MS (and other proprietary) file formats should only be supported in some neutral format with actual import/export tools. The reason is that the MS proprietary poison formats should not infext Linux tools directly. For much of this personal data in documents and especially in things like address books and such the data should be in some neutral format like XML in order for it to be maximally useful to multiple tools. A set of KDE specific formats (one for each tool) would not be terribly useful to the world either imho. There is no reason why the actual translation to/from neutral formats needs to be non-transparent once the program is configured to accept that kind of data transformation with suitable plug-ins.
What is the effect of a government, tax financed patent? Can the NSA dictate terms for use of this technology for the next 25 years or so and charge licensing fees? The technology is useful for many information filtering tasks besides Echelon style eavesdropping. The general technique should no more be patented than other software algorithms.
On the privacy thing I have assumed for quite some time now that as technology advances there is no such thing as real privacy. I am afraid we need to get use to that idea and work to insure there will not be ensuing abuse of other rights when almost all details of our lives are semi-public.
Being a programmer is one thing. Being an actual CS seems to take a bit more. Turning out code is something one with the requisite skills and temperament can do. Actually thinking deeply about the problems and classes of problems and coming up with new and better or at least state-of-the-art solutions goes a bit beyond just being a programmer. But I agree that a degree or two or three will not guarantee you have this extra something. It makes it more likely you have at least been exposed to more tools and challenges to think a bit more abstractly though.
Where the degree (MS or higher) really helps is getting into more research type projects and organizations. Some of these places won't even talk to you without a piece of paper regardless of how many people have witnessed you walking on water for how many years.
You better believe it. At least for the technical basics of the art like data structures, survey of languages, computability theory and automata, database, compiler/interpreters and OS. I consider these things crucial to be a well rounded hacker type. A lot of this you might pick up on your own if you're motivated. I picked up a better compiler education when I examined the texts for the basis to write a fairly substantial compiler myself rather than when just taking a class.
At 25 it is definitely not too late. My God, I will happily go back to school myself when money/time permits and I've been in the commercial software world for 20 years, much of it on the bleeding edge. I want to do some real reasearch beyond what I can cajole the current company or contract to see as necessary. I've run into too many interesting questions, puzzles and areas that are just itching to be explored more deeply over the years.
They could even make formatting structure out-of-line with the main etext if they want to preserve that format. The formating would have relative position pointers into the text (assuming it changes seldom if ever). Making the text more useable in different contexts and preferences would go a long way to gathering more interest.
Some of the items listed which are quite costly to develop and must recoup those costs to be doable at all. Especially, hi-tech agriculture products, and vaccines and medicines. Much software is also not cheap to produce in terms of effort, scarcity of sufficiently skilled people, creativity and so on. That something is needful for life is not an argument that it must be free of cost to the user or be fully in the public domain.
I do agree that somethings are absurd to patent like the human genome and most software. I do not agree it is wrong to copyright a book or a software work or a record or that simply making such digital makes copyright wrong. Much depends on the nature of the copyright and what its terms enable and disable. The authors and artists must be paid for their works as long as we have a system that functions on money at all. If this can be done w/o copyright then wonderful but we cannot rush to cry "public domain" w/o bothering to take care of paying our creators and investors well for their efforts. Not if we wish to continue to receive the fruits of those efforts.
Equating Open Source with communism or any other form of sociall/government organization is silly. Open Source has zero to do with that old "to each according to his ability, to each according to his need" BS. That slogan does not work, never has and never will. The problem is who determines what is whose ability and need? The practical implications suck big time. Open Source is simply that, Open Source. It says that real effective software must include source and at least some rights to do various things with the source. Of and by itself that hasn't a thing to do with any isms. Even Stallman says that free software does not mean at zero price necessarily. It means you are free to do with it what you will once you have acquired it.
I don't believe that any state should be able to declare that some aspect of science or computers or software is the officially sanctioned state version. In non-communist countries we have the wonderful right to laugh our arse off if any political critters try such a thing and tell them to go take a leap into political oblivion. I will not play the game of oohing over the billion odd enslaved Chinese who are now potential users. I don't profit by slavery or claim it is pragmatic to see moves by a slave state toward something I believe should be more used as a good thing.
