I remember Scotty explaining to LaForge one day that his secret was to always exaggerate the time it takes to do something by a factor of eight. Usually this made him look like a hero, but he always had that built in time for dealing with the unforeseen problem. It was one of the more important lessons I learned from Star Trek.
But the real value that James Doohan brought to his role, and that I value from his life was his enthusiasm. The positiveness he brought to his role came from the heart. He embodied and lived the "Can Do" attitude. Hopefully it was every bit as infectious as the rest of Star Trek and we will all remember that we too 'can do!'
I would have thought that with the holes they were going for a swiss, but the color says cheddar. I've never seen Google have anything questionable in their art before this, but if it had been up to me I would have stuck with a real swiss image. If they really wanted to be clever about it, they could have image mapped the holes to the craters for that added surrealistic effect!
As a child of the 60's-70's and expatriot for a few yers there I can honestly say that while the re-runs of Star Trek and Hogan's Heros were nice, that the world of BBC & ITV really opened my eyes to the possibility of what television could be. Those few of us who have enjoyed the site and shared it with our friends and our children have known for some time that this is a precious site, an important part of our cultural heritage and one of the few shining fortresses in the vast wasteland of planetary television. How unfortunate that BBC has chosen to turn it's back on the those of us who can really appreciate it. Let's hope that they are going to offer the content on a series of data CD's or DVD's. One day I'd like to enrich the veiwing experience of my grandchildren (if I should be so lucky!) with information like this. It was a real eye opener for my son to learn more about Gerry Anderson, The Prisoner and all those PCS reruns of Dr. Who. Let us further home that the BBC realizes that those few 700,000 are people who care, and appreciate both the content and the effort to bring it forward in such a nice manner.
I've been a proponent of cell phones for the deaf for a long time. Unfortunately the phone companies just don't get it. The idea is that they would need a plan with no minutes, but lots of text messaging. They can use sms instead of relay calls to customers who can accept sms, and use the relay call network for others who don't.
f the blindness and/or deafness is sensory organ related as opposed to brain related it may be possible to explore the concept of meta-visual and meta-audio metaphors. Take, as a starting example the 'phantom limb' effect. In this example, the user no longer has a limb, but frequently experiences the 'presence' of the limb. People who experience this can give you a detailed 'impression' of what is happening with the missing limb. Clearly, in these cases, the related area of the brain is fully capable of functionality.
Recently, reports of the Utah Hand show examples where the nerves of the missing arm are re-routed, subcutaneously to the chest, where they can both exert and receive electrical signals, allowing for both command and feedback impulses to be transmitted to/from the brain.
If we look the example of a person blinded in later life, they have a rich history of visual experience in which to draw upon, which can be accessed through both tactile and auditory senses if present. Imagine a tactile display, about the size of a display screen, in which there are tiny, air inflatable or piezoelectric cells which can be mapped above membrane switches. The tactile surface could express braille and represent the controls available on the normal display via a driver and mapping software. The locations of the items on the display would represent the windows of the open applications. A window could also represent a tactile keyboard. Now that we have an input, output device, we can enrich the environment further, by adding voices to relay the bulk of the information. I say voices, because the applications should speak in different voices so that the user has an audio metaphor to represent different applications and tell them apart. By using 3-d surround sound acoustical technology, an applications voice could appear to be spatially located. A window with a message in the background may be quieter. And the voices should have different intonations depending on what is being done, perhaps a slightly more stern or flat voicing for a dialog box, and something more friendly for informational messages.
The operating system differs here, because it displays the only the controls on the tactile pad, and displays the application content via the sound system and a synthesizer.
By modularizing the various application flows and dialogs, the system should be configurable for a variety of differently challenged individuals. Similarly, the same PC should be able to inter operate over the net with other similar and dissimilar devices. So a deaf poet could work with a blind musician to create a song, and their producer could listen and make suggestions in real time.
I recommend Slashdot to my friends and I'd be the first one to say that I enjoy Slashdot. And I'll admit that occasionally it's more than an entertainment. There are articles and comments that really do deserve moderations like Insightful, Informative and Funny. And there is 'other content' which is patiently moderated Offtopic or Flamebait. I think that it is the principle responsibility of Slashdot to insure that the entire spectrum of freedom of speech be maintained, without giving in to any outside pressures from the greater journalistic public, which is most certainly, by in large, run at the editorial whim of a very few people. People who's self appointed task it is to 'select what is important' (from their own personal point of view) and, in so doing, to deprive us of the experience of the world at large.
