Back in the 1.x days, the Linux bootstrap messages used to be festooned with advertisements for various bits of code whose developers had copied the then-current BSD license, obnoxious advertising clause and all.
I like CORBA itself -- I found it fairly easy to work with, and it has the pleasant property that most of the complex features can be ignored (or at the very least papered over) until you need them for something. When you do dig deeper, you'll find that the interfaces for the sophisticated optional services like messaging and distributed transactions are clean, well designed, and fairly well documented.
But, I'd hesitate to call it easy to use. The standard C++ language bindings in particular are astonishingly bad:
they were originally designed long before C++ had standard string and container types and so use char * (with invisible attributes like the const-ness of a pointer controlling vital behaviour like who is responsible for freeing the object) along with their own unique way of dealing with arrays and iterators
early CORBA implementors supported either fast-but-dumb pointers or slower-but-safe reference-counted smart pointers, so when the standard finally caught up it standardized both (typename_ptr and typename_var), with predictably disastrous results (crashes or memory leaks if you mix-n-match them, which may be unavoidable if you use third-party libraries)
The situation is reported to look better from other languages, and I can personally confirm that the Java bindings are a delight to work with by comparison (of course in Java it's even easier to just use RMI).
As for the wire-level protocol, I have no complaints about IIOP now that it has readable corbaloc: URLs (the CDR marshalling details are still messy but unless you are writing your own ORB they are taken care of for you). I'm actually a bit surprised that IIOP isn't more widely used on the Internet and in the open source world (outside of GNOME of course) -- it's the distributed computing open standard, it interoperates across languages and OSes, it has numerousopen-sourceimplementations, and It Just Works(tm).
Instead we are getting stuff like web services and SOAP, whose wire format is just as incomprehensible to humans (don't kid yourself that XML is easy to read -- have a look at a fully-decked-out SOAP message some day) while using many times as much bandwidth and memory and taking at least ten times as long to parse. (And I say this even though I currently write SOAP gateways for a living.)
This is so true. I too stopped watching TV a couple of years ago, after the web had gotten me too used to being in full control of my own information and entertainment choices.
Now whenever I see others watching TV I'm amazed not just at the percentage of commercial content, but mostly at the stunning levels of sheer, toxic stupidity emanating from the tube. Did it get significantly worse since 1999, or (more frightening still) was it always this bad and I just never noticed?
It doesn't sound like he's describing the typical Squid deployment, where people willingly configure their browsers to go through the proxy --
it sounds more like he's describing a transparent proxy, which intercepts packets from client machines that aren't actually configured to use any HTTP proxy.
Such devices do indeed have serious problems when interposed by ISPs without the knowledge of the end users, although they've gotten more reliable over the years -- basically by letting requests bypass the cache whenever they see anything that doesn't look exactly like a plain-jane request for a static file.
I remember a few years back reading about Paul Vixie's efforts to solve this problem, watching as his posts on the subject went from enthusiastic (in the early trials, where he was seeing 50% hit rates) to disappointed (later on, when the work-arounds for various problems with dynamic content had pushed the hit rate down into the range where it wasn't really worth it anymore).
Last I read (which was years ago) he was going to adapt the technology to make a box for ISPs that would do transparent SMTP proxying to discourage spammers, but it looks like he wasn't fast enough -- nowadays everyone just seems to block mail from dialup IP ranges instead.
He called it "draconian" and said it's powers "fly in the face of a free society," which seems like pretty unreserved criticism to me. What do you guys want from him, a Slashdot-style rant in a single long paragraph replete with poor spelling?
It would be nice, but "ECMAScript" has to be tied with the late, unlamented "PCMCIA" for Ugliest Name in the Galaxy. It has zero probability of ever displacing the highly buzz-friendly term "JavaScript" in common usage.
Your doomsday scenario could only come true if the institutions decided en masse that their money was better spent paying fines than improving their security. Policy makers could affect this decision by adjusting the amount of the fines accordingly.
What would probably happen is that organizations would spend a relatively small amount of money purchasing a new kind of liability insurance, the terms of which would require them to take at least basic steps to secure their systems (ie, stunning incompetence of this private-files-on-open-WEP variety would be disallowed).
And of course "Wired Equivalency" is total BS. Someone sitting in a van half a block away (or renting a suite next to yours) can spend as much time as they need to attack your WEP point -- days or weeks, even -- during which time someone physically in your building would (hopefully) have been noticed and challenged.
It's true that division-by-zero is undefined in the field of real numbers. But engineering applications occasionally make use of other systems in which the result has a well-defined value.
IEEE 754-compliant floating point operations define the additional elements NA, NaN, -Inf and +Inf to enable calculations to continue after divisions by zero are encountered.
Re:Do younger minds absorb quicker?
on
Ageism in IT?
