Force a bunch of men to learn how to go to the moon and odds are one of them will come up with something pretty neat, even if you don't give a whit for space exploration. Velcro, anyone?
This has already happened.. the defence research & dev. organization has seen a small amount of their stuff useful in other areas.. supercomputers (really beowulf style clusters) help with weather forecasts after the US govt. banned their export.. of course they are also used for less desirable things. Some carbon-fiber material has been used to make prosthetic aids.. ref. here:
From: http://www.docuweb.ca/India/news/9907.html
Many of the technologies being developed under the umbrella of the Indian Defence Research Organisation are similarly rooted in reality - the need to be self sufficient in core technologies like inertial guidance and propulsion; in heat resistant '`composite'' materials and the design and fabrication of application specific integrated circuits (ASICS). Super computers based on massively parallel processing building blocks by both the Defence and Atomic research organisations have been fruitfully deployed on such applications like weather modeling and medical imaging. Meanwhile, in areas of core competency, Defence researchers are on the threshold of producing India's own version of the Light Combat Aircraft (LCA) - major challenge in avionics which has brought together over two dozen public and private agencies in a consortium venture. Some of the technologies that have gone into the fabrication of critical components for aircraft and missiles also from the core of civilian initiatives like the development of indigenous catheters, heart valves and prosthetic aids for the physically handicapped.
The NYT has done a disservice by presenting this as a tool for increasing trust in the web. It's actually a tool for preserving privacy of data.
A better example would be the analysis of medical records. Now, the IS people at HMOs like Kaiser Permanente have one overriding fear - that they will get up one morning and find that the NYT/WSJ reporting that medical records data has been compromised. As a result, they lock down this data and impose extreme security restrictions in their access. This denies the opportunity to perform data mining analysis, say by an outside contractor.
If you now pose the question as follows: how can an outside agency perform some data mining analysis (for instance infection rates of two diseases in particular areas) on this data without compromising individual security ? One nice answer to this is this technique from the IBM researchers.
So no, this isn't any "fairer" than the last page. It's just less full of complete untruths. Instead it has things that are technically true but not the whole story. A quite nice example of content free marketing.
Well I took a quick glance of the rest of the comments and saw that nobody responded to the claim on SMP support. It's true that that's something Windoze does better. I believe Linux's threading implementation can be much better as well.
The article says he has got an IM norm. That means he needs another norm to become an IM. Still, it's amazing improvement. Especially since I play chess recreationally (ie. I suck at it but I still play:-) it seems absolutely amazing!!
There have been other successful examples of individuals "crossing" sports. Jonty Rhodes who is a cricket international for South Africa, used to play for the South African hockey team as well. It shows as he was probably the best athlete in cricket in the nineties. In first decade of the 20th century, C.B.Fry represented England in both its cricket and soccer teams. Of course, Deion Sanders was reasonably successful at baseball to complement his football.
... but I think differently now. Well I'm in school again, but I worked for some years in the interim.
I think it's good to have a well-rounded education - for three reasons. (1) You really don't know what you're going to do in later years (2) It's pretty much the only time you will get a chance to read and study all kinds of "other" stuff (3) Often times things that you have read in some other field could tangentially affect you - or even if they don't affect you, some glimmerings of what you studied might suddenly make sense (4) Finally, things keep changing in the world.. when I was an undergraduate studying electrical engineering I objected to a course in "General Biology". Now Bio-informatics is a very hot field, and any computer scientist should be at least reasonably familiar with it - how do you build systems to solve such problems otherwise ?
Anyway, things will be very boring if you were only to study the stuff you'll do at work. And for the work stuff, there is little substitute to the real world in any case.
The instant you decide to move data and/or referential integrity from the DB into 'code', you've lost the battle. I can't believe someone would even suggest this. Sure, stored procedures are a plus and most times they aren't necessary, but you simply cannot have your integrity checks outside the DB.
Most people think that you need to have RI in the database only for consistency and it's "easier" or "faster" to handle it "in code".