If there is one thing that I can say of a Slashdot reader, it is that that reader has the freedom to chose what they want to read and how they want to interpret it, rather than the 'pre-digested' and outright biased reporting that is available from the media at large. This openness is the key to developing the independent, 'out of the box' thinking; the generalists of the evolving age of Information and Knowledge.
So kudos to Slashdot and their outspoken and many faceted readers.
Even if they 'automagically' re assemble every molecule, state, position and kinetic energy; there is no opportunity for the consciousness to be transferred.
Even compartmentalized, the fanny pack requires sliding the 'pack' portion around to get access to it, and then usually some digging around in order to find what you were looking for. The Utility Belt has a separate compartment for each item, and there is no need to slide the belt around. Believe me, I use a fanny pack to lug the stuff around, but it's no where near as efficent as a real Utility Belt would be.
The one feature of both Batman and Star Wars was the Utility Belt. We have Cell Phones, Pagers, PDAs, Music Players, Multi-Tools, and the list is growing. It will be quite a while before we can expect an all in one tool that will do it all. It takes a while after dressing to load up the belt with all this stuff and cram the extra media in the pocket along with wallet and keys. I've often thought that it would be nice to just have a Utility Belt that could hold all the stuff and be ready at a moments notice.
On a somewhat related matter, the physician's vest in Event Horizon also seems like a good idea. I'm considering a 'Red Green' and modifying a fishing vest to accommodate the usual tools I use for PC repair for family and friends. Just slip it on and I'm ready to tackle hardware and software problems!
Most computer science courses are designed at the lower levels to weed out the undesirable students. In the process students are usually forced to learn things that they will never use.
I would suggest that she think more along the lines of what she wants to do as a software engineer. There are a wide range of possibilities, and all the schooling in the world will do little more than get one's foot in the door.
By determining what actual work she wants to do with the degree she will be in (IMHO) a much better position to identify or assemble the plan that will both meet her needs and empower her to make substantial progress in real world (outside of school).
Nothing stimulates the mind like working on a problem that is so interesting that it borders on compulsion. Working on a project that one enjoys is a truly uplifting experience, while working on projects that are uninteresting can lead to depression and worse. When she can really visualize what she wants to be working on, it will become very clear what will be necessary to enable her to pursue her dreams.
While I don't mind cutting out the middleman, I think this stinks for the artist. Let the artists set up their own websites and accept a payment equivalent to their royalty. Let's kick out the guys that say who will and will not release music. There's an explosion of new music out there, some of which we may not appreciate, that is just waiting for the opportunity to get listened to. Let's break up the whole cartel, the RIAA, the radio stations and anything else that stands in the way of the freedom of musical expression, which ought to be covered in the 1st amendment.
Your point about the sparse matrix is well taken! That is one of the places where MUMPS excelled. The VA is doing amazing things these days. VistA is a project that is making the 'bleeding edge' a functional reality for many people in medicine.
Well, with Blu-Ray right around the corner, why would I want to invest in the almost outdated DVD of the inferior enhancement process just to turn around and have to buy it again. Now if it were episodes of 'Space Angel' I might have to reconsider!:^)
I remember when I could actually call Mr. Brown himself with crash dumps. During the first few weeks of release for the IBM PS2 version. It has come a long way and has a wonderful toolkit. I really like the visual implementation. It's really wonderful. You guys do good stuff there! For those who are interested in MUMPS I agree they should check out your site!
I can concede the most of the implementation VS model points, no problemo.
Either you understand your data or you don't. If you do, then you can express it with the relational model, which is the only complete model for the manipulation and retrieval of data. If you don't, then it really doesn't matter what you use.
Here too, expressing it in the 'model' is one point, but the implementation is another. I can actually build a mumps database in a relational database half dozen fields.
A unique numeric node ID
A pointer representing the ID of the parent node above this level
A pointer representing the ID of the previous block at this level
A pointer representing the ID of the next block at this level
A pointer representing the ID of the first Child node below this level
The Data Content of the node, represented as a string, which contains internal segments representing the subscripts or the data elements of a particular 'global'.
Top level directory blocks have '0' parents, bottem level data blocks have no children, etc.
I remember the 'root' block of a DSM MUMPS implementation would store the pointers to the first data and routine blocks in block 0, but it was referenced as block -16777216. Because the operating system was stored after the boot sector it was not unusual to see pretty big offsets to the first data and routine blocks. As I pointed out before the routine blocks were pretty normal. The top level held the list of routine names, each included a pointe r to the first block of the routine's text. If the routine overflowed the block, the 'next' pointer pointed to the next block of it.