OK, so I have a bad attitude. I am royally PO'd that I can get a simple, efficient WEB browser in Linux. Netscape locks up every few hours it seems like. The KDE built-in browser is used more by me for casual browsing because at least it doesn't lock up although it occassionaly (relatively rarely) crashes. But it doesn't know what to do with what I though was some fairly simple HTML like some form based stuff. Can't some of us who know about such beasties manage to beef this thing up a bit? It might be the only alternative for some of us. And while I am on the subject, what is with fonts and text size on Linux vs. Windows in Netscape? Why isn't the "larger/smaller" text selector enabled? Some sites use tiny fonts that don't respond to changing Netscape overall fonts or changing kfm font size either. But I can control the size of the displayed text in IE5. This is no joke to those of us with aging eyes who read a LOT online.
What is up with this browser stuff? It is not my area of anything approaching expertise or even competence. But dammit, this krap is making my favorite platform look pretty bad. I've stuck my head into Mozilla just a bit. First impression is it is way too bloated with everything in the known universe. My suggestion is to get a good basic browser working that is fully up to HTML4 standards, is fast and rock solid and only then add modules to do fancier things. Anyone doing this?
Why do I care that programs upload a guid? UUIDs are a very efficient tool for a lot more than keeping track of who is using the product and their various pieces of registration information. We are sort of schizoid about this stuff. Privacy is effectively dead through techological innovation. You are literally watched or watchable almost every moment of the day. Or didn't anyone notice these capabilities? We groan about RP getting our id and at the same time would like to live in a world that tailors itself to our likes and dislikes a bit more closely including advising us of various opportunities and products (at least when we want to know). This sort of stuff can't be done without gathering information and knowing who you are. Online business cannot be finalized without you effectively having a digital signature/fingerprint.
Perhaps the needed balance is the ability to simply say NO when we wish to or provide alternate minimum information.
Do you wish to insult all users of components? Have you not noticed the use of components in both KDE and to a large extent GNOME? Do you wish to insult all consultants? What do you think a lot of open source hackers do to make their daily bread? Why assume that consultants cannot consult based upon open source? Why insult all components and the idea of components just because you didn't like a particular set of components? Do you even understand what components are and are not?
To the extent your comment is true there are errors in the vendor's implementations of ORBS and OAs. The entire idea is interoperability including bridges between different vendor's ORBs. If this is broken then something is seriously wrong.
Being part of the open source community does not require that one be simple-minded. There is a lot more to distributed programming environments that actualy work than a simplistic set of transport and RPC mechanisms. Those of us that have spent much time doing real world distributed programming know this. There is nothing wrong with HTTP, DNS, XML and so on for what they are good for. But claiming they will take the place of tools designed for things they were never been to cover is plain wrong.
The idea that CORBA has been killed off by the WEB is utterly absurd. CORBA is a hell of a lot more than an RPC transport mechanism. CORBA calls, despite there often bemoaned slowness are much faster than any of the standard WEB technologies such as JSP, servlets, CGI, XML RPC proposals, SOAP and so on. Jave RMI does a fraction of what CORBA does but only for a single language environment. IIOP became meaningful to a hell of a lot of organizations judging by what is actually used. HTTP-based XML RPC is only a transport and RPC mechanism and not a terribly efficient one at that. It doesn't even address where the receiver finds the object/function addressed or how the arguments should be unmarshaled, much less transaction services, name services, persistent services and the rest addressed by CORBA. Nor does it specify what distributed objects and functionality is avialable in a standard way. It is frankly worse than useless to diss a technology without understanding what it attempts to address.
I don't believe it already exists. In evidence I point out that there is question about its existence. With a truly effective open source component repository the question would be very succinctly answered. This touches on an important point. One of the problems with such a repository is having a truly effective indexing system and some way of categorizing the semantic content of items to some level. The entire software industry awaits such a beastie.
Outside of some obvious markets like PDAs and appliances, why is there such emphasis on thin client? The computing power, network bandwidth, memory and so on are all rising pretty steeply. With technology like Java (only one of many such) that enables download of objects/classes as needed and competent end-nodes, who cares about thin-clients? What it seems to me is really needed is much more powerful peer-peer object communication and management. Too much thinking gets straight-jacketed into client vs. server.