The reality is that the presence of RI helps sophisticated query rewrite engines to do some very nifty transformations. So having RI actually helps you get better performing plans for read queries and hurts performance a little for writes.
but it's obviously
false given that Deep Blue never played anyone
but Kasparov.
Not true. Plenty of people have played DeepBlue the program. In fact I've played it too.. although I'm only a recreational player ! I lasted about 15 moves !:-)
Seriously, this was when I was a grad student and attended an IBM conference (CASCON'95). DeepBlue was an exhibit along with the research team. In the final day of the conference, DeepBlue played a top Canadian player and defeated him quite easily. Throughout the show, anyone could walk up to DeepBlue and play against it.. most people were quite apprehensive about it though. It had already beaten Kasparov in Philadelphia - the first game only. Kasparov bounced back.. this was Match 1. Apparently the 3rd game these guys misplaced the opening book and it still managed to do okay.
The difference between the DeepBlue I played was that it was on a stock RS/6000 machine (today's p-series e-Server). The beast that played Kasparov was a top of the line SP.
Not meaning to be condescending but this is the difference between academia and the real world. Porting a product to another platform is not just a matter of writing make. You have tons of issues to deal with: compiler idiosyncracies, byte-endian assumptions (yes they are bad, but there are cases where they become important), operating system interface differences.. the subtler the worse it is.
On top of all of this the real kicker is the amount of testing and release engineering you have to do on each platform before shipping it. Believe me you have to do this. Defects are a way of life with real products.
About this language, I'm sceptical of a new replacement for everything. Second, who uses Java on the client-side anyway ? The beauty of Java is that server-side stuff is cross-platform. Java on the client is just another language. I think we're doing just well for various approaches to separate content and logic.
Finally we've put in a whole load of internet infrastructure for dynamic caching. Separating content from logic will also allow us to be even more aggressive with caching fragments of web pages all over the 'net. I'm not sure how a proprietary solution will work..
IMO these guys are probably angling to be bought by Microsoft.
It's a real breeze, and I was impressed. As an IBMer who works on AIX, I used to run AfterStep instead of the abomination that is CDE.
Now with IBM's own "AIX Toolbox for Linux Applications" web page, it's trivial to download all the rpms necessary and just install KDE> You just need to use smitty to install rpm first !
Link at: www.ibm. om/servers/aix/products/aixos/linux/
This happened to me in my current job when I interviewed 3.5 years back. Nobody asked me a single serious technical question. They just took whatever I said about my previous job at face value.
I actually think that's bad. When you are interviewing at a place, you want to get a feel for how smart those folks are.. and a good engaging interview which makes you think and results in some sharp give-and-take is probably one of the best ways to impress a candidate about you and your organization.
Well at least until recently the job market was such that impressing the candidate was important:-)
I have two kinds of questions I ask. I call one "programming in the small" and the other "programming in the large".
For the first, the task is to design an interface for a C function that returns a null-terminated character string. The thing is that the length of the string is random and the caller has no apriori info about the string. Also the function is provided as part of a library so you can make NO assumptions of the environment the caller is in (for instance what memory manager it uses). This is a great question to test a variety of things like interface design skills, nitty-gritty knowledge of how memory-managers work, all the way upto callback functions. I never fail people on the test just because they don't get the ultimate answer I want. More important to me is how they proceed to try to answer the question. Of course at times there are people who are really bad, but I normally phone screen to avoid such folks.
My second question is a little different. I quiz them a little about the internet and the HTTP protocol and ask the candidate to tell me how webserver implementations would be different if HTTP was a state-ful protocol.
Finally I ask the candidate their favourite field of computer science and ask a question in that.
In general when I interview I look for people with good systems skills. Database internals experience is a plus.. but the market was so tough last year my gosh. I must have interviewed about 20 people to staff 3 positions.