The 'globals' were a bit different. First off the array index used key compression. The first array index in the block was complete. If there was any overlap in the index between the one before and the one after, it was indicated by a flagged numeric, for example if 'ABCDE' was followed by 'ABCDF' it would be stored as '0ABCDE'{pointer to block containing ABCDE's data}{delimiter}4F{pointer to ABCDF's data} etc.
In Mumps, if we stored these in global M these would be coded as:
The implementation was very precise, the containers wasted very little space, etc. A global that contained little or no data would contain its entire structure in 1 or 2 blocks. With 16K blocks that is a max of 32K for an empty global. It would absorb more than 24K of data in most cases before a block split was required, and the structure required no meta data other than the actual global storage representation itself. An empty database implementation of any RDBMS, say Access for example) is a huge waste of space. 92K just to have the empty container structure. Pointers to separate and segregate tables and fields special formatting and handling of each data type.
Enough of implementation, anyway. My chief complaint is that the relational model is inflexible in it's implementation by definition. MUMPS provided the programmer the ability to navigate the data in a tree format. If you wanted to just 'tack' something in somewhere, just like creating a new synapse in your brain, you could. By convention you could explore and extend it as a tree. If special code was needed to access the new data, it could be included in the data structure itself, just as it can in your brain. I see the RDBMS guys just throw up their hands as the number of tables in their databases become unmanageable and that they have to 'just say no.' to projects that just don't fit their 'understanding' of how their implementation requires them to do things. As long as I stick to object stores for data that 'fails to normalize' properly and continues to grow exponentially I do fine. I find that RDBMS is great for large collections of things that are all basically the same, and I routinely use them for those tasks. But for working with an extensive amount of dissimilarly structured data it's about as effective as ray tracing in COBOL. Yes, it can be done in the model and in the implementation, but who would want to maintain it?;^)
MUMPS didn't lack data integrity. It just didn't follow the relational model, as you pointed out. Mumps was based on binary trees. In fact the whole big thing was one big tree. At the top were two blocks, one for globals and the other for routines. The routines were the simple one, just like a file directory. The data on the other hand was pretty damn complex, with top level, middle level and data level blocks.
Your problem with external indexing would seem to indicate that your global structure could have been better. I have never seen a good mumps structure that failed to represent a self referential index when ordered through at a particular level.
As for the whitespace, well I adapted quite well to Python, I guess, in part because of that decade of MUMPS.
In any case it is not the language, but the structure of the mind that can appreciate it that makes a programmer productive. You seem to feel pretty comfortable with the relational model. When I want to do something like catalog my music, I don't want to only store a name and an artist and a song, I want to explore the influences, histories, etc, call it multideminsional model. It's the same with medicine. There are nuances which the relational model will never be able to accomidate. Sure it's good at what it does, but it will never think like a human being. And that's what we're all going to want sooner or later.
II respectfully disagree. I used MUMPS for a decade before switching. If you look at the MUMPS command structure, the pattern matching and the ability to use indirection, you will see that an even finer level of control was achieved. For example, using a relational database, containing different structures, some finite, like an integer and some not so finite, like a blob, the relational database has to make a much larger number of disk hits just to locate the data. You also don't have to return the entire result of a query before being able to act on records. In MUMPS you can advance through the records one at a time to process them, again much less work for the system to achieve the same goals. With MUMPS and 16K of memory on the server and a VT100 as the client, you could do in text mode, things that you have to have a 1GHz+ PC and web browser to do. I've dismantled record crashed from RDBMS and MUMPS systems, the MUMPS is much easier to recover, the RDBMS you're pretty much screwed. MUMPS hasn't been made obsolete either, just look at CACHE, it's nothing but MUMPS underneath. RDBMS look great when all you're dealing with is similar in nature, a table of this, a table of that. Now let's record your medical record in it' suddenly you discover that for dissimilarly structured data the RDBMS is about as effective as a bazooka at a watchmaker's shop. Someone sold you a bill of goods my friend, time to think outside of the box.
MUMPS had one data type, a string and one data structure, an array. Commands abbreviable to a single letter. MUMPS ran whole Hospital Information Systems on DEC PDP-11's where MUMPS was both the Operating System and the Language and was called Digital Standard MUMPS or DSM for short. It was way ahead of it's time, it predated SQL, and was one of it's progenitors. Anyone who offers a survey of databases and fails to include it is seriously myopic. Google 'MUMPS Language' for the tip of the iceberg.