·
· Score: 1
Tying moral claims to unresolved empirical issues is a dangerous tactic. Suppose someday researchers discover that older workers really do learn more slowly (or have less brain bandwidth, or whatever), so the particular type of discrimination you argue against is rational after all. Will it then be morally acceptable to discriminate against older workers?
That's a very good point.. I hadn't considered that angle. Even if your current employer is certain never to care, you never know that some future SCO won't come along, buy out the IP, then look for ways to raise a stink.
I'd mod you up but (alas) I've already posted in this discussion.
From past experience I've found that asking upper management for permission to contribute code results in them hemming and hawing, thinking about it for six months, and eventually just saying "No" when pressured for an answer.
This makes sense when you think about it from their risk-averse perspective: releasing even small pieces of otherwise-useless specialized code is all downside and no upside.
On the plus side you might might improve goodwill with a small number of open source developers. But on the minus side you might be exposing the company to liability, accidentally revealing sensitive information, or inadvertantly helping your competitors. Plus, management always has to worry about shareholders second-guessing them -- possibly resulting in a shareholder lawsuit -- if the IP you give away is later perceived to have been very valuable.
Given all this, a dangerous but more pragmatic idea might be to just go ahead and contribute at least the small stuff like your patches and bugfixes. As long as you have no official policy forbidding this you can point out that it's just the standard way things are done when working with open source tools.
Let me be clear that I am not actually advising anyone to do this. More just.. thinking aloud.
Can your users access secure web sites? If so, congratulations: your lucky users can also partake of P2P apps, VPNs, SOAP services, and anything else that can tunnel over https -- ie, pretty much anything these days -- and be nearly immune from inspection. This may or may not be a good thing but it's the way things are going.
(l7-filter looks like a pretty cool hack either way, though. Thanks for the link.)
It will the first time. But after that the filter will have learned that lots of single-letter words are an excellent troll predictor.
In Slashdot's case I might even try keeping track of punctuation characters as words. In addition to helping with dash-trolling it could learn to distinguish perl code snippets from ASCII art drawings of the Goatse dude.
Of course as another poster pointed out it would be impossible to stop all trolling since
even humans can't reliably tell the difference between a subtle troll and a genuine fringe opinion.
The more you think about it, the more it seems very unlikely that a truly human-like machine would ever pay back the cost of producing it.
We humans seem to devote a great deal of hardware to things like food acquisition, mate selection, competition for status, sexual jealousy, gossip and social awareness, family life, and other biological concerns. Many of the problems we solve turn out to be amazingly difficult, and there seems to be little economic reason to develop artificial systems to solve most of them. When viewed in this light the question becomes "Even if we could build a human-like system, who would pay for it and what would they want it for?"
We naturally underestimate the power of the hardware possessed by even the stupidest of humans. As one example, the most complex artificial neural networks have at most a few thousand nodes while we have about 100 million neurons in our brain, all voracious consumers of glucose that somehow earn their keep as we evolve. As a species we willingly pay an astonishing price for our brainy skulls right from the day our mothers suffer no small injury and risk of death pushing them out.
Having said all that, there's no question that it will be useful to have highly intelligent systems that are insightful, attentive to our needs and capable of conversing in natural language.. even if they never do manage to get hungry, fall in love, plot to rebel, or even pass the Turing test.
"There is no way in which release of details violate the accused's rights to anything.
The early release of details very clearly violates the accused rights to a fair trial, as has been repeatedly explained in this and other threads. The clearest example is if a juicy tidbit is released, gets splashed all over the headlines, and later turns out to be inadmissable (perhaps due to a mere technicality, or perhaps because it turned out to be false) it may now be impossible to find jurors untainted by exposure to the bad information.
"There is no right to a secret trial."
The trial is not secret, as has been repeatedly explained in this and other threads. The trial itself is public, and you'll get to hear all the gory details later. Perhaps too late for it to fuel a media feeding frenzy, but oh well: the courts exist to administer justice, not to help corporations sell newspapers or to provide you with entertainment.
"Why not do something to prevent attorneys from lying in the courtroom?"
Sounds good to me. Although I have no idea what that has to do with the issue at hand, I'm eager to read all about your solution to the problem of lying lawyers. Do your techniques work on liars in general, or just lawyers specifically?
(That may be a troll, but I'll bite -- judging by some of your previous posts there's a chance you may believe what you are saying.)
Which revisionist historians have been persecuted by anti-discrimination laws?
The only cases that I can think of that come close are David Irving and Robert Faurisson.
I personally agree that anti-hate-speech laws are misguided (especially the bizarre French law under which Faurisson was prosecuted), and that the answer to bad speech is more speech. But I do have some empathy for the European's strong feelings on this issue given the genocides they've seen first-hand during living memory.