IANAC (Cryptographer:-), but isn't this essentially like a cryptography problem ? The data in the tapes is binary data in 0s and 1s. We have a bunch of printouts with presumably contiguous segments matching presumably contiguous segments of the tape data. Surely we could take a small segment of the printouts and run it over a moving window of tape data segments till we can establish the mapping between the binary data and the printout information.
Yes.. this is probably a stupid wacky idea.. but hey at least I entertained you.
Tex is a very interesting program. Don Knuth's book "Tex: The Program" and paper in "Software, Practice and Experience" are required reading for any serious software developer. Just for the fun of it:-)
In an earlier avatar in grad school, I used to do research in software testing and reliability, and my professors were early pioneers in the field. In the days when they were in grad school (mid-seventies) the first questions they were trying to answer is "What kind of bugs do programmers make ?" - the reason being they were trying to come up with effective strategies to test programs that eliminate the more common and more serious defects first. These guys used to stand outside undergrad computing labs and grab printouts from students who left to analyze for errors !!
Later on in his career, my advisor spent six months with Knuth at Stanford to come up with a categorization of the various kinds of errors in programs.
Naw.. this is just religious bs. I know non-technical people who work at Sun and bitch and bitch about having to use StarOffice. Bottom-line is that their productivity is way down. Especially those who have to work with the outside world.
I used to be religious before, but I'm quite agnostic now. The important thing is to be able to do your job well and fast. "Eating your own dogfood" is _only_ useful to weed out defects in the product. Not to satisfy some meaningless jihad.
I mean, Sun's gotta get a clue. They are a non-entity in the client space. Why force their non-techies to use Sun boxes ?
Finally don't even talk about costs. I'm sure it's much much cheaper to get the employees x86 boxes from whoever. More importantly, think of all those powerful SPARC chips (which were designed for I/O intensive systems) being wasted on X !! They are best used with Sun's customers.
Just to correct the article. The SEI (Software Engineering Institute) referred to in the article has nothing to do with the Feds.
The SEI is part of Carnegie Mellon University. The folks at SEI developed the "Capability Maturity Model" and this is the model the article talks about.
I used to do research in Software Engineering in a prior life and got frustrated with it. "Mainstream" CS folks are extremely snobbish and treat most S/W Engg Research as "very soft".
Conversely, the industry has very little time for the wonkish papers put out by the software engineering community. Ignored by the rest of CS and by industry, software engineering is certainly the poor step-child.
In a sense, the field has itself to blame. Much of the "research" in software engineering gets done with toy programs (I was quite guilty of this), and in many ways didn't relate to the real world. Sometimes however, good tools and good practices did come out.. however, getting individual developers to use these tools is quite difficult.
In the real world I experienced working at an extremely chaotic technology company, where people didn't even bother to write one-pager design documents. Now I work for one of the biggest technology companies and we have a reasonable balance between process and keeping the creativity going. Certainly I can tell from personal experience that the process has helped us find countless defects at an earlier stage in the game. At the same time, the flexibility we have with the process (individual developers are encouraged to come up with their schedules as opposed to being laid down with the law by management) has kept us retain our creative edge, which we do need to fight our tough competitors.
Actually OS/390 runs the S/390 machine. And to be accurate, the zOS runs the z-series e-server:-)
VM is the "sandbox" in which Linux runs.
Furthermore, it is possible for Linux to run on the S/390 (z-series) bare metal, and have a single Linux image on one huge big-ass mainframe box. It's just pretty stupid to do that. It would waste the tremendous capacity of a mainframe.
Unfortunately, in a Java server side environment when you have about 300 threads with the same priority of the same process, this policy doesn't work too well. The rub lies in the fact that the scheduler treats essentially treats each of the 300 threads as though they are 300 processes, and applies the algorithm to select the next thread to execute to every process in the list. There was a result shown by some guys from IBM that a rude hack like letting the scheduler to always choose a thread at random ends up performing much better for the server-side Java benchmarks. Of course, this is only a rude hack for pedagogical purposes.
