SysV init that is used in Linux is configurable enough to do all those things by itself -- just put everything in inittab. OTOH, "init scripts" definitely won't exist, or will be completely different from what we are accustomed to.
The point here isn't that we're expecting russia to do something we shouldn't, it's that russia has been stalling and stalling saying "Okay, okay, no problems comrade, just a little more time and we'll have the module ready. And how about a little more cash while we're at it?"
How often NASA itself manages to fit into originally planned cost and do everything without delays? And was there ever as heavy as this criticism of NASA for anything short of blatant mission failure?
Re:...and if problems are related...
on
NASA Gets Smart
·
· Score: 2
I don't have the weight of our Service Module in front of me, but I don't think NASA would make such a hollow threat unless it were true and we had the capability to launch our SM.
Why not? It's not like they are going to actually launch it -- they will wait until Proton will be ready (it definitely will take less time than building a new module) and will demand that their module will be launched, even if at that time Russian one will be already in orbit. Not really useful technically but makes them look better politically.
The new (almuninum-lithium) Super Lightweight External Tank goes a long way to launching heavier payloads.
It definitely will be more expensive than Proton, and still unlikely will put shuttle into a different class of space vehicles. In addition to that there will be a problem, how to attach that thing to the shuttle and keep the whole thing balanced.
...and if problems are related...
on
NASA Gets Smart
·
· Score: 2
...not to module itself but to Proton, how is NASA going to launch their new module, that I can't expect weighting less than Russian one? Using a giant slingshot? By three shuttles, tied together by a rope? By assembling all Republicans together and praying? The whole point of using Proton was that it's the only currently available vehicle that can launch heavy modules reasonably cheap, and the whole space program, including the station is already underfunded.
...however you may want to separate scripts that do database access and ones that don't into different "Applications" in fhttpd config to reduce the number of persistent connections to the database.
Thats a load of BS. In October 1998, INS doubled the fees for all applications. Now they charge $220 for Adjustment of Status (which is what Linus is waiting on) (up from $100). Where is this extra money going?
Nowhere. With the amount of bureaucracy involved INS in nowhere close to being able to support itself from application fees -- it needs money from government to be able to operate.
People who put down sysadmins should be forced to wake up at 3 a.m. to fix an outage every couple of weeks until they retract their statement. If they never retract it, good, more sleep for everyone else. Hrmph.
Please, look, at the comment, to which I was answering -- exact opposite of what you are referring to.
How about this one? Lots of people want to be "Operators" because they get to spend time with their family, and don't get paged at three in the morning or during their daughters 8th birthday party just to come in to work to fix a a bug they introduced because they were just stupid? We also have hobbies that don't include computers, or computer games, and usually involve more than 4 people.
And why should anyone other than you care? Your "hobbies" most likely contribute nothing to lives of others, yet the price, society pays for supporting overpaid drones becomes rather noticeable. Why should I care, how wide was someone's smile at his daughter's birthday party if that person was responsible for network (or phone, or power) outage?
I personally think you're ranting a bit, and not experienced the ease with which it is to develop cross platform tools using XML for data interchange. Try it - you might like it. And if you don't, switch back to CORBA with all the nasties in there, or COM or some other supposedly "cross platform" method of data interchange. And write your own parsers for your own mini-format. There's More Than One Way To Do It (tm).
You completely missed the point. XML is just fine as interchange format -- as I have said, MIME is more wasteful, comma-separated lists are too simple, and key-value pairs are both. The problem is that it has "formal" DTD and is being used for standardization and declaration of formats for applications -- something where semantics (substance) must be primary and actual format (form) serves it. It's clearly unsuitable for this goal, and allows all kinds of abuse.
Parsers are simple, no one even writes them by hand anymore for anything more complex than comma-separated list. Semantics is complex, and every protocol has its own one.