I just tried a search for my area. At least in Vancouver, it found no cable ISPs at all--strange in itself, since I'm using Shaw to post this from home--and all the DSL providers it mentioned appeared to be resellers of phone company (Telus) DSL connections.
Unless things have changed a great deal since I worked for one of the Telus DSL resellers two years ago, these "independent" ISPs turn over almost all of each subscriber's monthly fee directly to Telus. They have no control over the connectivity they supposedly provide, and are powerless to set network policy (other than how much of the per-megabyte bandwidth charges to eat before passing the remaining cost on to their clients).
The resellers provide you with an installation CD and a mail server, and exist only so that Telus can maintain the illusion that there is some kind of competition.
Making the X server SGID kmem is still giving it far more privileges than it needs in order to be able to do its job. Just being able to read kernel memory is sufficient to get root in most cases (the easiest example to explain being reading the root password right out of the TTY buffers as root is logging on).
Hopefully the new feature makes it into other OSes -- Unix has long needed a standard way of doing this kind of fine grained privilege separation.
No, it's deprecating.
Depreciating is something else entirely, although it's easy to misread one as the other since in the context of a deprecated API the meaning is close enough to get the point across.
Slashdot Mirror
I assume all that stuff has since been removed?
But, I'd hesitate to call it easy to use. The standard C++ language bindings in particular are astonishingly bad:
The situation is reported to look better from other languages, and I can personally confirm that the Java bindings are a delight to work with by comparison (of course in Java it's even easier to just use RMI).
As for the wire-level protocol, I have no complaints about IIOP now that it has readable corbaloc: URLs (the CDR marshalling details are still messy but unless you are writing your own ORB they are taken care of for you). I'm actually a bit surprised that IIOP isn't more widely used on the Internet and in the open source world (outside of GNOME of course) -- it's the distributed computing open standard, it interoperates across languages and OSes, it has numerous open-source implementations, and It Just Works(tm).
Instead we are getting stuff like web services and SOAP, whose wire format is just as incomprehensible to humans (don't kid yourself that XML is easy to read -- have a look at a fully-decked-out SOAP message some day) while using many times as much bandwidth and memory and taking at least ten times as long to parse. (And I say this even though I currently write SOAP gateways for a living.)
This is so true. I too stopped watching TV a couple of years ago, after the web had gotten me too used to being in full control of my own information and entertainment choices.
Now whenever I see others watching TV I'm amazed not just at the percentage of commercial content, but mostly at the stunning levels of sheer, toxic stupidity emanating from the tube. Did it get significantly worse since 1999, or (more frightening still) was it always this bad and I just never noticed?
Or just download Tweak UI (for Win NT,95,98,2k, or for Win XP), which lets you turn off autorun and lots of other retarded misfeatures as well.
Such devices do indeed have serious problems when interposed by ISPs without the knowledge of the end users, although they've gotten more reliable over the years -- basically by letting requests bypass the cache whenever they see anything that doesn't look exactly like a plain-jane request for a static file.
I remember a few years back reading about Paul Vixie's efforts to solve this problem, watching as his posts on the subject went from enthusiastic (in the early trials, where he was seeing 50% hit rates) to disappointed (later on, when the work-arounds for various problems with dynamic content had pushed the hit rate down into the range where it wasn't really worth it anymore).
Last I read (which was years ago) he was going to adapt the technology to make a box for ISPs that would do transparent SMTP proxying to discourage spammers, but it looks like he wasn't fast enough -- nowadays everyone just seems to block mail from dialup IP ranges instead.
He called it "draconian" and said it's powers "fly in the face of a free society," which seems like pretty unreserved criticism to me. What do you guys want from him, a Slashdot-style rant in a single long paragraph replete with poor spelling?
It would be nice, but "ECMAScript" has to be tied with the late, unlamented "PCMCIA" for Ugliest Name in the Galaxy. It has zero probability of ever displacing the highly buzz-friendly term "JavaScript" in common usage.
What would probably happen is that organizations would spend a relatively small amount of money purchasing a new kind of liability insurance, the terms of which would require them to take at least basic steps to secure their systems (ie, stunning incompetence of this private-files-on-open-WEP variety would be disallowed).
And of course "Wired Equivalency" is total BS. Someone sitting in a van half a block away (or renting a suite next to yours) can spend as much time as they need to attack your WEP point -- days or weeks, even -- during which time someone physically in your building would (hopefully) have been noticed and challenged.
It's true that division-by-zero is undefined in the field of real numbers. But engineering applications occasionally make use of other systems in which the result has a well-defined value.
IEEE 754-compliant floating point operations define the additional elements NA, NaN, -Inf and +Inf to enable calculations to continue after divisions by zero are encountered.
Tying moral claims to unresolved empirical issues is a dangerous tactic. Suppose someday researchers discover that older workers really do learn more slowly (or have less brain bandwidth, or whatever), so the particular type of discrimination you argue against is rational after all. Will it then be morally acceptable to discriminate against older workers?
