How do you know you NEED all these features? Have you prototyped the system yet? Have you done your UML diagrams yet? I think that most of the languages you mentioned could fit the bill (of course, this forum is heavily non-M$, so expect to see VB downplayed).
A while ago I did a comparative study of the graphical design tools on which UML is based. My conclusion was that the idea was a pretty bad one and all of them became more trouble than they were worth as they attempted to track every feature of C++ or such graphically. As the projects grew in scope the diagrams became less and less useful.
I was recently forced to use UML, it appears to me to be worse in every important respect than it's predecessors. In addition to the added complexity of now tracking multiple languages UML has lost any coherence the input languages had. UML certainly does not fit well with XML Schema which has a particularly complex data model. In the end I rolled my own graphical markup which people seemed to like but it probably worked because I was using it to present a design rather than create one and I absorbed a bunch of our coding conventions into the notation so the notation was for the subset of XML Schema we used rather than every last feature.
I have noticed that UML and its ilk tend to appeal to people who are brought up on databases and make the mistake of thinking the entity relationship model is useful. Predicate calculus and typed set theory are vastly more powerful in my experience. If I see a bunch of schemas written in Z or VDM I can understand them pretty quickly, I can also see the mapping from the schema to code.
As to the original question, it appears bogus to me. Much more important that what is the most fashionable and feature rich language is what language is going to have support over the life of the project. Java is a definite, C# is almost but not quite guaranteed to be arround.
I syspect that the question is not posed to get an honest reply. The question appears much more likely to be intended to beat the drum for operator overloading.
As such it is worth remembering that Java abandoned operator overloading for good reasons. The C++ approach was just too hard to optimize and lead to buggy and unreliable compilers and code. C# may have got the mix right, I seem to recall there being some limited operator overloading mechanism with a lot of restrictions.
That's a convenient excuse for the USPTO. ``Gosh... it'd be too hard to really do our job of verifying that this claim is original. So we'll let the courts sort it out.'' Or do they honestly believe that the prior art has to exist in a patent description?
The USPTO uses two contradictory arguments acording to need. When the complaint is made that they issue trivial patents they claim that they cannot perform an adequate review. When the point is made that every other country requires publication and a public opposition period prior to issue they claim that their review is more than adequate.
I doubt that the case will succeed. The patent claims are for a specific type of hypertext linking model which is actually the opposite of the URL method used in the Web. What BT 'invented' was the Prestel database driven hypertext model in which all the links were managed in a central database. The innovation of the Web was to discard the stupid database, and the whining hypertext weenies and embed links directly into the page.
While the Web approach appears 'obvious' today, back in the early days of the Web the hypertext community wrote long boring articles on why central databases are the only way to go. Tim could not even get published in the hypertext litterature, until the next year when they asked him to be the keynote speaker.
Prestel served up a group of numbered pages that could reference other pages by the page number. While this was an innovative idea in 1603 when the idea was first proposed the use of page numbers to reference other parts of a text has been standard practice for at least 350 years. It certainly was not the Web and certainly was not hypertext as most people understand it.
Please note that the folk going on about BT's status as a public/private company are both off base. At the time the innovative Prestel system was invented and the patent filed, BT was a government monopoly and part of the Post Office. BT only began their moneygrubbing patent demands after privatisation.
My company is committed to not growing, and it's amazing that I found so few other companies with the same princples, given the obvious success of the idea. ArsDigita is just one of any number of companies that went through the same trajectory.
Actually there are lots of professional services companies with the same approach. A lot of legal, accounting and engineering practices have a 'no growth' or 'slow growth' policy.
The reason is that professional services companies do not become more profitable as they become larger. A PSO company sells the professional skills of its employees. In a typical PSO company there is a pyramid of expertise. For each senior partner there might be two junior partners, four associates and the same number of support staff. That ratio does not change much if there are a hundred senior partners or one.
The problem for the PSO company is that a senior partner has to be paid pretty well or else they will go off and start their own outfit. At the bottom of the tree an associate might be paid $400 a day and be charged out at $1,600 - a markup of 400%. But at the top of the tree a senior partner is likely to be paid more than their charge out rate, their real job being to bring work into the practice. A senior partner might charge out 100 days a year (i.e ahlf the normal rate) at $5,000 a day, but they can keep busy 4 associates at a profit of $6,400 a day and 2 juniors at a profit of $3,600 a day, so they actually contribute to the practice an average of $12,500 a day and take out maybe $6,000.
