Hi!
Happy Monday! Why is Science Fiction called out as its own category in the Duey Decimal system? Joseph Campbell is my understanding. I'm not sure how many years ago Campbell died but Science Fiction had no torch passing, unfortunately. Heinlien, Herbert, Daw, Zelazny all had ties to Campbell. Mainly in Campbells editorial review of Sci Fi magazines. Who buys Sci Fi magazines these days? Or Sci Fi stories? Roger Zelzany started his own publishing company because he said back in the 1980's the book publishers cracked down and declared a 300 page minimum per book. Prior to that many "books" were nothing but elongated Sci Fi stories or a collection of stories. These books were only 100 pages and sold for $2. Campbell required research into mythology, religion, history and science as the foundation for a good Sci Fi story. Unfortunately, that's a lot of work. But you can still find it. Try out one of the later series by Jerry Pernoule and Larry Niven. Lots of science and history. You learn something via a good Sci Fi story. Essentialy, Campbell had a Sci Fi community established that has since dissolved into history.
Hi! Happy Sunday!
Ok, here is what I recall from the lecture.
Registry problems.
Security is at the file level. Any software that needs installed needs to update the registry. Either the registry has to be world writable or the registry is only writable by Admin? How does a non-Admin user update the registry when installing software? For this reason, some companies don't allow users to install any software on MS machines. Any application which updates the registry can overwrite any an all settings in the registry.
DLL model. Operating system DLL's are expected to be updated by applications. For example, a very common problem is that a game program will need to update the sound or graphics DLL. However, the new DLL will not have been tested with other pieces of the useer's system, including other games. Countless friends of mine have trashed the sound or video on their machines by installing games. With regards to Unix application software doesn't install a new version of X or sound drivers.
The problem with the Apple model in general is known in UNIX as the "set uid" problem which does exist in Unix. Programs can be designated to always run as a Admin user by anyone. For many reasons this is weak security (I won't elaborate) and most admins perfer the "sudo" model where a user is allowed to temporarily become Admin for a command using sudo. The distinction is that the use of the command via sudo is under finer granularity of control. However, sudo is a security problem and in general admins should limit its use.
Are you a security expert? I'm not. However, I did attend a security lecture by security experts where Unix (in general) was compared to Windows (in general) and the conclusion was that Unix (in general) was far more secure than Windows (in general). Why? The DLL model is insecure. The registry model is insecure. The user groups (Admin vs. Local) are insecure. By default, Windows still enourages the initial user to have Admin privileges and by default Unix does not. AS someone who runs Linux, my "user" account has normal priveleges and I use "su" and "sudo" for those times I need admin privileges. Windows doesn't even have "su" or "sudo" capabilities. You have to log out and log back in? Most reasonable people grow tired of this in no short order make themselves Admin. In UNIX, I find su and sudo very reasonable to work with and have never made my user the root equivalent. The list of *fundamental* differences went on and on. Never once during the lecture was a distinction between various *nix flavors and Windows flavors made. Bottom line? Unix was designed from the get-go with security in mind. Windows is patching a fundamentally insecure system.
Cheap research with tax payer matching funds
on
Big Company on Campus
·
· Score: 3, Interesting
Over the years UNIX has benefited greatly from the fact that Universities like Berkeley, MIT and Stanford published research because BSD was wide open.
In 1996, when I was a grad student at Berkeley in CS, Microsoft approached the Profs at Berkeley with the source code for NT. The idea was that Berkeley would do research on NT. Amazingly enough the proposal was considered. Rumor was, and I don't know this for a fact, that the only reason the deal fell through is that while Microsoft was willing to release 100% of the source, they weren't willing to relenquish copyright. Derived worked would be owned by Microsoft, even when published. Berkeley said no.
It is interesting then that Microsoft wants research done on.NET.
"The company concluded that to make.Net a success, it had to get academics involved. Not only would their imprimatur lend credibility to the technology, Microsoft would benefit from their technical expertise."