Does anyone else find it not at all a good sign that SUN is backing an office suite that is a nearly totally closed environment? There is no API, no COM, no CORBA or RMI access to the applications and data within Star Office AFAIK. So exactly what does it mean when the SO folk hero talks of bringing SO functionality to the WEB and the rest of the world? Exactly how will this be done without open APIs? If this is the story then there is merit to the notion that SUN is attempting to out-Microsoft Microsoft. At least MS had the sense to use and publish COM APIs to most of its office products.
I also get quite a few "Unexpected (or what it Unknown) error encountered. Data is being saved" crashers from SO 5.1 on RH Linux 6.0 SMP.
I have run into this annoying problem myself a lot. IMHO it is seriously broken that email programs, calendar apps, address books and so on take the user's info and encode it into some form that is not useable outside (exceptions for standard unix email format) the proprietary program that first touched and "managed" the information. What I plan to hack on some day RSN is representing this info in a non-proprietary format as XML with code to go from/to various proprietary formats. Eventually GUI minded folk can write these tools that take the XML as their native format.
What is the big deal? SOAP is a simple and fairly intuitive way to do object messaging and data exchange across the WEB. It is perfectly useable with Java, CORBA, Python, COM or whatever else ultimately handles execution of the appropriate method on the specified object. It will not replace any of these. It is only a communication protocol. It says nothing about the many services necessary to have a true distributed object environment. Now that the proposal is out there is no way for MS to control who makes use of the basic idea.
Also, CORBA always has been non-proprietary and OS independent.
If we have a common data format (XMLish mayhap) for e-wallets and if we have strong encryption on the content and standardized management protocols then there would be no reason any particular vendor should have a monopoly just as there is no reason for any vendor to have a monopoly on checking accounts. The information should not be stored centrally with Microsoft or anyone else. Some of the new smartcards that are readable wirelessly by computer might be a choice for keeping this info out of centralized hands.
Re:Case against consultoids
on
Managing Geeks
·
· Score: 1
As a geek I am quite tired of the notion that any company owns me or my brain. At most companies the infrastructure level that I am most interested in hacking/improving and the general software environment are actually pretty common to tons of specific companies and applications. I would much rather develop improvements at company A and see them implemented and then go do something similar at company B and spread the knowledge to others to disseminate the techniques more quickly. Being the property of a particular company generally limits the good stuff to that company's property. This is wrong. Also being attached permanently to one company subjects one to cycles where they want to use the talents of their geeks to fight current fires without giving the geeks room to deal with what the really important issues are. This frustrates mightily. I am not sure precisely how to structure my career yet but I do not wish to ever be in a position again where I produce something that I truly believe in only to have one company own it and mothball it if they don't understand what to do with it and are afraid to spin it off.
Isn't one of the great things that not only are great things happening and we can be part of them but that we can actually be comfortable financially in the process? Sometimes we seem to have bought into the old myth that says to make money is inherently evil.
Someone please correct me if I am wrong, but I think the biggest thing missing on Linux today is a good software component technology. CORBA does part of the job but as far as I understand it (and my knowledge is limited here) the equivalent of INPROCs either can't be done or are seldom done and there is a bit more overhead involved in CORBA use than in COM in the MS world. Without a good component technology and reasonably sound interface specifications I am afraid that continual bifurcation and confusion of libraries will result. Tied into this problem is another area I don't have the full skinny on in Linux yet. The ELF binary format supports real DLLs allowing full runtime dynamic linking to the best of my understanding. Yet most instructions on incorporating feature XYZ into my application ask me to linking against some libXYZ.so. With true dynamic linking this is not really necessary and simply creates another dependency to be broken later on.
Is there a general plan for good lightweight component software that takes advantage of true DLLS (among other things) on Linux? Is this plan at a central enough level that it will be generally available rather than dependent on a particular development flavor like KDE or GNOME or a particular Linux distribution?
WHAT?? I've been in this business for over 2 decades and I know for a fact that there were no "good ole days" in software. And any designer/programmer who has done a moderately large software system knows why object oriented technologies are a blessing. Of course they are more often than not not used well but this isn't the fault of OO per se. I more blame the fact that making better programming tools/environments/methodology just isn't a high profit proposition. So we end up automating everyone's work but our own.