Oracle makes the query plan at execution time (and caches it for efficiency) this means that if the nature of your data changes, or you add new indices, you have to re-compile the queries stored in DB2 or it will continue to use inefficient query plans. I consider this extremely stupid.
Not really fair.. yes, we cache compiled access sections and that's a good thing because most stuff these days is dynamic sql. In the old days you had more static sql and we used to rule.
More to the point, if you add indexes, or the nature of the data changes, just issue RUNSTATS. That will automagically invalidate the access plans in the dynamic section cache and the next time the query gets executed it will _automagically_ get recompiled (transparent to the user) using the new statistics.
I do not suggest that there are no uses for ODBMS systems. Certainly the example you give for CAD systems is good.
Yes, the evaluations I did were circa 1997. It's quite possible that better things have come since then. However, I have also interviewed a couple of engineers from Versant and their description of their systems left a lot to be desired - of course that could just be them.
Oh btw, I not only did evaluations, I also played around looking at the internals of one of the academic systems.
I'm sure my employer uses/resells ODBMS systems somewhere. Big deal.. if my comments suggested that I think ODBMS systems will fade away - I don't. I just think to suggest that they will replace relational systems is stupid. I think we both agree there.
I agree with a lot of what you say.. however:
An Object Database is a far more general solution, it is childs play to make an ODBMS pretend to be an RDBMS the opposite is what is hard.
This is laugable ! Yes, it's child's play to get an ODBMS support "relational" operations - but it's impossible to support relational operations in any efficient manner even coming close to what a relational engine can do. Yes I'm biased, but I also have the ability to view engineering problems with objectivity. I submit that anybody who thinks performance isn't an important consideration for a relational system doesn't really understand the business and what relational customers want.
General solutions can't be optimized for specific usage, because they are general. By the time the relational companies are done building ORDBMS that really compete with us..
We aren't really trying to compete with you if you are providing persistence for objects in an interactive application for CAD design. We are competing with you in caching solutions for web applications. I strongly believe that an ODBMS solution isn't scalable for a high-volume web application. You might get some performance wins when your load is low, but one thing you have to understand is the difference between performance and scalability. How on earth do you plan to support shared-nothing (or shared-disk) parallelism for instance ? I know of large trading sites which have clusters of 16 mainframe boxes which are white-hot. You cannot pump that kind of transaction volume through ODBMS systems.
ODBMS and RDBMS systems each have their own places. There is a small area where they overlap - and in that overlap area I believe ORDBMS will win because they offer the best of one and a "reasonable" enough amount of the other. The pure ODBMS market will definitely survive. How many vendors will be left, I'm not sure.
The question is, "why did you go to CS ?". If you went to CS just for "a job" or "a career" I can understand how you feel today. I went to CS because I was fascinated not just by technology, but by the thought of building technology.
Today I have a great job building technology and doing cutting edge research. Sure, I don't work for a dot-com and won't be a zillionaire but I have a bunch load of fun. So much fun that I'm considering going back to school and getting a Ph.D.
Slashdot Mirror
Force a bunch of men to learn how to go to the moon and odds are one of them will come up with something pretty neat, even if you don't give a whit for space exploration. Velcro, anyone?
.. the defence research & dev. organization has seen a small amount of their stuff useful in other areas .. supercomputers (really beowulf style clusters) help with weather forecasts after the US govt. banned their export .. of course they are also used for less desirable things. Some carbon-fiber material has been used to make prosthetic aids .. ref. here:
This has already happened
From: http://www.docuweb.ca/India/news/9907.html
Many of the technologies being developed under the umbrella of the Indian Defence Research Organisation are similarly rooted in reality - the need to be self sufficient in core technologies like inertial guidance and propulsion; in heat resistant '`composite'' materials and the design and fabrication of application specific integrated circuits (ASICS). Super computers based on massively parallel processing building blocks by both the Defence and Atomic research organisations have been fruitfully deployed on such applications like weather modeling and medical imaging. Meanwhile, in areas of core competency, Defence researchers are on the threshold of producing India's own version of the Light Combat Aircraft (LCA) - major challenge in avionics which has brought together over two dozen public and private agencies in a consortium venture. Some of the technologies that have gone into the fabrication of critical components for aircraft and missiles also from the core of civilian initiatives like the development of indigenous catheters, heart valves and prosthetic aids for the physically handicapped.