This is just euphumism for buying cheap research. While $500 million dollars may seem a lot, its nothing compared to the 4 billion of internal expenditure. What are they getting for that 4 billion? My bet would be that if University profs and students start innovating on.NET then that $500 million will pay out much more rewards then the 4 billion of internal dollars. Microsoft is just wanting cheap yet better research with tax payer matching funds.
Ha! If we want to talk conspiracy theories, my favorite is that Microsoft is behind SCO. The SCO lawsuit, then, is a conspiracy by Microsoft against Open Source.
Baahh! You say! Here's the Evidence:
So far, the only major company [August 8, 2003]
to publicly endorse SCO's claims by taking a license is Microsoft, the company that analysts say has the most to lose from the spread of Linux.
IBM used the notion of FUD, once coined for IBM but now synonymous for Microsoft, to finger Microsoft as the culprit via this usage:
In an internal memo obtained by internetnews.com targeted for IBM's sales force, Bob Samson, vice president of systems sales in IBM's Systems Group, discussed his company's thrust behind the SCO suit.
"We see no merit in their claims and no supporting facts," Samson said. "Significantly, IBM counter sued SCO on a range of issues. Simply put, SCO's scheme is an attempt to profit from its limited rights to a very old UNIX operating system by introducing fear, uncertainty and doubt into the marketplace."
There you have it. Microsoft is SCO.
Have you indulged in your favorite conspiracy theory today?
Cheers!
-Mybrid
Hi!
Happy Sunday! The founding fathers had a lot to say about freedom. One thing they generally agreed upon is that "free" doesn't mean free from responsibility. You have resposibilities as a citizen based on services rendered to you by the state. This is why you have to pay taxes. You are not free of this responsibility. Yet, if you read the GPL it states " This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE."
Imagine, if you will, elevators were shipped with such a "free" notion? Who would ride in such an elevator? Imagine if a Nuclear Power plant used GPL software. Paul Aoki has a link on his website where he states, "Prototypes have a life of their own. Some University POSTGRES applications that gave us pause." and he goes on to list:...that were out of the ordinary. (These are not urban legends, these are all based on bug reports or other support requests we received first-hand.)
" * cruise missile "threat assessment" system (Johns Hopkins APL / U. S. Navy Tomahawk Program Office, 1991) " * "evaluation of automatic target recognizer (ATR) algorithms" (U. S. Army Night Vision & Electro-Optics Directorate, 1993) " * "a jet engine measuring system" (General Electric, 1993) " * "an asteroid detection project, which aims at discovering earth-grazing asteroids which are potentially dangerous for the Earth." (Observatoire de la Cote d'Azur, 1993) " * "Geoinformation Systems for the problems of the Chernobyl accident" (TechnoSoft, Ukraine, 1992) and other unspecified applications (Russian Nuclear Safety Institute, Moscow, 1993)" (http://db.cs.berkeley.edu/~aoki/.admin/pg apps.htm l) Now, imagine Chernobyl melts down because of some GPL software? Is this an appropriate use of the term free, free of responsibility?
But things are even more absurd. The Open Source community would have everyone believe that software shipped without any warranty is more secure than software shipped with warranty? Really? If a company warranties its software for fitness and could be sued for the melt-down of Chernobyl, then said company would supply less secure software than Open Source who is free of such a responsibility? Postgres is NOT GPL license, btw, it is Berkeley License, but still "free".
What the Open Source community and FSF is asking the public to do is to "trust" without responsibility. Hey, ride this elevator controlled by GPL software, but you can't sue us if the thing breaks. Why should the public do this?
The lawsuit FSF should be worried about is not SCO, but rather the notion that one can ship software without responsibility. While nothing directly in the Constitution states this is illegal, Constitutional writings by the founding fathers make it clear that responsibility comes with freedom. Thus you can't yell "fire" in a crowded movie house. The question remains, can you write "free" software that causes fire in crowded movie house and not be sued?