Why everyone is trying to use XML for every possible application while XML itself is not very well "standardizable"? It has no predefined way to attach any formal (or even not so formal) description of semantics of the data to DTD -- I would understand if I was able to attach pieces of, say, portable C code into "the definition", and say that everyone who wants to support my format can compile code, extracted from my "definition" using some standard parser/converter, link it with standard parser library, feed the same DTD to that parser, and the result will be a "skeleton" of compliant with my standard input/output/display/... procedure. But right now we have only trendy-sounding TLA for simple "open tag -- recursion -- close tag" format that isn't much better than anything else, but differs from any other format in rather spectacular way -- no one so far produced completely compliant and usable parser for it in compiled language (no, gnome-xml isn't compliant -- unicode conversion from charsets, other than hardcoded in the source, shows its ugly head).
I understand the need for standardization. I understand that comma-separated values or plain key-value list poorly represent complex nature of the data. I understand that HTML standard committees royally screwed up under the pressure of companies. I understand that in general text is cheap. I understand that XML at least provides some means to show structure and attributes of the data (but so did RFC-822 + MIME more than ten years ago -- just with a bit more waste of space). But sorry, this feeble attempt of meta-standardization just doesn't _do_ enough to justify itself now. Semantics of the data still should be defined in English, and quality of definitions that I see declines rapidly. It helps with displaying that data, but displaying is a microscopic, almost unnoticeable piece of any serious data processing. Semantics still has to be handled by "manually" written, rewritten, ported everywhere and debugged programs that actually are supposed to know what to do with data. Programmers still can't derive any useful information about the data nature from DTD, and should rely on vague texts and their interpretations of it, so the effort, XML saves (writing a parser for arbitrary format) is a big fat zero compared to the real work programmer still has to do to make his program work. No way to do formal proof of anything except that data is formatted as it's supposed to. No way to derive testing procedure for implementation of the processing program. No anything that actually helps programmers to write a useful program and make sure that it works.
Parsers are written in the languages that are nice for demos and small web sites, but don't scale on anything large (what is it, a conspiracy of hardware manufacturers?). I can churn out XML-like meta-standards at the rate ten per week, but since all of them will share the same flaws, why would anyone care? Why do we see a lot of "uses" for XML, but no real progress in improving it in the most natural way -- standardizing the linking between format and semantics? It's possible to keep XML as it is -- it's good enough to define some "canonical" form, the data is (or can be) kept, but without a useful way to handle semantics it's dead.
I am afraid that this situation is created on purpose -- there are already some formats of data that have semantics attached. The problem is, they are proprietary, tied to platforms, languages and architectures. They have semantics, however the formats, they use for data exchange are unnecessarily cryptic or hard to serialize to the stream of bytes. By keeping proprietary "guts" with algorithms, object models, transaction-level protocols and adding "open" formatting of the data vendors get the best of the both worlds -- no one but them can make any sense with the data (both implementation of data handling and the objects-handling engine itstlf are closed-source -- say, COM), but they look "open" and "nice".
PalmIII (,e,x) aren't obsolete -- PalMVII has its wireless network support, but in places like SF Bay Area one would rather use smaller PalmIII and attach Ricochet modem when necessary, like I do -- as opposed to others, Ricochet is fast, and flat rate. And, of course, there are a lot of people, who don't need any wireless connectivity on their PDA, especially if it makes PDA more bulky.
Is there a higher cost associated with using these products? i.e. now your email client can have RSA built in but the cost of the security is passed on to the consumer.
It was free (using existing free libraries), and it will remain free (using the same free libraries). It only changes one thing -- now every individual had to download them from abroad because RSA had no resources to prosecute against users but could prosecute against distributors. When patents will expire distributors will include everything in their CDs/ FTP sites/whatever, so users will get everything from them.
As for commercial software -- what commercial software? Who in his right mind will use encryption products that don't pass through scrutiny that only open source makes possible?
The reason standards bodies don't avoid patented technology is that they have to concentrate on producing a good, rather than absolutely free, standard. Often there is no alternative to do anything but go with a patented soultion because no other solutions are avaliable. For instance, I challenge you to find a non-patented technique for audio compression that even approaches systems like MP3 and AAC...