The reason PSO firms have to grow is that as the people at the base of the pyramid get more experience the only way of maintaining profitability is to either increase recruitment at the base or to restrict the number of promotions. That is why a lot of accounting companies have 'up or out' policies. If you don't make the next step in the ladder by a certain date you are told to look for work elsewhere.
All in all VC should never be funding pure PSO companies. Most PSO companies are organized as partnerships for good reason - the company itself actually has very little value, the value is all in the knowledge of the employees. And those employees can and will walk out the door with it unless they feel they are rewarded for it.
PSO can add a lot to a growth company's bottom line, but only as a supplement to product, not as a replacement. In general the markets tend to be sniffy about companies that make more than 20-25% of their revenues from PSO. An Oracle or an IBM can make a lot of money from PSO, even run a pure PSO division. But that works because the customers (and consultants) know that they are buying more than just the consultant's time.
The other problem with PSO is that it is economically a pretty risky proposition. A typical PSO company operates on a margin or 20%. If there is a downturn they don't have much scope to cut costs without layoffs.
Wrong. The Euro is not gold backed. If it was it would not have collapsed in the year after introduction while the price of gold was rising.
No industrialised country can function with a gold backed currency. In an industrialised country the economy expands with improved production over time. But the production of yellow metal is pretty constant. You have a quantity that has exponential growth tied to one that has only linear growth. The result is an economic slump.
This is why Karl Marx wept when he heard of the '49 gold strike. The additional gold supply would allow the capitalist economies to expand and the revolution would be delayed for a decade or more. The theory of the inevitable communist revolution was based in large part on the belief that the gold standard was essential.
Elitism will not work. Because if people have the ability and the time, but no job, they will sit around making high quality software and giving it away for free. And that poses quite a little problem if you have a similar product and want to charge for it, now doesn't it?
Not really, true elite programmers and system architects are very rare and their productivity is massively greater than that of a merely 'good' or 'average' hacker.
If an elite programmer has a problem getting a job it is because they have a personality problem or a substance problem that is so extreeme they are not likely to offer much to an OSS collective either.
That does not mean that there are no elite programmers doing OSS stuff, but most of the ones I know are doing OSS because they made a bundle from the dotcom boom and don't know what to do with the rest of their life.
I don't think that mere incremental improvement of OSS products is going to threaten the paid software model. For MS Office to be displaced there has to be a discontinuous change, not an incremental one. Matching Word feature for feature as you suggest is a pointless waste of time. Mr Softy has shown that he knows how to win that game.
I hate Linux and Open Source in general with a passion
I hate the weenie aspect. There are plenty of OSS folk who are reasonable and knowledgeable. At one time I wrote a lot of OSS code.
Then there are a bunch of maga-bores who have adopted OSS as their life message and religion. Talking to them is like talking to a born again Christian or a Taleban fanatic, everything comes back to their own all embracing pet theory. It is like talking to a Libertarian or a Marxist, they have an economic theory and it explains everything and what is more if you don't agree with them you are [damned to hell for eternity/ a legitimate target for murder/ a complete imbecile/ an oppressor of the masses].
I don't think the talk was actually anti-open source, the guy took potshots at plenty of people's sacred cows. But the only one that slashweeniedom is going to talk about is OSS because it is soooooooo interesting
The problem is with what the weenies think, the problem is weenieism. As Karl Popper pointed out most of the misery in the world is caused by people who have an absolute belief in their possesion of an absolute truth.
I gave up on OSS once it became popular and the weenies moved in.
An absurd fallacy. Perhaps for fetchmail or hello, world! or other,
similarly sized projects, but nowhere else.
The cause of most longstanding security bugs is almost always a major error of design that is not easily corrected.
Take for example sendmail and Outlook, two programs that have become the Typhoid Mary of security. The problem in both cases is a failure of code/data separation. Outlook encourages users to open random content sent from random addresses and we wonder why viruses are a problem. Sendmail has a crazy macro processing language built in that is completely unnecessary for managing SMTP, the numerous buffer overflows in this code only serve to hide the structural problem.
The other problem that the talk highlights is that ideology is the cause of the problem and not the solution. sendmail has been opensource for almost twenty years and shows no signs of being fully debugged yet - and that despite the fact that it is an application whose requirements have decreased over time. Who cares about routing bang path mail or X.400 these days? The code to support that stuff should have been excised years ago.
The reason I like source code is that it allows me to extend a program to do something I need it to do that it doesn't do yet. Even then I would much prefer a well written API to source code, but source code plus API plus documentation is best of all.
This is completely untrue. First of all nobody is holding a gun to anyone's head to program in Java.
That is not what I said. Java only allows you the choice to use it or not. If you try to develop a version of the language that meets your needs that Sun does not like, Sun will sick lawyers on you.