"Hey, get in our GPL/Open Source elevator, we are screaming at you that its safer than the one under warranty. Money is not an incentive. The fact that you can't sue us for bad code has no bearing on our motivation to write quality code."
Hi!
Happy Friday! I'm not a security expert. However, I have read security articles and MS always pales in comparison to any Unix flavor.
Remember the Cisco virus a couple of weeks ago? What OS do you think all the Cisco routers run? What percentage of market share does Cisco have? Imagine, a world where Cisco routers ran on Windows? How often do you hear of Cisco worms? If Cisco ran on Windows?
Fact is, when companies want something reliable and secure Windows is not the first choice.
This re-design is not technologically feasible or even possible at the enterprise level without (1) a high degree of design coordination, (2) access to expensive and sophisticated design and testing equipment; (3) access to UNIX code, methods and concepts; (4) UNIX architectural experience; and (5) a very significant financial investment.
Well, I guess somebody should tell that to Bill Joy when he rewrote Berkely UNIX from scratch in one summer in the early 1980's. The "UNIX Fast File System", TCP/IP and other Berkeley UNIX innovations didn't come from "well trained UNIX engineers". It came from professors and students at Berkeley and eventually every University in the country was doing research on UNIX and contributing to it. What about Ingres, the first relational database? Industry or research? Let's see, RISC architecture and RAID storage came from: 1.) Industry or 2.) Dave Patterson and students at UC Berkeley.
Hi! Mmm, something as complex as this cries out for a usability study and task analysis. Are all entries equally likely to be selected? One of the things that annoys me with many web sites that want your address is that "USA" is alphabetized as a country selection then it ends up at the bottom of hundreds of entries. Make the common case fast as they say. If 90% of the time "USA" is selected then 1.) USA should be the default and 2.) It should be at the top of the list. Only you know the tasks your users are performing. Odds are the 80/20 rule applies where 80% of the time only 20% of the possibilities are used. The 20% should then be at the top so to speak.
But without further context, its hard to say what the solution is to this.
Excellent responses. Thanks! 1.) Dialects of XML (including XHTML) are just as readable as older style HTML.
This can be said of ANY language if the dialect is a small enough subset. However, XHTML doesn't allow for mistakes. For example, what if I have a start tag but no end? Something a human might do but XML doesn't tolerate. One can argue that HTML is plagued with malformed HTML, but the flip side is that make its more accessible to humans. HTML is forwards compatible, undefined tags are ignored. Try that in XHTML.
1.) Tab delimted files are horrible are representing hiearchical relationships.
Tab delimited files don't represent any relationship. It's just a container for data represented in a text format. The relationship is seperate from the data itself, where it should be. The relationship is stored in a schema somewhere. This points to a problem with XML data, the schema, or parts of it, is shipped with every record in the form of markup. Even tab delimited files are inefficient because if an integer is one byte, effeciency would dictate it is laughable to ship one byte of data represented as ASCII text where the integer 10000 takes 5 bytes instead of only one. If efficiency for data is desired, then binary representation is the only choice.
1.) I'm hoping this is a Troll. Not at all. There are two disparate tasks XML claims to integrate that are a.) semantic schema information of data and b.) presentation formating. XML markup of the data itself is inefficient for delivering mass quantities of data if you just consider the binary->text coversion and disregard markup. With regards to layout, fonts, borders and presentations, Postscript is much richer.
As Socrates once quipped, "the unXamined life isn't worth living." Let's examine the life of HTML. HTML has a distinct advantage over any XML derivative: it is human readable and understandable. Children can learn HTML . The popularity of the WEB is in huge part due to the accessiblity of HTML by a huge audience that is far, far larger than just professional programmers. Quite the opposite for XML.