As for the future, I think that while OO databases are all the vogue, that O databases, where each bit of information is represented by it's own object and will have the ability to demonstrate autonomous agency is a good area for research. Then we can just let the data speak for itself!;^)
"Every person takes the limits of their own field of vision for the limits of the world." - Arthur Schopenhauer
I've been using LZMP for about three years. Linux Zope MySQL and Python. The whole thing results in a JavaScript rendering of the task that needs to be accomplished. The Client runs it in their browser and the server gets back the result. If a third party is involved that 'absolutely needs' XML we can accommodate them, but otherwise there has never been a compelling reason to go through the overhead of wrapping the data in XML just to unwrap it on the other end.
I'm heavily in favor of having the client do the brunt of the work, and letting the server just dish it out and file it. Servers are still pretty expensive, and most PC's are above 1 GHz these days, so speed is a lot less of an issue. Another issue is that I like pushing the minimum amount of code and data out to the client that will get the job done. I hate it when someone has to load an XML driven ActiveX or Java client tools when all they need is an input field.
And pushing all that 'fancy tool code' is usually where security becomes an issue. "Hey it was there on the toolbar, I just dragged and dropped it onto the layout of the web page and hooked it up to the database table field..." I've heard that one before. Notice there's no really effective version of Moore's Law for software? That is other than this:
"While the hardware will be twice as capable and half the cost, the software will double in cost and be half as effective, so there's really no progress at all!"
Slashdot Mirror
A potential base for mining the Oort Cloud? I wonder if Wan is already there?
"Microsoft has received a patent for Communications. We are now no longer
allowed to communicate without a license."
There is a commotion outside.
"What's that sound?"
The door explodes inward and the room is immediately filled with smoke and
overrun with stormtroopers.
A large and ominous voice booms out.
"You are forbidden to communicate."
I am stunned! I don't know what to think!
Then just as suddenly the first wave of stormtroopers are felled, one by one in
an unimaginably short flash of time.
"Don't know what to think a voice says?"
I look up and see Jeff Brazos towering above the carnage.
"That's fine by me. I just patented Thought!"
i $l(exc,"""") k=$p(exc,"""",2)
s k=$l(exc) i $e(exc,1,1)="""",$e(exc,k,k)="""" s exc=$e(exc,2,k-1)
is a really bad way to do it. It would be better to do it like this:
i $l(exec,"""") k=$p(exc,"""",2)
it would be better to pick good examples, not crappy ones.
Don't bash a language that you don't understand.
But the real value that James Doohan brought to his role, and that I value from his life was his enthusiasm. The positiveness he brought to his role came from the heart. He embodied and lived the "Can Do" attitude. Hopefully it was every bit as infectious as the rest of Star Trek and we will all remember that we too 'can do!'
I would have thought that with the holes they were going for a swiss, but the color says cheddar. I've never seen Google have anything questionable in their art before this, but if it had been up to me I would have stuck with a real swiss image. If they really wanted to be clever about it, they could have image mapped the holes to the craters for that added surrealistic effect!
Just be sure not to let it take over the world, OK?
God Save The Queen!
I've been a proponent of cell phones for the deaf for a long time. Unfortunately the phone companies just don't get it. The idea is that they would need a plan with no minutes, but lots of text messaging. They can use sms instead of relay calls to customers who can accept sms, and use the relay call network for others who don't.
</off-topic>
Recently, reports of the Utah Hand show examples where the nerves of the missing arm are re-routed, subcutaneously to the chest, where they can both exert and receive electrical signals, allowing for both command and feedback impulses to be transmitted to/from the brain.
If we look the example of a person blinded in later life, they have a rich history of visual experience in which to draw upon, which can be accessed through both tactile and auditory senses if present. Imagine a tactile display, about the size of a display screen, in which there are tiny, air inflatable or piezoelectric cells which can be mapped above membrane switches. The tactile surface could express braille and represent the controls available on the normal display via a driver and mapping software. The locations of the items on the display would represent the windows of the open applications. A window could also represent a tactile keyboard. Now that we have an input, output device, we can enrich the environment further, by adding voices to relay the bulk of the information. I say voices, because the applications should speak in different voices so that the user has an audio metaphor to represent different applications and tell them apart. By using 3-d surround sound acoustical technology, an applications voice could appear to be spatially located. A window with a message in the background may be quieter. And the voices should have different intonations depending on what is being done, perhaps a slightly more stern or flat voicing for a dialog box, and something more friendly for informational messages.
