
The claims are astoundingly verbose and hard to decypher even for a patent, but as far as I can tell they are:
A distributed environment comprised of multiple servers, each may have many users. To save bandwidth, information is sent only once per server rather than duplicated for each user. The communications may be multicast or point-to-point. The environment can support applications. There is a global timebase, and every event is timestamped.
Users specify what information they're interested in, and events have parameters which can be filtered on. Only information the user's interested on will be sent to them. The system can be object oriented, and objects can have types, which can be used to filter.
Events and objects in the system may have position, priority, data rate etc associated with them, and these can be used to filter on.
The priorities of multiple users on a server can be coalesced into groups for easier filtering.
Information sent to a remote server can be pre-filtered by the sender according to its knowledge of the remote server's needs, bandwidth. Also, information recieved by a server is filtered to remove unwanted info.
Servers can be grouped according to shared preferences.
The multi-user environment can be grouped into regions, where at least one server administers each region. Regions can overlap and there is the concept of 'neighboring' regions. Filtering based on this geographic knowledge. Regions may be grouped. Does this sound patentable to YOU?
Way cool! Here's hoping this is just the first of many.
A very large proportion of free software in existence was developed wholly or partly by people whose time was being paid for by taxes. A large proportion of stuff produced by people at Universities fits into this category.
The only difference is that much of this development was done as a side project, or as tools for a main project. The only difference being discussed here is making the free software the MAIN part of a grant application rather than just a subsidiary benefit.
Who employed RMS originally and paid for much of the original GNU development? MIT, a university, and therefore at least partly, government taxation.