Again total bullshit. If it were only Sun's interest the JCP wouldn't exist. If it were only Sun's interest Oracle wouldnt have created a native compiler for Java in the Oracle DB.
My experience of JCP is that Sun decides who gets to chair your group then the chair goes off to do what they damn well please. It is not an open process at all. Oracles's java compiler is clearly in the interests of Sun as well as Oracle.
Also, it'd be great if you could point out how exactly Java has become "anti-Microsoft" in any way except by Microsoft itself?
There is nothing to prevent people writing to a portable sub-set of C# if they chose to do so. That is very different from Java which denies choice, programmers are forced to program to the lowest common denominator.
The problem with Java is that the language will always be limited to the subset of functionality that it serves Sun's interests to provide. Optimised i86 code and direct access to native O/S features don't meet those interests. So they will never be first class features of the language, even if Sun does not sick its lawyers onto companies that try to provide them.
While C# reflects Microsoft's interests those are much more closely tied to those of developers. It is not in Microsoft's interest to limit functionality in any way.
This sounds like FUD. He didn't really post any examples about what kind of problems C# has for security, that would have been helpful.
I really would hope that Joy would be above stooping to a FUD strategy. But apparently not.
What Joy does is to conflate two separate issues. The first is network code, the second is adding runtime code safety to C.
Nobody disputes the fact that network code should be safe. Microsoft's security policies don't let you download unsafe code by default and the CLI runtime checker kicks in in precisely the same way it does in Java.
That is not the same as saying that there is no use in any circumstance for unsafe code. The vast majority of the Linux kernel is 'unsafe code' if you want to be pejorative. Java simply refuses to ever let you write a piece of code of that type in Java and tries to make interfacing to that type of code as hard as possible. Microsoft on the other hand have rather a lot of legacy C++ code that they want to access even though they can't rewrite every line in C#. I think that Microsoft's interest here is closer to that of most developers.
As for the MIME type FUD, it is a real pity that sun deliberately bypassed the MIME type system in HTTP to prevent firewalls being able to block Java. I made the protest at the time, they ignored it.
Equally criticising Microsoft for not implementing the Javascript security model is a bit rich since few people outside Netscape would agree that Javascript has a security model. By their own admission Javascript was thrown together in a fortnight and thrown out the door with almost no checking.
Sun would do much better to consider how they are going to survive as a specialist hardware manufacturer as Linux comoditizes the UNIX workstation market rather than spend their time engaged starting stupid arguments with Microsoft. As Balmer himself points out lots of companies have lost their way by focussing on Microsoft's business model instead of their own.
The problem with Java is that it is a closed, proprietary language whose primary design criteria has become 'get Microsoft'. In the process Java has been deliberately made less useful to windows programmers, which means the vast majority.
I think that Sun is playing right into Microsoft's hands by balkanizing software development into Java and Windows camps. History suggests that the larger developer pool will win out in the end. A stupid, stupid strategy.
Communicating Sequential Processes. It is an algebra for describing parallel processes that communicate with each other. The key feature of CSP is that when the processes communicate they rendezvous, that is if A is trying to talk to B and B is not ready then A will wait until B is ready, contrawise B will wait if A is not ready.
The main application of CSP in program language design was the inmos Occam programming language.
The advantage of CSP is that it is much easier to write parallel programs because communication between processes is handled in the same way as communication between processors. The type of structures that have to be kludged up out of semaphores and shared memory in UNIX/C or Windows/C can be provided as native language features.
What CSP ammounts to is a start on developing the equivalent of structured programming constructs for network computing applications. In practice you need to add a few extra structures such as FSM support and structures to handle termination and fault handling is a decent fashion.
Dealling with the responses is an exercise in what we called at the MIT AI lab 'mass listening'. It is very hard to correlate that volume of response in a usefull fashion. But it is done every week for the President and on a smaller scale each member of congress.
I am not surprised at the breakdown of the messages, except that the number of messages rejected as 'opinion' (7,000) sounds rather low if anything. The number of form letters (3,000)also sounds like it on the low side.
I doubt that anyone in the administrationis going to treat the messages as 'votes' [what start a lawsuit to stop them being counted? - Ed]. The number of messages on both sides will have been inflated by 'astroturf' (fake grass roots) campaigns by Microsoft, Sun, AOL etc. Fortunately messages of that type tend to be easier to spot than the people who purchase the campaigns think.