But what prevents standard bodies from demanding everyone who wants to have their technology included in the standard, to abandon patents, or provide free blanket license? Then companies will have choices -- keep technology proprietary, or have it "endorsed" as a standard (and benefitting from already having that technology _implemented_) at the price that they no longer can put any restrictions or licensing fees on it?
RFCs about IP and related protocols don't contain patented algorithms, and TCP/IP seems to turn out just fine without them. OTOH, ITU "standardizes" all kinds of proprietary shit, and this is why telephony is such a mess when it comes to interoperability.
If people said "What about VCR's?" I'd just say, in this case, "Well, you can't copy VCR tapes either, and the only reason that VCRs are legal is for "time-shifting" programs. You can't show em publicly and you can't give out copies."
Wrong.
It's legal to produce and keep one copy of any copyrighted product, one legally possesses. Only distribution of such copies is illegal.
Even if VCRs were useful mostly for illegal activity, it would not be sufficient to ban them.
Linux is a versatile OS but it has many shortcomings which the pundits and many linux enthusiasts want to ignore - the hardware drivers are not on par with their NT counterparts, nor is the support infrastructure there. There are some drivers which are rock solid under linux, but the majority of them have quirks, bugs, and I'd say upwards of 50% are in "perma-beta". Further, the tcp/ip lock-spin problem as surfaced in the mindcraft testing seems to prove that linux does what it was designed to do: run well on *well supported* commodity hardware, do so with good stability, and makes an excellent server for home / small business use. However, for mega corporations and so-called "e-commerce" - it's lacking.
The configuration in Mindcraft tests that "discovered" problem with networking was absolutely unrealistic -- multiple 100-megabit ethernet cards that on the server. The only justification for such a configuration is being a router because in all other cases one gigabit card is more efficient, and it never was an intention to make system do tricks that in all sane configurations are reserved to routers and switches.
In sane configuration with one faster card "problem" never occurs, and definitely data-pumping from disk to ethernet is not a limiting factor on "e-commerce" site's performance with any high-end hardware and any networking-capable OS -- processing speed, resource allocation, scheduler and local I/O are.
Maybe I don't understand something about/. moderation, but the above poster is absolutely correct, the people talking about the result taking a chunk out of big bang theory are making a catagoreical error; certainly doesn't deserve a score of 0. Why the hell do some people just HAVE to comment on something regardless of the fact that they know they don't know what they are talking about?
The article wasn't moderated -- all anonymous articles start with the score 0, as opposed to normal users. Moderators later can moderate anonymous articles up, but this one was just posted, so it remained with its initial score.
Isn't that like saying Microsoft can write better software than the Open SOurce community?
I have meant that grandma that need a technical support to find a "Print" button is unlikely to make a greeting card on computer that will be in any way superior to Hallmark card bought in the store, with text that she can easier write with a normal pen, but yes, Open Source community never was too prominent in greeting cards business.
Take a minute and critically (that means no demagogic knee-jerking kids) about what life would be like if Microsoft didn't exist tomorrow. Are you Open Source guys ready to give free tech support to every 90 yr old granny who wants to print a birthday card? I sure hope so in such a circumstance...
This is why specialized support companies exist. I personally won't do even phone support for my own fhttpd for every non-technical user, but this is the reason why open source software is profitable for businesses -- someone else can have its own support infrastructure make money on it.
As for every grandma trying to print a greeting card, no one would be hurt if she wouldn't be able to -- Hallmark makes better ones anyway. One who is interested in selling software to users, whose support doesn't justify the distribution cost should deal with this problem, not developers.
Slashdot Mirror
SysV init that is used in Linux is configurable enough to do all those things by itself -- just put everything in inittab. OTOH, "init scripts" definitely won't exist, or will be completely different from what we are accustomed to.
t'd be nice to use the ssh client on a PDA, but why would you want to run the daemon on the road?