That's a very good point.. I hadn't considered that angle. Even if your current employer is certain never to care, you never know that some future SCO won't come along, buy out the IP, then look for ways to raise a stink.
I'd mod you up but (alas) I've already posted in this discussion.
This makes sense when you think about it from their risk-averse perspective: releasing even small pieces of otherwise-useless specialized code is all downside and no upside.
On the plus side you might might improve goodwill with a small number of open source developers. But on the minus side you might be exposing the company to liability, accidentally revealing sensitive information, or inadvertantly helping your competitors. Plus, management always has to worry about shareholders second-guessing them -- possibly resulting in a shareholder lawsuit -- if the IP you give away is later perceived to have been very valuable.
Given all this, a dangerous but more pragmatic idea might be to just go ahead and contribute at least the small stuff like your patches and bugfixes. As long as you have no official policy forbidding this you can point out that it's just the standard way things are done when working with open source tools.
Let me be clear that I am not actually advising anyone to do this. More just.. thinking aloud.
Can your users access secure web sites? If so, congratulations: your lucky users can also partake of P2P apps, VPNs, SOAP services, and anything else that can tunnel over https -- ie, pretty much anything these days -- and be nearly immune from inspection. This may or may not be a good thing but it's the way things are going.
(l7-filter looks like a pretty cool hack either way, though. Thanks for the link.)
It will the first time. But after that the filter will have learned that lots of single-letter words are an excellent troll predictor.
In Slashdot's case I might even try keeping track of punctuation characters as words. In addition to helping with dash-trolling it could learn to distinguish perl code snippets from ASCII art drawings of the Goatse dude.
Of course as another poster pointed out it would be impossible to stop all trolling since even humans can't reliably tell the difference between a subtle troll and a genuine fringe opinion.
Of course that's 100 billion neurons, not millon. Oops.
The more you think about it, the more it seems very unlikely that a truly human-like machine would ever pay back the cost of producing it.
We humans seem to devote a great deal of hardware to things like food acquisition, mate selection, competition for status, sexual jealousy, gossip and social awareness, family life, and other biological concerns. Many of the problems we solve turn out to be amazingly difficult, and there seems to be little economic reason to develop artificial systems to solve most of them. When viewed in this light the question becomes "Even if we could build a human-like system, who would pay for it and what would they want it for?"
We naturally underestimate the power of the hardware possessed by even the stupidest of humans. As one example, the most complex artificial neural networks have at most a few thousand nodes while we have about 100 million neurons in our brain, all voracious consumers of glucose that somehow earn their keep as we evolve. As a species we willingly pay an astonishing price for our brainy skulls right from the day our mothers suffer no small injury and risk of death pushing them out.
Having said all that, there's no question that it will be useful to have highly intelligent systems that are insightful, attentive to our needs and capable of conversing in natural language.. even if they never do manage to get hungry, fall in love, plot to rebel, or even pass the Turing test.
With every X-Box purchase, MS helpfully provides chips containing both the encryption keys and the instructions for using them.
Use as much encryption as you want, it won't help much when you are giving the key to the attacker.
If a service goes down in the middle of a cluster and no client connections arrive before it comes back up, does it make any downtime?
Which revisionist historians have been persecuted by anti-discrimination laws? The only cases that I can think of that come close are David Irving and Robert Faurisson.
Irving ran afoul of British libel law, not hate speech law, and Faurisson (1) can hardly be called a historian and (2) was in no way merely attempting to "correct errors in the historical record.".
I personally agree that anti-hate-speech laws are misguided (especially the bizarre French law under which Faurisson was prosecuted), and that the answer to bad speech is more speech. But I do have some empathy for the European's strong feelings on this issue given the genocides they've seen first-hand during living memory.
Unless things have changed a great deal since I worked for one of the Telus DSL resellers two years ago, these "independent" ISPs turn over almost all of each subscriber's monthly fee directly to Telus. They have no control over the connectivity they supposedly provide, and are powerless to set network policy (other than how much of the per-megabyte bandwidth charges to eat before passing the remaining cost on to their clients).
The resellers provide you with an installation CD and a mail server, and exist only so that Telus can maintain the illusion that there is some kind of competition.
Making the X server SGID kmem is still giving it far more privileges than it needs in order to be able to do its job. Just being able to read kernel memory is sufficient to get root in most cases (the easiest example to explain being reading the root password right out of the TTY buffers as root is logging on).
Hopefully the new feature makes it into other OSes -- Unix has long needed a standard way of doing this kind of fine grained privilege separation.
Dirct link to the high rez quicktime is here.
No, it's deprecating. Depreciating is something else entirely, although it's easy to misread one as the other since in the context of a deprecated API the meaning is close enough to get the point across.