The bulk of the messages will simply repeat each other and standard positions fed to people by the media (including slashdot). I suspect that the 48 'substantive' comments are mainly the briefs written by industry lawyers to support one party or another. I strongly suspect however that it is the case that practically every idea expressed in the 22,000 contributions is covered in the 48 'substantive' contributions. Identifying a small number of contributions that put all the important issues well is a tremendous service to people trying to read the materials.
Taking the feedback as email will have helped sorting to an enormous degree. But a structured forum with some form of moderation could have helped the feedback further, collapsing repetative positions down to one instance and such. The moderation need not have been on the slashdot model in which there is a single pool of moderators, there could be twin panels of moderators representing each side. After all posting troll comments and pornography would do nothing for either side unless they wanted to discredit the dabate.
Finally the cost of publication at $400 a page does not seem unreasonable, it is roughly equivalent to the cost of printing and distributing about 1,000 copies. That is not much more than one per senator, congressman, state AG, party affected and news organization.
Granted, you could pick XML as a more recent thing that's going to have long-term improving effects. But it's not really a huge innovation, and it may turn out in 5 years that it's not as relevant as everyone thinks it's going to be.
I don't think anyone should get a Turing for XML, the inventive step took place in SGML about 15 years earlier. XML is simply a cleanup of SGML that removes incompetence, idiocy and illogic from the SGML design. The basic principle is the same though. SGML was too badly executed to give Goldfarb any awards.
Rather than give Dahl and co an award I would much rather people looked at their ideas and acted on them. People slap each other on the back over their use of 'object oriented' languages, but Java and C# both offer only a small part of the power of the OOP model and C++ offers the wrong parts.
The modern OOP languages offer only the data structuring concepts, Simula also had a message passing concept which has largely been lost. It took me ages to work out that the reason the message passing explanation in the C++ book was so hard to follow is that it is actually irrelevant to C++.
A true fine grained parallelism model with message passing would be very useful at this point in time when a lot of programming projects involve networks and GUIs. The pthreads model is too low level. It is a pity that the developers did not take note of the fact that a year after inventing monitors, Hoare invented CSP which is a more powerful model.
To give concrete examples of the benefit of deep grained parallelism and message passing consider how often Windows halts waiting for input in some dialog box in a different window or how often an X-Window will be left unrefreshed during processing. The reason is that doing the job properly requires multithreading and the pthread model is to cumbersome for most developers to deal with.
Secondly, I'm constantly amazed at how people mis-read our stats page....
Regardless of anything else, using these number to declare that one thing is more secure than another is a mistake. Based on our numbers, why didn't they declare that everyone should run MacOS for security?
I think the problem is that people were misreading the numbers both ways. The use of the raw Bugtraq numbers against Windows was always a canard, the use of the raw figures in the reverse direction is a canard.
The article's argument is sufficient to demonstrate that the 'Linux is more secure' argument is false, but insufficient to prove 'Windows is more secure'. As you point out few bugs are found on the MacOs, that is not surprising since these days they hold MacWorld in a telephone booth and in any case just how many security holes did Edison have in his desk lamp?
The problem is that security really is complex, it certainly is not a linear issue and it is completely determined by your operating environment. An O/S configuration to secure server will almost certainly prevent a user getting useful work done.
As a security professional who is pretty well known in the field, I can tell you that both the Linux is more secure and Windows more secure religious arguments are wrong.
Windows cannot currently compete in the real high end security configuration where we strip down the O/S to run only the services that are absolutely essential. However Microsoft make no secret of the fact they are working on a platform of that type. If I could find a way to audit that work I would rather buy a secure kernel in than have to spend seven figure sums doing the strip down in house.
The multiple eyes argument in favor of Open software only works up to a point. The problem is that you rely on the defenders being more vigilant than the attackers, that is not always the case. Although these days the trend in hacking has been to go for the binaries rather than muck arround with source code.
The biggest problem with Windows is the predeliction for supporting active code in email messages. But Microsoft is not the only company that does not understand the importance of code/data separation. Sun and Netscape have both been guilty of equally eggregious abominations.
In order to meet C2, the NT box can't be connected to a network, a serial connection, or a modem. Well, you can, but you can't allow anybody access to it, same thing.
That is a consequence of the C2 standard which was written by the military for the US govt in the days before networking.
C2 was obsolete before the Web existed. Back in 1993 when I was asked to do a security audit of the Web standards against the Orange book I concluded that the standard was no help at all.
The other reason that C2 is not very useful is that the main concern in Orange book is partitioning multiple users data on the same machine. These days each user has their own machine, a one person computer that does not meet C2 mandatory access control requirements can be perfectly secure - look at a Palm or Pocket PC or a smartcard.
A lot of people are very hesitant to install a whole set of libraries to run only one application -- almost no matter how good the app is -- when there are 'good enough' alternatives for the standard libs they already have.