There is TG ssh for PalmOS -- I use it often with Metricom/Ricochet modem.
The point here isn't that we're expecting russia to do something we shouldn't, it's that russia has been stalling and stalling saying "Okay, okay, no problems comrade, just a little more time and we'll have the module ready. And how about a little more cash while we're at it?"
How often NASA itself manages to fit into originally planned cost and do everything without delays? And was there ever as heavy as this criticism of NASA for anything short of blatant mission failure?
I don't have the weight of our Service Module in front of me, but I don't think NASA would make such a hollow threat unless it were true and we had the capability to launch our SM.
Why not? It's not like they are going to actually launch it -- they will wait until Proton will be ready (it definitely will take less time than building a new module) and will demand that their module will be launched, even if at that time Russian one will be already in orbit. Not really useful technically but makes them look better politically.
The new (almuninum-lithium) Super Lightweight External Tank goes a long way to launching heavier payloads.
It definitely will be more expensive than Proton, and still unlikely will put shuttle into a different class of space vehicles. In addition to that there will be a problem, how to attach that thing to the shuttle and keep the whole thing balanced.
...not to module itself but to Proton, how is NASA going to launch their new module, that I can't expect weighting less than Russian one? Using a giant slingshot? By three shuttles, tied together by a rope? By assembling all Republicans together and praying? The whole point of using Proton was that it's the only currently available vehicle that can launch heavy modules reasonably cheap, and the whole space program, including the station is already underfunded.
...however you may want to separate scripts that do database access and ones that don't into different "Applications" in fhttpd config to reduce the number of persistent connections to the database.
Sure theres that mpegtv player for linux, its better but its SLOW compared with windows. It gets around 15fps on an mpeg encoded at 30.
Hint: don't run it through X server emulator on Windows box, epecially over 10baseT network with misconfigured routers, especially under a DoS attack.
What a smart guy! Too bad, he didn't know that to get citizenship this way one has to get a green card first.
In 1993, in Colorado I have opened bank account immediately after arriving on B-2 (visitor) visa, much earlier than I've got SSN, H-1B visa, etc.
Thats a load of BS. In October 1998, INS doubled the fees for all applications. Now they charge $220 for Adjustment of Status (which is what Linus is waiting on) (up from $100). Where is this extra money going?
Nowhere. With the amount of bureaucracy involved INS in nowhere close to being able to support itself from application fees -- it needs money from government to be able to operate.
People who put down sysadmins should be forced to wake up at 3 a.m. to fix an outage every couple of weeks until they retract their statement. If they never retract it, good, more sleep for everyone else. Hrmph.
Please, look, at the comment, to which I was answering -- exact opposite of what you are referring to.
How about this one? Lots of people want to be "Operators" because they get to spend time with their family, and don't get paged at three in the morning or during their daughters 8th birthday party just to come in to work to fix a a bug they introduced because they were just stupid? We also have hobbies that don't include computers, or computer games, and usually involve more than 4 people.
And why should anyone other than you care? Your "hobbies" most likely contribute nothing to lives of others, yet the price, society pays for supporting overpaid drones becomes rather noticeable. Why should I care, how wide was someone's smile at his daughter's birthday party if that person was responsible for network (or phone, or power) outage?
I personally think you're ranting a bit, and not experienced the ease with which it is to develop cross platform tools using XML for data interchange. Try it - you might like it. And if you don't, switch back to CORBA with all the nasties in there, or COM or some other supposedly "cross platform" method of data interchange. And write your own parsers for your own mini-format. There's More Than One Way To Do It (tm).
You completely missed the point. XML is just fine as interchange format -- as I have said, MIME is more wasteful, comma-separated lists are too simple, and key-value pairs are both. The problem is that it has "formal" DTD and is being used for standardization and declaration of formats for applications -- something where semantics (substance) must be primary and actual format (form) serves it. It's clearly unsuitable for this goal, and allows all kinds of abuse.