The operating system differs here, because it displays the only the controls on the tactile pad, and displays the application content via the sound system and a synthesizer.
By modularizing the various application flows and dialogs, the system should be configurable for a variety of differently challenged individuals. Similarly, the same PC should be able to inter operate over the net with other similar and dissimilar devices. So a deaf poet could work with a blind musician to create a song, and their producer could listen and make suggestions in real time.
If there is one thing that I can say of a Slashdot reader, it is that that reader has the freedom to chose what they want to read and how they want to interpret it, rather than the 'pre-digested' and outright biased reporting that is available from the media at large. This openness is the key to developing the independent, 'out of the box' thinking; the generalists of the evolving age of Information and Knowledge.
So kudos to Slashdot and their outspoken and many faceted readers.
Even if they 'automagically' re assemble every molecule, state, position and kinetic energy; there is no opportunity for the consciousness to be transferred.
Even compartmentalized, the fanny pack requires sliding the 'pack' portion around to get access to it, and then usually some digging around in order to find what you were looking for. The Utility Belt has a separate compartment for each item, and there is no need to slide the belt around. Believe me, I use a fanny pack to lug the stuff around, but it's no where near as efficent as a real Utility Belt would be.
On a somewhat related matter, the physician's vest in Event Horizon also seems like a good idea. I'm considering a 'Red Green' and modifying a fishing vest to accommodate the usual tools I use for PC repair for family and friends. Just slip it on and I'm ready to tackle hardware and software problems!
I would suggest that she think more along the lines of what she wants to do as a software engineer. There are a wide range of possibilities, and all the schooling in the world will do little more than get one's foot in the door.
By determining what actual work she wants to do with the degree she will be in (IMHO) a much better position to identify or assemble the plan that will both meet her needs and empower her to make substantial progress in real world (outside of school).
Nothing stimulates the mind like working on a problem that is so interesting that it borders on compulsion. Working on a project that one enjoys is a truly uplifting experience, while working on projects that are uninteresting can lead to depression and worse. When she can really visualize what she wants to be working on, it will become very clear what will be necessary to enable her to pursue her dreams.
Best of luck!
While I don't mind cutting out the middleman, I think this stinks for the artist. Let the artists set up their own websites and accept a payment equivalent to their royalty. Let's kick out the guys that say who will and will not release music. There's an explosion of new music out there, some of which we may not appreciate, that is just waiting for the opportunity to get listened to. Let's break up the whole cartel, the RIAA, the radio stations and anything else that stands in the way of the freedom of musical expression, which ought to be covered in the 1st amendment.
Your point about the sparse matrix is well taken! That is one of the places where MUMPS excelled. The VA is doing amazing things these days. VistA is a project that is making the 'bleeding edge' a functional reality for many people in medicine.
Well, with Blu-Ray right around the corner, why would I want to invest in the almost outdated DVD of the inferior enhancement process just to turn around and have to buy it again. Now if it were episodes of 'Space Angel' I might have to reconsider! :^)
I remember when I could actually call Mr. Brown himself with crash dumps. During the first few weeks of release for the IBM PS2 version. It has come a long way and has a wonderful toolkit. I really like the visual implementation. It's really wonderful. You guys do good stuff there! For those who are interested in MUMPS I agree they should check out your site!
Either you understand your data or you don't. If you do, then you can express it with the relational model, which is the only complete model for the manipulation and retrieval of data. If you don't, then it really doesn't matter what you use.
Here too, expressing it in the 'model' is one point, but the implementation is another. I can actually build a mumps database in a relational database half dozen fields.
Top level directory blocks have '0' parents, bottem level data blocks have no children, etc.
I remember the 'root' block of a DSM MUMPS implementation would store the pointers to the first data and routine blocks in block 0, but it was referenced as block -16777216. Because the operating system was stored after the boot sector it was not unusual to see pretty big offsets to the first data and routine blocks. As I pointed out before the routine blocks were pretty normal. The top level held the list of routine names, each included a pointe r to the first block of the routine's text. If the routine overflowed the block, the 'next' pointer pointed to the next block of it.
The 'globals' were a bit different. First off the array index used key compression. The first array index in the block was complete. If there was any overlap in the index between the one before and the one after, it was indicated by a flagged numeric, for example if 'ABCDE' was followed by 'ABCDF' it would be stored as '0ABCDE'{pointer to block containing ABCDE's data}{delimiter}4F{pointer to ABCDF's data} etc.