I argue XML is a broken dead-end. 1.) XML is not readable by any audience other than computer professionals. 2.) Since XML is for computer professionals then then it begs the question: does it fit that audience? I argue it is too cumbersome and verbose and therefore an eventual dead-end. Why do I need to see <phone> 655-7970 </phone> everytime? Just give me the schema and a tab delimited file of records that fit the schema?
If the idea is to give greater flexibility for document presentation, you'll want a PDF format or a fully compliant SGML implementation via a TOOL that doesn't require you to understand the SGML or PDF language itself. When was the last time you interacted and editted PDF directly? Why should you? The only person who needs to understand the SGML or PDF is the developer of the TOOL.
Why don't we replace XML with Postscript? Especially if I'm dealing indirectly with the language via some tool. After all Postcript is Turing complete and can excercise the full capability of the CPU and machine, something XML is not.
Mmm, some companies have different application clients. Business logic can arguably be done in the database itself or in a middle tier share by all clients. A non-trivial database form interaction that I've written sends all the raw strings from the posted HTML input back to the database with no client business logic. The server post scripts simply marshalls the user input into the database call. This has a couple of advantages: 1.) Client simply collects user input and passes it to the database. 2.) Business logic is owned by the database. 3.) Thin client such as handheld devices and telephony applications can share the same business logic. 4.) Data validation happens where it has too anyway -- in the database (PL/SQL code).
Of course this requires a round-trip to the server so some very simple checking for obvious stuff is done in Javascript on the client.
Finally, if your editing serveral tables on one form one questions the design of your application. Does your user understand it?
A WebBrowser and HTML forms should not be used for complex data sets as this was never the intent. Something with application triggers would be more suitable; a 4 GL form language like Power Builder suitable to the task should be used.
Take a tip from Albert Einstein, "Make something as simple as possible, but no simpler."
The current state of affairs is that most Web App servers rely on JSP's. Are we going to embed XForms in JSP's? IBM's WebSphere can not serve up raw XML and convert it to HTML due to performance issues if your web site has any real traffic. Precompiled JSP's are the only game in town. What if the precompiled JSP has XML/XForm?
XML is still not delivered to browsers in general -- we still rely on HTML converted from JSP/XML on the server. Introducing XForms is putting the cart before the horse. XML should be the lingua franca for serving up web content before building even more on top of it.We haven't learned how to walk with delivering native XML to the browser to spend any energy on yet another extension of XML.
Finally, we all know Microsoft extends everything. Given they own 90% of the browser market you can't claim anything till Monopoly State weights in. This is just a headache waiting to happen.
What we have here is a tug of war between skilled doctors of the present who want to continue the knowledge gap they enjoy over patients, especially self-directed patients, and ordinary patients who want to have complete control over their bodies. It's a contest of wills. The trend of healthcare has been to further and further decentralize power by putting tools in the hands of patients at the expense of doctors, and designing those tools for ease of use for those with less and less training GoboHealthCare seems to get that.
In all cases with technology, aspects of it evolve from requiring an expert into that of individual use. Of course there are experts who do not want to relinquish any control (or money). However, I don't believe anyone would want to live without doctors or similarly computer experts.
HCI teaches that "task analysis" is the perspective of human/computer engineering. Designing the file system to meet "tasks" of the user is not such a high priority in the HCI world. Instead, I feel it is far more important to improve "Find" or "Search" so that these tools not only inspect file names but content within, aka Google, as default behavior. GoboLinux should focus on a Find and Search dialog and leave the existing file system intact.
Hi!
Einstein stated the first law of sotware engineering, "Make something as simple as possible, but no simpler."
There is some notion in some of the threads here that perhaps a file system can be made simpler than it really is. The file system is a very complicated roadmap to the computer.
Anyone who claims that something with hundreds of thousands of files can be made "simple of enough for user who doesn't know computers" is sadly misdirected and misinformed.