Slashdot Mirror
MS (and other proprietary) file formats should only be supported in some neutral format with actual import/export tools. The reason is that the MS proprietary poison formats should not infext Linux tools directly. For much of this personal data in documents and especially in things like address books and such the data should be in some neutral format like XML in order for it to be maximally useful to multiple tools. A set of KDE specific formats (one for each tool) would not be terribly useful to the world either imho. There is no reason why the actual translation to/from neutral formats needs to be non-transparent once the program is configured to accept that kind of data transformation with suitable plug-ins.
What is the effect of a government, tax financed patent? Can the NSA dictate terms for use of this technology for the next 25 years or so and charge licensing fees? The technology is useful for many information filtering tasks besides Echelon style eavesdropping. The general technique should no more be patented than other software algorithms.
On the privacy thing I have assumed for quite some time now that as technology advances there is no such thing as real privacy. I am afraid we need to get use to that idea and work to insure there will not be ensuing abuse of other rights when almost all details of our lives are semi-public.
Being a programmer is one thing. Being an actual CS seems to take a bit more. Turning out code is something one with the requisite skills and temperament can do. Actually thinking deeply about the problems and classes of problems and coming up with new and better or at least state-of-the-art solutions goes a bit beyond just being a programmer. But I agree that a degree or two or three will not guarantee you have this extra something. It makes it more likely you have at least been exposed to more tools and challenges to think a bit more abstractly though.
Where the degree (MS or higher) really helps is getting into more research type projects and organizations. Some of these places won't even talk to you without a piece of paper regardless of how many people have witnessed you walking on water for how many years.
You better believe it. At least for the technical basics of the art like data structures, survey of languages, computability theory and automata, database, compiler/interpreters and OS. I consider these things crucial to be a well rounded hacker type. A lot of this you might pick up on your own if you're motivated. I picked up a better compiler education when I examined the texts for the basis to write a fairly substantial compiler myself rather than when just taking a class.
At 25 it is definitely not too late. My God, I will happily go back to school myself when money/time permits and I've been in the commercial software world for 20 years, much of it on the bleeding edge. I want to do some real reasearch beyond what I can cajole the current company or contract to see as necessary. I've run into too many interesting questions, puzzles and areas that are just itching to be explored more deeply over the years.
They could even make formatting structure out-of-line with the main etext if they want to preserve that format. The formating would have relative position pointers into the text (assuming it changes seldom if ever). Making the text more useable in different contexts and preferences would go a long way to gathering more interest.
Some of the items listed which are quite costly to develop and must recoup those costs to be doable at all. Especially, hi-tech agriculture products, and vaccines and medicines. Much software is also not cheap to produce in terms of effort, scarcity of sufficiently skilled people, creativity and so on. That something is needful for life is not an argument that it must be free of cost to the user or be fully in the public domain.
I do agree that somethings are absurd to patent like the human genome and most software. I do not agree it is wrong to copyright a book or a software work or a record or that simply making such digital makes copyright wrong. Much depends on the nature of the copyright and what its terms enable and disable. The authors and artists must be paid for their works as long as we have a system that functions on money at all. If this can be done w/o copyright then wonderful but we cannot rush to cry "public domain" w/o bothering to take care of paying our creators and investors well for their efforts. Not if we wish to continue to receive the fruits of those efforts.
Irrelevant and hardly on the same scale in current times especially. By your logic all abuses would be ignored due to no one being perfect.
Equating Open Source with communism or any other form of sociall/government organization is silly. Open Source has zero to do with that old "to each according to his ability, to each according to his need" BS. That slogan does not work, never has and never will. The problem is who determines what is whose ability and need? The practical implications suck big time. Open Source is simply that, Open Source. It says that real effective software must include source and at least some rights to do various things with the source. Of and by itself that hasn't a thing to do with any isms. Even Stallman says that free software does not mean at zero price necessarily. It means you are free to do with it what you will once you have acquired it.
I don't believe that any state should be able to declare that some aspect of science or computers or software is the officially sanctioned state version. In non-communist countries we have the wonderful right to laugh our arse off if any political critters try such a thing and tell them to go take a leap into political oblivion. I will not play the game of oohing over the billion odd enslaved Chinese who are now potential users. I don't profit by slavery or claim it is pragmatic to see moves by a slave state toward something I believe should be more used as a good thing.