The NYT has done a disservice by presenting this as a tool for increasing trust in the web. It's actually a tool for preserving privacy of data.
A better example would be the analysis of medical records. Now, the IS people at HMOs like Kaiser Permanente have one overriding fear - that they will get up one morning and find that the NYT/WSJ reporting that medical records data has been compromised. As a result, they lock down this data and impose extreme security restrictions in their access. This denies the opportunity to perform data mining analysis, say by an outside contractor.
If you now pose the question as follows: how can an outside agency perform some data mining analysis (for instance infection rates of two diseases in particular areas) on this data without compromising individual security ? One nice answer to this is this technique from the IBM researchers.
Well I took a quick glance of the rest of the comments and saw that nobody responded to the claim on SMP support. It's true that that's something Windoze does better. I believe Linux's threading implementation can be much better as well.
Would love to be disabused of course ..
The article says he has got an IM norm. That means he needs another norm to become an IM. Still, it's amazing improvement. Especially since I play chess recreationally (ie. I suck at it but I still play :-) it seems absolutely amazing!!
There have been other successful examples of individuals "crossing" sports. Jonty Rhodes who is a cricket international for South Africa, used to play for the South African hockey team as well. It shows as he was probably the best athlete in cricket in the nineties. In first decade of the 20th century, C.B.Fry represented England in both its cricket and soccer teams. Of course, Deion Sanders was reasonably successful at baseball to complement his football.
I think it's good to have a well-rounded education - for three reasons. (1) You really don't know what you're going to do in later years (2) It's pretty much the only time you will get a chance to read and study all kinds of "other" stuff (3) Often times things that you have read in some other field could tangentially affect you - or even if they don't affect you, some glimmerings of what you studied might suddenly make sense (4) Finally, things keep changing in the world
Anyway, things will be very boring if you were only to study the stuff you'll do at work. And for the work stuff, there is little substitute to the real world in any case.
The instant you decide to move data and/or referential integrity from the DB into 'code', you've lost the battle. I can't believe someone would even suggest this. Sure, stored procedures are a plus and most times they aren't necessary, but you simply cannot have your integrity checks outside the DB.
Most people think that you need to have RI in the database only for consistency and it's "easier" or "faster" to handle it "in code".
The reality is that the presence of RI helps sophisticated query rewrite engines to do some very nifty transformations. So having RI actually helps you get better performing plans for read queries and hurts performance a little for writes.
And I live in the valley too .. unfort I will be on vacation and will be out of town.
Not true. Plenty of people have played DeepBlue the program. In fact I've played it too .. although I'm only a recreational player ! I lasted about 15 moves ! :-)
Seriously, this was when I was a grad student and attended an IBM conference (CASCON'95). DeepBlue was an exhibit along with the research team. In the final day of the conference, DeepBlue played a top Canadian player and defeated him quite easily. Throughout the show, anyone could walk up to DeepBlue and play against it .. most people were quite apprehensive about it though. It had already beaten Kasparov in Philadelphia - the first game only. Kasparov bounced back .. this was Match 1. Apparently the 3rd game these guys misplaced the opening book and it still managed to do okay.
The difference between the DeepBlue I played was that it was on a stock RS/6000 machine (today's p-series e-Server). The beast that played Kasparov was a top of the line SP.
On top of all of this the real kicker is the amount of testing and release engineering you have to do on each platform before shipping it. Believe me you have to do this. Defects are a way of life with real products.