The second law of software engineering is "you have to be careful if you don't know where you are going because you might not get there." -Yogi Berra
Hi! Happy Tuesday! I have to say that estimation of projects is still 100 lines per day per developer. As hardware has gotten more sophisticated so have software tools. In the end we are right back where we started, 100 lines of code per day. Call this Brook's law, "The advancements in hardware and software offset each other such that productivity remains a constant 100 lines of code per day per developer."
Hi! Happy Tuesday! C++ easily allows one to overlaod the "++" operator or any other operator. However, there is a difference between theory and practice and in practice overloading operators like "++", "[]", or "=" tends to cause more confusion then add readibility. This is especially true when such an operator can through and exception. I'm of the opinion that the book review and postings illustrate the underlying problem with object oriented programming: most people just don't get it and it is usually implemented so poorly the benefits are rarely gained (as opposed to using procedural languages like C).
Slashdot Mirror
Hi!
Happy Monday! Why is Science Fiction called out as its own category in the Duey Decimal system? Joseph Campbell is my understanding. I'm not sure how many years ago Campbell died but Science Fiction had no torch passing, unfortunately. Heinlien, Herbert, Daw, Zelazny all had ties to Campbell. Mainly in Campbells editorial review of Sci Fi magazines. Who buys Sci Fi magazines these days? Or Sci Fi stories? Roger Zelzany started his own publishing company because he said back in the 1980's the book publishers cracked down and declared a 300 page minimum per book. Prior to that many "books" were nothing but elongated Sci Fi stories or a collection of stories. These books were only 100 pages and sold for $2.
Campbell required research into mythology, religion, history and science as the foundation for a good Sci Fi story. Unfortunately, that's a lot of work. But you can still find it. Try out one of the later series by Jerry Pernoule and Larry Niven. Lots of science and history. You learn something via a good Sci Fi story. Essentialy, Campbell had a Sci Fi community established that has since dissolved into history.
Security is at the file level. Any software that needs installed needs to update the registry. Either the registry has to be world writable or the registry is only writable by Admin? How does a non-Admin user update the registry when installing software? For this reason, some companies don't allow users to install any software on MS machines. Any application which updates the registry can overwrite any an all settings in the registry.
Cheers! -Mybrid
Are you a security expert? I'm not. However, I did attend a security lecture by security experts where Unix (in general) was compared to Windows (in general) and the conclusion was that Unix (in general) was far more secure than Windows (in general). Why? The DLL model is insecure. The registry model is insecure. The user groups (Admin vs. Local) are insecure. By default, Windows still enourages the initial user to have Admin privileges and by default Unix does not. AS someone who runs Linux, my "user" account has normal priveleges and I use "su" and "sudo" for those times I need admin privileges. Windows doesn't even have "su" or "sudo" capabilities. You have to log out and log back in? Most reasonable people grow tired of this in no short order make themselves Admin. In UNIX, I find su and sudo very reasonable to work with and have never made my user the root equivalent. The list of *fundamental* differences went on and on. Never once during the lecture was a distinction between various *nix flavors and Windows flavors made. Bottom line? Unix was designed from the get-go with security in mind. Windows is patching a fundamentally insecure system.
Over the years UNIX has benefited greatly from the fact that Universities like Berkeley, MIT and Stanford published research because BSD was wide open. In 1996, when I was a grad student at Berkeley in CS, Microsoft approached the Profs at Berkeley with the source code for NT. The idea was that Berkeley would do research on NT. Amazingly enough the proposal was considered. Rumor was, and I don't know this for a fact, that the only reason the deal fell through is that while Microsoft was willing to release 100% of the source, they weren't willing to relenquish copyright. Derived worked would be owned by Microsoft, even when published. Berkeley said no.
It is interesting then that Microsoft wants research done on .NET.
This is just euphumism for buying cheap research. While $500 million dollars may seem a lot, its nothing compared to the 4 billion of internal expenditure. What are they getting for that 4 billion? My bet would be that if University profs and students start innovating onThe SCO lawsuit, then, is a conspiracy by Microsoft against Open Source.