OK, so I have a bad attitude. I am royally PO'd that I can get a simple, efficient WEB browser in Linux. Netscape locks up every few hours it seems like. The KDE built-in browser is used more by me for casual browsing because at least it doesn't lock up although it occassionaly (relatively rarely) crashes. But it doesn't know what to do with what I though was some fairly simple HTML like some form based stuff. Can't some of us who know about such beasties manage to beef this thing up a bit? It might be the only alternative for some of us. And while I am on the subject, what is with fonts and text size on Linux vs. Windows in Netscape? Why isn't the "larger/smaller" text selector enabled? Some sites use tiny fonts that don't respond to changing Netscape overall fonts or changing kfm font size either. But I can control the size of the displayed text in IE5. This is no joke to those of us with aging eyes who read a LOT online.
What is up with this browser stuff? It is not my area of anything approaching expertise or even competence. But dammit, this krap is making my favorite platform look pretty bad. I've stuck my head into Mozilla just a bit. First impression is it is way too bloated with everything in the known universe. My suggestion is to get a good basic browser working that is fully up to HTML4 standards, is fast and rock solid and only then add modules to do fancier things. Anyone doing this?
Why do I care that programs upload a guid? UUIDs are a very efficient tool for a lot more than keeping track of who is using the product and their various pieces of registration information. We are sort of schizoid about this stuff. Privacy is effectively dead through techological innovation. You are literally watched or watchable almost every moment of the day. Or didn't anyone notice these capabilities? We groan about RP getting our id and at the same time would like to live in a world that tailors itself to our likes and dislikes a bit more closely including advising us of various opportunities and products (at least when we want to know). This sort of stuff can't be done without gathering information and knowing who you are. Online business cannot be finalized without you effectively having a digital signature/fingerprint.
Perhaps the needed balance is the ability to simply say NO when we wish to or provide alternate minimum information.
Do you wish to insult all users of components? Have you not noticed the use of components in both KDE and to a large extent GNOME?
Do you wish to insult all consultants? What do you think a lot of open source hackers do to make their daily bread? Why assume that consultants
cannot consult based upon open source? Why insult all components and the idea of components just because you didn't like a particular
set of components? Do you even understand what components are and are not?
To the extent your comment is true there are errors in the vendor's implementations
of ORBS and OAs. The entire idea is interoperability including bridges between different
vendor's ORBs. If this is broken then something is seriously wrong.
Being part of the open source community does not require that one be simple-minded.
There is a lot more to distributed programming environments that actualy work than a simplistic set
of transport and RPC mechanisms. Those of us that have spent much time doing real world distributed
programming know this. There is nothing wrong with HTTP, DNS, XML and so on for what they are good
for. But claiming they will take the place of tools designed for things they were never been to cover is plain
wrong.
The idea that CORBA has been killed off by the WEB is utterly absurd. CORBA is a hell of a lot more than an RPC transport mechanism. CORBA calls, despite there often bemoaned slowness are much faster than any of the standard WEB technologies such as JSP, servlets, CGI, XML RPC proposals, SOAP and so on. Jave RMI does a fraction of what CORBA does but only for a single language environment. IIOP became meaningful to a hell of a lot of organizations judging by what is actually used. HTTP-based XML RPC is only a transport and RPC mechanism and not a terribly efficient
one at that. It doesn't even address where the receiver finds the object/function addressed or how the arguments should be unmarshaled, much less transaction services, name services, persistent services
and the rest addressed by CORBA. Nor does it specify what distributed objects and functionality is avialable in a standard way. It is frankly worse than useless to diss a technology without
understanding what it attempts to address.
I don't believe it already exists. In evidence I point out that there is question
about its existence. With a truly effective open source component repository
the question would be very succinctly answered. This touches on
an important point. One of the problems with such a repository is having
a truly effective indexing system and some way of categorizing the semantic
content of items to some level. The entire software industry awaits such a
beastie.
Outside of some obvious markets like PDAs and appliances, why is there such emphasis on thin client? The computing power, network bandwidth, memory and so on are all rising pretty steeply. With technology like Java (only one of many such) that enables download of objects/classes as needed and competent end-nodes, who cares about thin-clients? What it seems to me is really needed is much more powerful peer-peer object communication and management. Too much thinking gets straight-jacketed into client vs. server.