In Mumps, if we stored these in global M these would be coded as:
^M('ABCDE')={ABCDE's data}
^M('ABCDF')={ABCDF's data}
The implementation was very precise, the containers wasted very little space, etc. A global that contained little or no data would contain its entire structure in 1 or 2 blocks. With 16K blocks that is a max of 32K for an empty global. It would absorb more than 24K of data in most cases before a block split was required, and the structure required no meta data other than the actual global storage representation itself. An empty database implementation of any RDBMS, say Access for example) is a huge waste of space. 92K just to have the empty container structure. Pointers to separate and segregate tables and fields special formatting and handling of each data type.
Enough of implementation, anyway. My chief complaint is that the relational model is inflexible in it's implementation by definition. MUMPS provided the programmer the ability to navigate the data in a tree format. If you wanted to just 'tack' something in somewhere, just like creating a new synapse in your brain, you could. By convention you could explore and extend it as a tree. If special code was needed to access the new data, it could be included in the data structure itself, just as it can in your brain. I see the RDBMS guys just throw up their hands as the number of tables in their databases become unmanageable and that they have to 'just say no.' to projects that just don't fit their 'understanding' of how their implementation requires them to do things. As long as I stick to object stores for data that 'fails to normalize' properly and continues to grow exponentially I do fine. I find that RDBMS is great for large collections of things that are all basically the same, and I routinely use them for those tasks. But for working with an extensive amount of dissimilarly structured data it's about as effective as ray tracing in COBOL. Yes, it can be done in the model and in the implementation, but who would want to maintain it? ;^)
So, we just think differently...
Your problem with external indexing would seem to indicate that your global structure could have been better. I have never seen a good mumps structure that failed to represent a self referential index when ordered through at a particular level.
As for the whitespace, well I adapted quite well to Python, I guess, in part because of that decade of MUMPS.
In any case it is not the language, but the structure of the mind that can appreciate it that makes a programmer productive. You seem to feel pretty comfortable with the relational model. When I want to do something like catalog my music, I don't want to only store a name and an artist and a song, I want to explore the influences, histories, etc, call it multideminsional model. It's the same with medicine. There are nuances which the relational model will never be able to accomidate. Sure it's good at what it does, but it will never think like a human being. And that's what we're all going to want sooner or later.
II respectfully disagree. I used MUMPS for a decade before switching. If you look at the MUMPS command structure, the pattern matching and the ability to use indirection, you will see that an even finer level of control was achieved. For example, using a relational database, containing different structures, some finite, like an integer and some not so finite, like a blob, the relational database has to make a much larger number of disk hits just to locate the data. You also don't have to return the entire result of a query before being able to act on records. In MUMPS you can advance through the records one at a time to process them, again much less work for the system to achieve the same goals. With MUMPS and 16K of memory on the server and a VT100 as the client, you could do in text mode, things that you have to have a 1GHz+ PC and web browser to do. I've dismantled record crashed from RDBMS and MUMPS systems, the MUMPS is much easier to recover, the RDBMS you're pretty much screwed. MUMPS hasn't been made obsolete either, just look at CACHE, it's nothing but MUMPS underneath. RDBMS look great when all you're dealing with is similar in nature, a table of this, a table of that. Now let's record your medical record in it' suddenly you discover that for dissimilarly structured data the RDBMS is about as effective as a bazooka at a watchmaker's shop. Someone sold you a bill of goods my friend, time to think outside of the box.
As for the future, I think that while OO databases are all the vogue, that O databases, where each bit of information is represented by it's own object and will have the ability to demonstrate autonomous agency is a good area for research. Then we can just let the data speak for itself! ;^)
"Every person takes the limits of their own field of vision for the limits of the world."
- Arthur Schopenhauer
I'll have it doing a lot more than just playing games in no time!
I'm heavily in favor of having the client do the brunt of the work, and letting the server just dish it out and file it. Servers are still pretty expensive, and most PC's are above 1 GHz these days, so speed is a lot less of an issue. Another issue is that I like pushing the minimum amount of code and data out to the client that will get the job done. I hate it when someone has to load an XML driven ActiveX or Java client tools when all they need is an input field.
And pushing all that 'fancy tool code' is usually where security becomes an issue. "Hey it was there on the toolbar, I just dragged and dropped it onto the layout of the web page and hooked it up to the database table field..." I've heard that one before. Notice there's no really effective version of Moore's Law for software? That is other than this:
"While the hardware will be twice as capable and half the cost, the software will double in cost and be half as effective, so there's really no progress at all!"
And you can quote me on that!