Baahh! You say! Here's the Evidence:
There you have it. Microsoft is SCO. Have you indulged in your favorite conspiracy theory today?
Cheers!
-Mybrid
Hi! ...that were out of the ordinary. (These are not urban legends, these are all based on bug reports or other support requests we received first-hand.)
g apps.htm l)
Happy Sunday! The founding fathers had a lot to say about freedom. One thing they generally agreed upon is that "free" doesn't mean free from responsibility. You have resposibilities as a citizen based on services rendered to you by the state. This is why you have to pay taxes. You are not free of this responsibility. Yet, if you read the GPL it states " This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE."
Imagine, if you will, elevators were shipped with such a "free" notion? Who would ride in such an elevator? Imagine if a Nuclear Power plant used GPL software. Paul Aoki has a link on his website where he states, "Prototypes have a life of their own. Some University POSTGRES applications that gave us pause." and he goes on to list:
" * cruise missile "threat assessment" system (Johns Hopkins APL / U. S. Navy Tomahawk Program Office, 1991)
" * "evaluation of automatic target recognizer (ATR) algorithms" (U. S. Army Night Vision & Electro-Optics Directorate, 1993)
" * "a jet engine measuring system" (General Electric, 1993)
" * "an asteroid detection project, which aims at discovering earth-grazing asteroids which are potentially dangerous for the Earth." (Observatoire de la Cote d'Azur, 1993)
" * "Geoinformation Systems for the problems of the Chernobyl accident" (TechnoSoft, Ukraine, 1992) and other unspecified applications (Russian Nuclear Safety Institute, Moscow, 1993)"
(http://db.cs.berkeley.edu/~aoki/.admin/p
Now, imagine Chernobyl melts down because of some GPL software? Is this an appropriate use of the term free, free of responsibility?
But things are even more absurd. The Open Source community would have everyone believe that software shipped without any warranty is more secure than software shipped with warranty? Really? If a company warranties its software for fitness and could be sued for the melt-down of Chernobyl, then said company would supply less secure software than Open Source who is free of such a responsibility? Postgres is NOT GPL license, btw, it is Berkeley License, but still "free".
What the Open Source community and FSF is asking the public to do is to "trust" without responsibility. Hey, ride this elevator controlled by GPL software, but you can't sue us if the thing breaks. Why should the public do this?
The lawsuit FSF should be worried about is not SCO, but rather the notion that one can ship software without responsibility. While nothing directly in the Constitution states this is illegal, Constitutional writings by the founding fathers make it clear that responsibility comes with freedom. Thus you can't yell "fire" in a crowded movie house. The question remains, can you write "free" software that causes fire in crowded movie house and not be sued?
"Hey, get in our GPL/Open Source elevator, we are screaming at you that its safer than the one under warranty. Money is not an incentive. The fact that you can't sue us for bad code has no bearing on our motivation to write quality code."
Hi!
Happy Friday! I'm not a security expert. However, I have read security articles and MS always pales in comparison to any Unix flavor.
Remember the Cisco virus a couple of weeks ago? What OS do you think all the Cisco routers run? What percentage of market share does Cisco have? Imagine, a world where Cisco routers ran on Windows? How often do you hear of Cisco worms? If Cisco ran on Windows?
Fact is, when companies want something reliable and secure Windows is not the first choice.
Cheers!
-Mybrid
This re-design is not technologically feasible or even possible at the enterprise level without (1) a high degree of design coordination, (2) access to expensive and sophisticated design and testing equipment; (3) access to UNIX code, methods and concepts; (4) UNIX architectural experience; and (5) a very significant financial investment.
Well, I guess somebody should tell that to Bill Joy when he rewrote Berkely UNIX from scratch in one summer in the early 1980's. The "UNIX Fast File System", TCP/IP and other Berkeley UNIX innovations didn't come from "well trained UNIX engineers". It came from professors and students at Berkeley and eventually every University in the country was doing research on UNIX and contributing to it. What about Ingres, the first relational database? Industry or research?