The main reason that Mosaic was the first mega successful Web browser was that it was the first to use a GUI toolkit that did not look like crap. It was not actually the first Motif browser but it was the first with Motif look and feel. The other browsers looked like science projects.
Ten years later it is quite possible that Motif's time has come and gone. The Motif look is somewhat dated and the OSF licensing model is certainly dated. Unless it was released as open source sometime I didn't notice you still have to pay for Motif which pretty much rules it out in the Linux world. I don't think that the chances of survival for the non-open source Unix world are very good these days.
What puzzles me is that these toolkits still need a rack of twenty manuals each of which is six inches thick. Its only a goddam menu system!
this one is the truth, not the parent, yet this post get -1 and the partne 3, informative. what a load of crap. that is exactly what gore said "I took the initiative in creating the internet." idiot slashcrappers.
There is a very determined group of people who want to project their particular version of events as 'the truth'. So someone at GOP spin control thinks the issue important enough to tell their people to log into slashdot, see ifthey have mod points and mod people down. And no, I am not making this up, they really do do that sorta thing these days, pretty pathetic eh?
Kinda makes the pissing contest between Kleinrock et. al. over who was the sole inventor of packet switching sound small beer. I mean who actually cares who did what? None of them invented the whole thing, the fact that there are three possible contenders points to the fact that however necessary the design step was, it would have been discovered somehow at that time. It is not like Hawking radiation or Einstein's relativity which could easily have waited several decades longer but for a flash of inspiration.
It is a shame that the GOP have to pay dweebs to sit in front of computers to repeat lies. Gore never claimed to invent the Internet.
Gore correctly and truthfully took credit for getting us the money to develop the Internet. He was also very helpful in the development of the Web. The endorsement of the Web by the Whitehouse had a massive effect on commercial use. It was also the final nail in the coffin for 'Interactive TV'.
Of course if a lie is repeated often enough people will eventually mistake it for the truth. This particular lie was invented because the GOP was frightened of the comparison between Gore who had achieve a lot and their empty suit of whom the best they could say was he would not interfere with his advisers.
The key to the cost comparison is that RAM supports more queries per second than disk. Supporting the number of queries per second using disk would require a lot more duplicates of the data to support the query rate.
The cost differential between RAM and disk has been eroding for some time, particularly if you compare RAM with SCSI disk. While the price of IDE had dropped, SCSI is still premium priced for the business market, even though there is no reason why a SCSI controller should cost a cent more than IDE.
A 80Gb SCSI-160 drive costs $800, RAM costs $150 for a 512Mb DIMM. So Disk costs $10 per Gb compared to $300.
The problem with the raw comparison is that you still need a lot of RAM to service a large disk, caching etc. There is also a limit on the amount of disk data one CPU can effectively manage. From experience I can asure folk that that limit is certainly less than 80Gb if the lokups are frequent!
So when you add the cost of a CPU and box into the equation the RAM solution is gong to look much better. I doubt that a single CPU could effectively manage more than 4Gb of disk data, but 4Gb of RAM data is quite viable. And you probably need at least 1Gb of RAM to support the disk data in any case so the all RAM solution looks good.
For most database applications RAM wins hands down. On top of the cost of the disk you have to count on
The cost of an Oracle license ($100K +++)
The cost of a whiny Oracle DBA ($100K/pa)
The cost of an equally whiny SQL programer to interface your code to the crack pot SQL data model ($100K/pa)
The cost of licenses for GUI based schema design tools etc. etc. for the whiny SQL types
Trips to CostCo for Malox
The main problem for the RAM route is getting persistence on transactions. So you need some secondary storage in case of power failure or disaster. This could be tape, but ironically disk is cheaper to run these days than almost all tape systems. A 40Gb cartridge for a tape drive can easily cost $150, which is more than an IDE disk drive that outperforms on practically every level (probably even longevity).
The key is that you use your secondary storage to write out the transaction log, you don't attempt to maintain the data structure on disk like SQL databases do. For high reliability you use a complete duplicate of the system to provide your first level backup with disaster recovery at a remote site.
Rules out Bush too on all three counts. (And also Clinton (for the cheating))
Hey if Dufus could not be bothered to show up at his base to fly really kewl military fighter jets what makes you think he would be interested in going into space?
Besides they recently added a pretzel test to the medical. If you have a dodgy ticker the last thing you should be doing is going in for unnecessary G-forces.
If you'd read the article, you'd realize that that $0.02 per hour applies only to service providers. They aren't going to charge owners of consumer devices an hourly fee to use their players
I have extensive dealings with that industry. I don't do the actual business negotiations but I am in a position where they would make their requirements known.