What are distributed objects that are OS and platform independent good for? Presumably you already know the answer. (Hint: it is NOT "nothing")
Does anyone else find it not at all a good sign that SUN is backing an office suite that is a nearly totally closed environment? There is no API, no COM, no CORBA or RMI access to the applications and data within Star Office AFAIK. So exactly what does it mean when the SO folk hero talks of bringing SO functionality to the WEB and the rest of the world? Exactly how will this be done without open APIs? If this is the story then there is merit to the notion that SUN is attempting to out-Microsoft Microsoft. At least MS had the sense to use and publish COM APIs to most of its office products.
I also get quite a few "Unexpected (or what it Unknown) error encountered. Data is being saved" crashers from SO 5.1 on RH Linux 6.0 SMP.
I have run into this annoying problem myself a lot. IMHO it is seriously broken that email programs, calendar apps, address books and so on take the user's info and encode it into some form that is not useable outside (exceptions for standard unix email format) the proprietary program that first touched and "managed" the information. What I plan to hack on some day RSN is representing this info in a non-proprietary format as XML with code to go from/to various proprietary formats. Eventually GUI minded folk can write these tools that take the XML as their native format.
What is the big deal? SOAP is a simple and fairly intuitive way to do object messaging and data exchange across the WEB. It is perfectly useable with Java, CORBA, Python, COM or whatever else ultimately handles execution of the appropriate method on the specified object. It will not replace any of these. It is only a communication protocol. It says nothing about the many services necessary to have a true distributed object environment. Now that the proposal is out there is no way for MS to control who makes use of the basic idea.
Also, CORBA always has been non-proprietary and OS independent.
If we have a common data format (XMLish mayhap) for e-wallets and if we have strong encryption on the content and standardized management protocols then there would be no reason any particular vendor should have a monopoly just as there is no reason for any vendor to have a monopoly on checking accounts. The information should not be stored centrally with Microsoft or anyone else. Some of the new smartcards that are readable wirelessly by computer might be a choice for keeping this info out of centralized hands.
As a geek I am quite tired of the notion that any company owns me or my brain. At most companies the infrastructure level that I am most interested in hacking/improving and the general software environment are actually pretty common to tons of specific companies and applications. I would much rather develop improvements at company A and see them implemented and then go do something similar at company B and spread the knowledge to others to disseminate the techniques more quickly. Being the property of a particular company generally limits the good stuff to that company's property. This is wrong. Also being attached permanently to one company subjects one to cycles where they want to use the talents of their geeks to fight current fires without giving the geeks room to deal with what the really important issues are. This frustrates mightily. I am not sure precisely how to structure my career yet but I do not wish to ever be in a position again where I produce something that I truly believe in only to have one company own it and mothball it if they don't understand what to do with it and are afraid to spin it off.
Isn't one of the great things that not only are great things happening and we can be part of them but that we can actually be comfortable financially in the process? Sometimes we seem to have bought into the old myth that says to make money is inherently evil.
Someone please correct me if I am wrong, but I think the biggest thing missing on Linux today is a good software component technology. CORBA does part of the job but as far as I understand it (and my knowledge is limited here) the equivalent of INPROCs either can't be done or are seldom done and there is a bit more overhead involved in CORBA use than in COM in the MS world. Without a good component technology and reasonably sound interface specifications I am afraid that continual bifurcation and confusion of libraries will result. Tied into this problem is another area I don't have the full skinny on in Linux yet. The ELF binary format supports real DLLs allowing full runtime dynamic linking to the best of my understanding. Yet most instructions on incorporating feature XYZ into my application ask me to linking against some libXYZ.so. With true dynamic linking this is not really necessary and simply creates another dependency to be broken later on.
Is there a general plan for good lightweight component software that takes advantage of true DLLS (among other things) on Linux? Is this plan at a central enough level that it will be generally available rather than dependent on a particular development flavor like KDE or GNOME or a particular Linux distribution?
WHAT?? I've been in this business for over 2 decades and I know for a fact that there were no "good ole days" in software. And any designer/programmer who has done a moderately large software system knows why object oriented technologies are a blessing. Of course they are more often than not not used well but this isn't the fault of OO per se. I more blame the fact that making better programming tools/environments/methodology just isn't a high profit proposition. So we end up automating everyone's work but our own.