About this language, I'm sceptical of a new replacement for everything. Second, who uses Java on the client-side anyway ? The beauty of Java is that server-side stuff is cross-platform. Java on the client is just another language. I think we're doing just well for various approaches to separate content and logic.
Finally we've put in a whole load of internet infrastructure for dynamic caching. Separating content from logic will also allow us to be even more aggressive with caching fragments of web pages all over the 'net. I'm not sure how a proprietary solution will work ..
IMO these guys are probably angling to be bought by Microsoft.
Now with IBM's own "AIX Toolbox for Linux Applications" web page, it's trivial to download all the rpms necessary and just install KDE> You just need to use smitty to install rpm first !
Link at: www.ibm. om/servers/aix/products/aixos/linux/
I actually think that's bad. When you are interviewing at a place, you want to get a feel for how smart those folks are .. and a good engaging interview which makes you think and results in some sharp give-and-take is probably one of the best ways to impress a candidate about you and your organization.
Well at least until recently the job market was such that impressing the candidate was important :-)
For the first, the task is to design an interface for a C function that returns a null-terminated character string. The thing is that the length of the string is random and the caller has no apriori info about the string. Also the function is provided as part of a library so you can make NO assumptions of the environment the caller is in (for instance what memory manager it uses). This is a great question to test a variety of things like interface design skills, nitty-gritty knowledge of how memory-managers work, all the way upto callback functions. I never fail people on the test just because they don't get the ultimate answer I want. More important to me is how they proceed to try to answer the question. Of course at times there are people who are really bad, but I normally phone screen to avoid such folks.
My second question is a little different. I quiz them a little about the internet and the HTTP protocol and ask the candidate to tell me how webserver implementations would be different if HTTP was a state-ful protocol.
Finally I ask the candidate their favourite field of computer science and ask a question in that.
In general when I interview I look for people with good systems skills. Database internals experience is a plus .. but the market was so tough last year my gosh. I must have interviewed about 20 people to staff 3 positions.
Oh btw .. _I_ am a H1-B worker ...
Yes .. this is probably a stupid wacky idea .. but hey at least I entertained you.
Interesting thing is that it categorizes the search results .. very nice. Especially the option to search patents.
In an earlier avatar in grad school, I used to do research in software testing and reliability, and my professors were early pioneers in the field. In the days when they were in grad school (mid-seventies) the first questions they were trying to answer is "What kind of bugs do programmers make ?" - the reason being they were trying to come up with effective strategies to test programs that eliminate the more common and more serious defects first. These guys used to stand outside undergrad computing labs and grab printouts from students who left to analyze for errors !!
Later on in his career, my advisor spent six months with Knuth at Stanford to come up with a categorization of the various kinds of errors in programs.
I used to be religious before, but I'm quite agnostic now. The important thing is to be able to do your job well and fast. "Eating your own dogfood" is _only_ useful to weed out defects in the product. Not to satisfy some meaningless jihad.
I mean, Sun's gotta get a clue. They are a non-entity in the client space. Why force their non-techies to use Sun boxes ?
Finally don't even talk about costs. I'm sure it's much much cheaper to get the employees x86 boxes from whoever. More importantly, think of all those powerful SPARC chips (which were designed for I/O intensive systems) being wasted on X !! They are best used with Sun's customers.
The SEI is part of Carnegie Mellon University. The folks at SEI developed the "Capability Maturity Model" and this is the model the article talks about.
I used to do research in Software Engineering in a prior life and got frustrated with it. "Mainstream" CS folks are extremely snobbish and treat most S/W Engg Research as "very soft".
Conversely, the industry has very little time for the wonkish papers put out by the software engineering community. Ignored by the rest of CS and by industry, software engineering is certainly the poor step-child.
In a sense, the field has itself to blame. Much of the "research" in software engineering gets done with toy programs (I was quite guilty of this), and in many ways didn't relate to the real world. Sometimes however, good tools and good practices did come out .. however, getting individual developers to use these tools is quite difficult.