Parsers are simple, no one even writes them by hand anymore for anything more complex than comma-separated list. Semantics is complex, and every protocol has its own one.
Why everyone is trying to use XML for every possible application while XML itself is not very well "standardizable"? It has no predefined way to attach any formal (or even not so formal) description of semantics of the data to DTD -- I would understand if I was able to attach pieces of, say, portable C code into "the definition", and say that everyone who wants to support my format can compile code, extracted from my "definition" using some standard parser/converter, link it with standard parser library, feed the same DTD to that parser, and the result will be a "skeleton" of compliant with my standard input/output/display/... procedure. But right now we have only trendy-sounding TLA for simple "open tag -- recursion -- close tag" format that isn't much better than anything else, but differs from any other format in rather spectacular way -- no one so far produced completely compliant and usable parser for it in compiled language (no, gnome-xml isn't compliant -- unicode conversion from charsets, other than hardcoded in the source, shows its ugly head).
I understand the need for standardization. I understand that comma-separated values or plain key-value list poorly represent complex nature of the data. I understand that HTML standard committees royally screwed up under the pressure of companies. I understand that in general text is cheap. I understand that XML at least provides some means to show structure and attributes of the data (but so did RFC-822 + MIME more than ten years ago -- just with a bit more waste of space). But sorry, this feeble attempt of meta-standardization just doesn't _do_ enough to justify itself now. Semantics of the data still should be defined in English, and quality of definitions that I see declines rapidly. It helps with displaying that data, but displaying is a microscopic, almost unnoticeable piece of any serious data processing. Semantics still has to be handled by "manually" written, rewritten, ported everywhere and debugged programs that actually are supposed to know what to do with data. Programmers still can't derive any useful information about the data nature from DTD, and should rely on vague texts and their interpretations of it, so the effort, XML saves (writing a parser for arbitrary format) is a big fat zero compared to the real work programmer still has to do to make his program work. No way to do formal proof of anything except that data is formatted as it's supposed to. No way to derive testing procedure for implementation of the processing program. No anything that actually helps programmers to write a useful program and make sure that it works.
Parsers are written in the languages that are nice for demos and small web sites, but don't scale on anything large (what is it, a conspiracy of hardware manufacturers?). I can churn out XML-like meta-standards at the rate ten per week, but since all of them will share the same flaws, why would anyone care? Why do we see a lot of "uses" for XML, but no real progress in improving it in the most natural way -- standardizing the linking between format and semantics? It's possible to keep XML as it is -- it's good enough to define some "canonical" form, the data is (or can be) kept, but without a useful way to handle semantics it's dead.
I am afraid that this situation is created on purpose -- there are already some formats of data that have semantics attached. The problem is, they are proprietary, tied to platforms, languages and architectures. They have semantics, however the formats, they use for data exchange are unnecessarily cryptic or hard to serialize to the stream of bytes. By keeping proprietary "guts" with algorithms, object models, transaction-level protocols and adding "open" formatting of the data vendors get the best of the both worlds -- no one but them can make any sense with the data (both implementation of data handling and the objects-handling engine itstlf are closed-source -- say, COM), but they look "open" and "nice".
PalmIII (,e,x) aren't obsolete -- PalMVII has its wireless network support, but in places like SF Bay Area one would rather use smaller PalmIII and attach Ricochet modem when necessary, like I do -- as opposed to others, Ricochet is fast, and flat rate. And, of course, there are a lot of people, who don't need any wireless connectivity on their PDA, especially if it makes PDA more bulky.
Is there a higher cost associated with using these products? i.e. now your email client can have RSA built in but the cost of the security is passed on to the consumer.
It was free (using existing free libraries), and it will remain free (using the same free libraries). It only changes one thing -- now every individual had to download them from abroad because RSA had no resources to prosecute against users but could prosecute against distributors. When patents will expire distributors will include everything in their CDs/ FTP sites/whatever, so users will get everything from them.
As for commercial software -- what commercial software? Who in his right mind will use encryption products that don't pass through scrutiny that only open source makes possible?