Let's see, RISC architecture and RAID storage came from: 1.) Industry or 2.) Dave Patterson and students at UC Berkeley.
Nobody I know of takes AIX seriously, do you?
Hi! Mmm, something as complex as this cries out for a usability study and task analysis. Are all entries equally likely to be selected? One of the things that annoys me with many web sites that want your address is that "USA" is alphabetized as a country selection then it ends up at the bottom of hundreds of entries. Make the common case fast as they say. If 90% of the time "USA" is selected then
1.) USA should be the default and
2.) It should be at the top of the list.
Only you know the tasks your users are performing. Odds are the 80/20 rule applies where 80% of the time only 20% of the possibilities are used. The 20% should then be at the top so to speak.
But without further context, its hard to say what the solution is to this.
Cheers!
-Mybrid
Excellent responses. Thanks!
1.) Dialects of XML (including XHTML) are just as readable as older style HTML.
This can be said of ANY language if the dialect is a small enough subset. However, XHTML doesn't allow for mistakes. For example, what if I have a start tag but no end? Something a human might do but XML doesn't tolerate. One can argue that HTML is plagued with malformed HTML, but the flip side is that make its more accessible to humans. HTML is forwards compatible, undefined tags are ignored. Try that in XHTML.
1.) Tab delimted files are horrible are representing hiearchical relationships.
Tab delimited files don't represent any relationship. It's just a container for data represented in a text format. The relationship is seperate from the data itself, where it should be. The relationship is stored in a schema somewhere. This points to a problem with XML data, the schema, or parts of it, is shipped with every record in the form of markup. Even tab delimited files are inefficient because if an integer is one byte, effeciency would dictate it is laughable to ship one byte of data represented as ASCII text where the integer 10000 takes 5 bytes instead of only one. If efficiency for data is desired, then binary representation is the only choice.
1.) I'm hoping this is a Troll.
Not at all. There are two disparate tasks XML claims to integrate that are a.) semantic schema information of data and b.) presentation formating. XML markup of the data itself is inefficient for delivering mass quantities of data if you just consider the binary->text coversion and disregard markup. With regards to layout, fonts, borders and presentations, Postscript is much richer.
Cheers!
-Mybrid
As Socrates once quipped, "the unXamined life isn't worth living."
Let's examine the life of HTML. HTML has a distinct advantage over any XML derivative: it is human readable and understandable. Children can learn HTML . The popularity of the WEB is in huge part due to the accessiblity of HTML by a huge audience that is far, far larger than just professional programmers. Quite the opposite for XML.
I argue XML is a broken dead-end.
1.) XML is not readable by any audience other than computer professionals.
2.) Since XML is for computer professionals then then it begs the question: does it fit that audience? I argue it is too cumbersome and verbose and therefore an eventual dead-end. Why do I need to see <phone> 655-7970 </phone> everytime? Just give me the schema and a tab delimited file of records that fit the schema?
If the idea is to give greater flexibility for document presentation, you'll want a PDF format or a fully compliant SGML implementation via a TOOL that doesn't require you to understand the SGML or PDF language itself. When was the last time you interacted and editted PDF directly? Why should you?
The only person who needs to understand the SGML or PDF is the developer of the TOOL.
Why don't we replace XML with Postscript? Especially if I'm dealing indirectly with the language via some tool. After all Postcript is Turing complete and can excercise the full capability of the CPU and machine, something XML is not.
Mmm, some companies have different application clients. Business logic can arguably be done in the database itself or in a middle tier share by all clients.
A non-trivial database form interaction that I've written sends all the raw strings from the posted HTML input back to the database with no client business logic. The server post scripts simply marshalls the user input into the database call. This has a couple of advantages:
1.) Client simply collects user input and passes it to the database.