I don't think that they are likely to accept the type of usage fees suggested. If I were in their situation I would never accept them if I had the option of using MPEG2 instead, even if the choice cost me money in the short term.
It is a visibility issue as much as anything. I would not allow anyone else to know how much I was streaming or allow them to put their monitoring infrastructure in my racks.
The majors distributors must realise that they can get better terms by simply waiting to deploy.
I once shared an office with Jim Gettys who declared that he had to have dual monitors on his desk because he was the guy 'who wrote the silly code to do it' [in X-Windows].
Since then I have been looking for an excuse for a second display. Until recently however the thought of paying for dual 18" LCD displays was just too much and now the model I have is no longer made so if I bought a second one it would not match. Like what is the point in having kewl stuff if it looks crappie? Also the demise of 3DFX means that I would also have to get a new PCI monitor card to drive the thing.
I agree with the other posters about not really wanting my PC built into my display. My computer system lives out of sight about 5 ft from my desk and is connected to the desk by 2 cables, the monitor cable and the USB cable. I have a USB keyboard, mouse and CDROM drive on the desk
Idealy I would move the computer into another room altogether 'cos the fan is pretty loud.
I think that before I start spending more money on decorating the office that the NASDAQ needs to go up above 3000 or so.
A while back some of us were sitting round at the IETF discussing the various techno gadgets we had bought recently. Jeff Schiller went first and showed everyone a ham radio the size of a matchbox. Then Steve Bellovin showed people his watch running Linux. All I had was a cheezy Palm VII running a Lisp machine emulator.
The last guy to go is one of those crypto dudes who wears all black. He holds out his hand and taps his palm a few times. Then after a brief pause he starts speaking to someone as if on the phone, which it turned out he was, this dude had a cell phone implanted into his palm and skull!
Anyway we continued drinking for some time (it was IETF after all) and the dude asked us to watch his laptop for a while while he went to the little boys room. We had some more drinks and were about to leave when someone pointed out that the dude had not returned yet. So I went off to the bathroom to find him.
I find the dude bent over the can with his legs stretched out and a bog roll stuck up his ass. Immediately I think the dude has been mugged. "Hey whats up, you OK?" I ask. "No I'm fine", the dude replies "I'm just waiting for a fax".
Not according to this [surrey.ac.uk] it's not. The patent was granted in 1985. 1985+20=2005. Or you could read at Unisys [unisys.com] itself.
Bzzt! The law is 17 years from issue or 20 years from filing, whichever is longer. The filing date on 4,558,302 is June 20, 1983, the issue date is December 10, 1985.
So the patent expires on the later of Dec 10, 2002 or June 20 2003.
What I don't quite understand is why anyone would use MPEG 4 under the proposed license instead of MPEG 2. Chances are that devices will have to support MPEG2 far into the future. It is not very likely that MPEG4 will offer such a devastating improvement in performance that many will be paying 2 cents an hour.
The problem with razor and blades type business models is that they are only good for the seller. Given the choice the customer will strategise to avoid razor and blades type models. Polaroid has gone chapter 11 because people prefer to pay $300 for a good digital camera that costs $0.00 per shot rather than pay $30 for a Polaroid camera and $1 for every shot they take.
Slashdot Mirror
A while ago I did a comparative study of the graphical design tools on which UML is based. My conclusion was that the idea was a pretty bad one and all of them became more trouble than they were worth as they attempted to track every feature of C++ or such graphically. As the projects grew in scope the diagrams became less and less useful.
I was recently forced to use UML, it appears to me to be worse in every important respect than it's predecessors. In addition to the added complexity of now tracking multiple languages UML has lost any coherence the input languages had. UML certainly does not fit well with XML Schema which has a particularly complex data model. In the end I rolled my own graphical markup which people seemed to like but it probably worked because I was using it to present a design rather than create one and I absorbed a bunch of our coding conventions into the notation so the notation was for the subset of XML Schema we used rather than every last feature.
I have noticed that UML and its ilk tend to appeal to people who are brought up on databases and make the mistake of thinking the entity relationship model is useful. Predicate calculus and typed set theory are vastly more powerful in my experience. If I see a bunch of schemas written in Z or VDM I can understand them pretty quickly, I can also see the mapping from the schema to code.
As to the original question, it appears bogus to me. Much more important that what is the most fashionable and feature rich language is what language is going to have support over the life of the project. Java is a definite, C# is almost but not quite guaranteed to be arround.