Hey, Q! Someone is dissing you and your continuum...
The reason standards bodies don't avoid patented technology is that they have to concentrate on producing a good, rather than absolutely free, standard. Often there is no alternative to do anything but go with a patented soultion because no other solutions are avaliable. For instance, I challenge you to find a non-patented technique for audio compression that even approaches systems like MP3 and AAC...
But what prevents standard bodies from demanding everyone who wants to have their technology included in the standard, to abandon patents, or provide free blanket license? Then companies will have choices -- keep technology proprietary, or have it "endorsed" as a standard (and benefitting from already having that technology _implemented_) at the price that they no longer can put any restrictions or licensing fees on it?
RFCs about IP and related protocols don't contain patented algorithms, and TCP/IP seems to turn out just fine without them. OTOH, ITU "standardizes" all kinds of proprietary shit, and this is why telephony is such a mess when it comes to interoperability.
If people said "What about VCR's?" I'd just say, in this case, "Well, you can't copy VCR tapes either, and the only reason that VCRs are legal is for "time-shifting" programs. You can't show em publicly and you can't give out copies."
Wrong.
Linux is a versatile OS but it has many shortcomings which the pundits and many linux enthusiasts want to ignore - the hardware drivers are not on par with their NT counterparts, nor is the support infrastructure there. There are some drivers which are rock solid under linux, but the majority of them have quirks, bugs, and I'd say upwards of 50% are in "perma-beta". Further, the tcp/ip lock-spin problem as surfaced in the mindcraft testing seems to prove that linux does what it was designed to do: run well on *well supported* commodity hardware, do so with good stability, and makes an excellent server for home / small business use. However, for mega corporations and so-called "e-commerce" - it's lacking.
The configuration in Mindcraft tests that "discovered" problem with networking was absolutely unrealistic -- multiple 100-megabit ethernet cards that on the server. The only justification for such a configuration is being a router because in all other cases one gigabit card is more efficient, and it never was an intention to make system do tricks that in all sane configurations are reserved to routers and switches.
In sane configuration with one faster card "problem" never occurs, and definitely data-pumping from disk to ethernet is not a limiting factor on "e-commerce" site's performance with any high-end hardware and any networking-capable OS -- processing speed, resource allocation, scheduler and local I/O are.
Maybe I don't understand something about /. moderation, but the above poster is absolutely correct, the people talking about the result taking a chunk out of big bang theory are making a catagoreical error; certainly doesn't deserve a score of 0. Why the hell do some people just HAVE to comment on something regardless of the fact that they know they don't know what they are talking about?
The article wasn't moderated -- all anonymous articles start with the score 0, as opposed to normal users. Moderators later can moderate anonymous articles up, but this one was just posted, so it remained with its initial score.
..., saboteur Belluzzo, clueless directors/managers and suicidal pricing policy did to SGI.
Isn't that like saying Microsoft can write better software than the Open SOurce community?
I have meant that grandma that need a technical support to find a "Print" button is unlikely to make a greeting card on computer that will be in any way superior to Hallmark card bought in the store, with text that she can easier write with a normal pen, but yes, Open Source community never was too prominent in greeting cards business.
Take a minute and critically (that means no demagogic knee-jerking kids) about what life would be like if Microsoft didn't exist tomorrow. Are you Open Source guys ready to give free tech support to every 90 yr old granny who wants to print a birthday card? I sure hope so in such a circumstance...
This is why specialized support companies exist. I personally won't do even phone support for my own fhttpd for every non-technical user, but this is the reason why open source software is profitable for businesses -- someone else can have its own support infrastructure make money on it.
As for every grandma trying to print a greeting card, no one would be hurt if she wouldn't be able to -- Hallmark makes better ones anyway. One who is interested in selling software to users, whose support doesn't justify the distribution cost should deal with this problem, not developers.
Unlikely, but I think, journalists will copy wire services releases and add "Me, too" at the end.