In the real world I experienced working at an extremely chaotic technology company, where people didn't even bother to write one-pager design documents. Now I work for one of the biggest technology companies and we have a reasonable balance between process and keeping the creativity going. Certainly I can tell from personal experience that the process has helped us find countless defects at an earlier stage in the game. At the same time, the flexibility we have with the process (individual developers are encouraged to come up with their schedules as opposed to being laid down with the law by management) has kept us retain our creative edge, which we do need to fight our tough competitors.
VM is the "sandbox" in which Linux runs.
Furthermore, it is possible for Linux to run on the S/390 (z-series) bare metal, and have a single Linux image on one huge big-ass mainframe box. It's just pretty stupid to do that. It would waste the tremendous capacity of a mainframe.
The thing is that seatbelts reduce fatalities by 40% (even without airbags) in a crash. It's too huge a number to ignore.
Unfortunately, in a Java server side environment when you have about 300 threads with the same priority of the same process, this policy doesn't work too well. The rub lies in the fact that the scheduler treats essentially treats each of the 300 threads as though they are 300 processes, and applies the algorithm to select the next thread to execute to every process in the list. There was a result shown by some guys from IBM that a rude hack like letting the scheduler to always choose a thread at random ends up performing much better for the server-side Java benchmarks. Of course, this is only a rude hack for pedagogical purposes.
Not really fair .. yes, we cache compiled access sections and that's a good thing because most stuff these days is dynamic sql. In the old days you had more static sql and we used to rule.
More to the point, if you add indexes, or the nature of the data changes, just issue RUNSTATS. That will automagically invalidate the access plans in the dynamic section cache and the next time the query gets executed it will _automagically_ get recompiled (transparent to the user) using the new statistics.
Yes, the evaluations I did were circa 1997. It's quite possible that better things have come since then. However, I have also interviewed a couple of engineers from Versant and their description of their systems left a lot to be desired - of course that could just be them.
Oh btw, I not only did evaluations, I also played around looking at the internals of one of the academic systems.
I'm sure my employer uses/resells ODBMS systems somewhere. Big deal .. if my comments suggested that I think ODBMS systems will fade away - I don't. I just think to suggest that they will replace relational systems is stupid. I think we both agree there.
I agree with a lot of what you say .. however:
An Object Database is a far more general solution, it is childs play to make an ODBMS pretend to be an RDBMS the opposite is what is hard.
This is laugable ! Yes, it's child's play to get an ODBMS support "relational" operations - but it's impossible to support relational operations in any efficient manner even coming close to what a relational engine can do. Yes I'm biased, but I also have the ability to view engineering problems with objectivity. I submit that anybody who thinks performance isn't an important consideration for a relational system doesn't really understand the business and what relational customers want.
General solutions can't be optimized for specific usage, because they are general. By the time the relational companies are done building ORDBMS that really compete with us ..
We aren't really trying to compete with you if you are providing persistence for objects in an interactive application for CAD design. We are competing with you in caching solutions for web applications. I strongly believe that an ODBMS solution isn't scalable for a high-volume web application. You might get some performance wins when your load is low, but one thing you have to understand is the difference between performance and scalability. How on earth do you plan to support shared-nothing (or shared-disk) parallelism for instance ? I know of large trading sites which have clusters of 16 mainframe boxes which are white-hot. You cannot pump that kind of transaction volume through ODBMS systems.
ODBMS and RDBMS systems each have their own places. There is a small area where they overlap - and in that overlap area I believe ORDBMS will win because they offer the best of one and a "reasonable" enough amount of the other. The pure ODBMS market will definitely survive. How many vendors will be left, I'm not sure.
So I wrote the above ..
The great thing about TDS is that they make fun of the media organizations that cover the news as well !!
Today I have a great job building technology and doing cutting edge research. Sure, I don't work for a dot-com and won't be a zillionaire but I have a bunch load of fun. So much fun that I'm considering going back to school and getting a Ph.D.
So CS isn't necessarily that bad, you know.