I syspect that the question is not posed to get an honest reply. The question appears much more likely to be intended to beat the drum for operator overloading.
As such it is worth remembering that Java abandoned operator overloading for good reasons. The C++ approach was just too hard to optimize and lead to buggy and unreliable compilers and code. C# may have got the mix right, I seem to recall there being some limited operator overloading mechanism with a lot of restrictions.
The USPTO uses two contradictory arguments acording to need. When the complaint is made that they issue trivial patents they claim that they cannot perform an adequate review. When the point is made that every other country requires publication and a public opposition period prior to issue they claim that their review is more than adequate.
I doubt that the case will succeed. The patent claims are for a specific type of hypertext linking model which is actually the opposite of the URL method used in the Web. What BT 'invented' was the Prestel database driven hypertext model in which all the links were managed in a central database. The innovation of the Web was to discard the stupid database, and the whining hypertext weenies and embed links directly into the page.
While the Web approach appears 'obvious' today, back in the early days of the Web the hypertext community wrote long boring articles on why central databases are the only way to go. Tim could not even get published in the hypertext litterature, until the next year when they asked him to be the keynote speaker.
Prestel served up a group of numbered pages that could reference other pages by the page number. While this was an innovative idea in 1603 when the idea was first proposed the use of page numbers to reference other parts of a text has been standard practice for at least 350 years. It certainly was not the Web and certainly was not hypertext as most people understand it.
Please note that the folk going on about BT's status as a public/private company are both off base. At the time the innovative Prestel system was invented and the patent filed, BT was a government monopoly and part of the Post Office. BT only began their moneygrubbing patent demands after privatisation.
Actually there are lots of professional services companies with the same approach. A lot of legal, accounting and engineering practices have a 'no growth' or 'slow growth' policy.
The reason is that professional services companies do not become more profitable as they become larger. A PSO company sells the professional skills of its employees. In a typical PSO company there is a pyramid of expertise. For each senior partner there might be two junior partners, four associates and the same number of support staff. That ratio does not change much if there are a hundred senior partners or one.
The problem for the PSO company is that a senior partner has to be paid pretty well or else they will go off and start their own outfit. At the bottom of the tree an associate might be paid $400 a day and be charged out at $1,600 - a markup of 400%. But at the top of the tree a senior partner is likely to be paid more than their charge out rate, their real job being to bring work into the practice. A senior partner might charge out 100 days a year (i.e ahlf the normal rate) at $5,000 a day, but they can keep busy 4 associates at a profit of $6,400 a day and 2 juniors at a profit of $3,600 a day, so they actually contribute to the practice an average of $12,500 a day and take out maybe $6,000.
The reason PSO firms have to grow is that as the people at the base of the pyramid get more experience the only way of maintaining profitability is to either increase recruitment at the base or to restrict the number of promotions. That is why a lot of accounting companies have 'up or out' policies. If you don't make the next step in the ladder by a certain date you are told to look for work elsewhere.
All in all VC should never be funding pure PSO companies. Most PSO companies are organized as partnerships for good reason - the company itself actually has very little value, the value is all in the knowledge of the employees. And those employees can and will walk out the door with it unless they feel they are rewarded for it.
PSO can add a lot to a growth company's bottom line, but only as a supplement to product, not as a replacement. In general the markets tend to be sniffy about companies that make more than 20-25% of their revenues from PSO. An Oracle or an IBM can make a lot of money from PSO, even run a pure PSO division. But that works because the customers (and consultants) know that they are buying more than just the consultant's time.
The other problem with PSO is that it is economically a pretty risky proposition. A typical PSO company operates on a margin or 20%. If there is a downturn they don't have much scope to cut costs without layoffs.
Wrong. The Euro is not gold backed. If it was it would not have collapsed in the year after introduction while the price of gold was rising.
No industrialised country can function with a gold backed currency. In an industrialised country the economy expands with improved production over time. But the production of yellow metal is pretty constant. You have a quantity that has exponential growth tied to one that has only linear growth. The result is an economic slump.
This is why Karl Marx wept when he heard of the '49 gold strike. The additional gold supply would allow the capitalist economies to expand and the revolution would be delayed for a decade or more. The theory of the inevitable communist revolution was based in large part on the belief that the gold standard was essential.
Not really, true elite programmers and system architects are very rare and their productivity is massively greater than that of a merely 'good' or 'average' hacker.
If an elite programmer has a problem getting a job it is because they have a personality problem or a substance problem that is so extreeme they are not likely to offer much to an OSS collective either.
That does not mean that there are no elite programmers doing OSS stuff, but most of the ones I know are doing OSS because they made a bundle from the dotcom boom and don't know what to do with the rest of their life.