2.) Business logic is owned by the database.
3.) Thin client such as handheld devices and telephony applications can share the same business logic.
4.) Data validation happens where it has too anyway -- in the database (PL/SQL code).
Of course this requires a round-trip to the server so some very simple checking for obvious stuff is done in Javascript on the client.
Finally, if your editing serveral tables on one form one questions the design of your application.
Does your user understand it?
A WebBrowser and HTML forms should not be used for complex data sets as this was never the intent. Something with application triggers would be more suitable; a 4 GL form language like Power Builder suitable to the task should be used.
Take a tip from Albert Einstein, "Make something as simple as possible, but no simpler."
The current state of affairs is that most Web App servers rely on JSP's. Are we going to embed XForms in JSP's? IBM's WebSphere can not serve up raw XML and convert it to HTML due to performance issues if your web site has any real traffic. Precompiled JSP's are the only game in town. What if the precompiled JSP has XML/XForm?
XML is still not delivered to browsers in general -- we still rely on HTML converted from JSP/XML on the server. Introducing XForms is putting the cart before the horse. XML should be the lingua franca for serving up web content before building even more on top of it.We haven't learned how to walk with delivering native XML to the browser to spend any energy on yet another extension of XML.
Finally, we all know Microsoft extends everything.
Given they own 90% of the browser market you can't claim anything till Monopoly State weights in. This is just a headache waiting to happen.
What we have here is a tug of war between skilled doctors of the present who want to continue the knowledge gap they enjoy over patients, especially self-directed patients, and ordinary patients who want to have complete control over their bodies. It's a contest of wills. The trend of healthcare has been to further and further decentralize power by putting tools in the hands of patients at the expense of doctors, and designing those tools for ease of use for those with less and less training GoboHealthCare seems to get that.
In all cases with technology, aspects of it evolve from requiring an expert into that of individual use. Of course there are experts who do not want to relinquish any control (or money).
However, I don't believe anyone would want to live without doctors or similarly computer experts.
HCI teaches that "task analysis" is the perspective of human/computer engineering.
Designing the file system to meet "tasks" of the user is not such a high priority in the HCI world.
Instead, I feel it is far more important to improve "Find" or "Search" so that these tools not only inspect file names but content within, aka Google, as default behavior.
GoboLinux should focus on a Find and Search dialog and leave the existing file system intact.
Hi!
Einstein stated the first law of sotware engineering, "Make something as simple as possible, but no simpler."
There is some notion in some of the threads here that perhaps a file system can be made simpler than it really is. The file system is a very complicated roadmap to the computer.
Anyone who claims that something with hundreds of thousands of files can be made "simple of enough for user who doesn't know computers" is sadly misdirected and misinformed.
The second law of software engineering is "you have to be careful if you don't know where you are going because you might not get there."
-Yogi Berra
Cheers!
-Mybrid
In the future, the technologically super-advanced pass themselves off as gods creating a new political agenda, accelerationism.
That's forward thinking.
Hi! Happy Tuesday! I have to say that estimation of projects is still 100 lines per day per developer. As hardware has gotten more sophisticated so have software tools. In the end we are right back where we started, 100 lines of code per day. Call this Brook's law, "The advancements in hardware and software offset each other such that productivity remains a constant 100 lines of code per day per developer."
Hi! Happy Tuesday! C++ easily allows one to overlaod the "++" operator or any other operator. However, there is a difference between theory and practice and in practice overloading operators like "++", "[]", or "=" tends to cause more confusion then add readibility. This is especially true when such an operator can through and exception. I'm of the opinion that the book review and postings illustrate the underlying problem with object oriented programming: most people just don't get it and it is usually implemented so poorly the benefits are rarely gained (as opposed to using procedural languages like C).
What a trip. I'm impressed that our government spent money wisely. I'm all for this kind of stuff. Truly, truly impressive.