I don't think that mere incremental improvement of OSS products is going to threaten the paid software model. For MS Office to be displaced there has to be a discontinuous change, not an incremental one. Matching Word feature for feature as you suggest is a pointless waste of time. Mr Softy has shown that he knows how to win that game.
I hate the weenie aspect. There are plenty of OSS folk who are reasonable and knowledgeable. At one time I wrote a lot of OSS code.
Then there are a bunch of maga-bores who have adopted OSS as their life message and religion. Talking to them is like talking to a born again Christian or a Taleban fanatic, everything comes back to their own all embracing pet theory. It is like talking to a Libertarian or a Marxist, they have an economic theory and it explains everything and what is more if you don't agree with them you are [damned to hell for eternity/ a legitimate target for murder/ a complete imbecile/ an oppressor of the masses].
I don't think the talk was actually anti-open source, the guy took potshots at plenty of people's sacred cows. But the only one that slashweeniedom is going to talk about is OSS because it is soooooooo interesting
The problem is with what the weenies think, the problem is weenieism. As Karl Popper pointed out most of the misery in the world is caused by people who have an absolute belief in their possesion of an absolute truth.
I gave up on OSS once it became popular and the weenies moved in.
An absurd fallacy. Perhaps for fetchmail or hello, world! or other, similarly sized projects, but nowhere else.
The cause of most longstanding security bugs is almost always a major error of design that is not easily corrected.
Take for example sendmail and Outlook, two programs that have become the Typhoid Mary of security. The problem in both cases is a failure of code/data separation. Outlook encourages users to open random content sent from random addresses and we wonder why viruses are a problem. Sendmail has a crazy macro processing language built in that is completely unnecessary for managing SMTP, the numerous buffer overflows in this code only serve to hide the structural problem.
The other problem that the talk highlights is that ideology is the cause of the problem and not the solution. sendmail has been opensource for almost twenty years and shows no signs of being fully debugged yet - and that despite the fact that it is an application whose requirements have decreased over time. Who cares about routing bang path mail or X.400 these days? The code to support that stuff should have been excised years ago.
The reason I like source code is that it allows me to extend a program to do something I need it to do that it doesn't do yet. Even then I would much prefer a well written API to source code, but source code plus API plus documentation is best of all.
That is not what I said. Java only allows you the choice to use it or not. If you try to develop a version of the language that meets your needs that Sun does not like, Sun will sick lawyers on you.
Again total bullshit. If it were only Sun's interest the JCP wouldn't exist. If it were only Sun's interest Oracle wouldnt have created a native compiler for Java in the Oracle DB.
My experience of JCP is that Sun decides who gets to chair your group then the chair goes off to do what they damn well please. It is not an open process at all. Oracles's java compiler is clearly in the interests of Sun as well as Oracle.
There is nothing to prevent people writing to a portable sub-set of C# if they chose to do so. That is very different from Java which denies choice, programmers are forced to program to the lowest common denominator.
The problem with Java is that the language will always be limited to the subset of functionality that it serves Sun's interests to provide. Optimised i86 code and direct access to native O/S features don't meet those interests. So they will never be first class features of the language, even if Sun does not sick its lawyers onto companies that try to provide them.
While C# reflects Microsoft's interests those are much more closely tied to those of developers. It is not in Microsoft's interest to limit functionality in any way.
I really would hope that Joy would be above stooping to a FUD strategy. But apparently not.
What Joy does is to conflate two separate issues. The first is network code, the second is adding runtime code safety to C.
Nobody disputes the fact that network code should be safe. Microsoft's security policies don't let you download unsafe code by default and the CLI runtime checker kicks in in precisely the same way it does in Java.
That is not the same as saying that there is no use in any circumstance for unsafe code. The vast majority of the Linux kernel is 'unsafe code' if you want to be pejorative. Java simply refuses to ever let you write a piece of code of that type in Java and tries to make interfacing to that type of code as hard as possible. Microsoft on the other hand have rather a lot of legacy C++ code that they want to access even though they can't rewrite every line in C#. I think that Microsoft's interest here is closer to that of most developers.
As for the MIME type FUD, it is a real pity that sun deliberately bypassed the MIME type system in HTTP to prevent firewalls being able to block Java. I made the protest at the time, they ignored it.
Equally criticising Microsoft for not implementing the Javascript security model is a bit rich since few people outside Netscape would agree that Javascript has a security model. By their own admission Javascript was thrown together in a fortnight and thrown out the door with almost no checking.
Sun would do much better to consider how they are going to survive as a specialist hardware manufacturer as Linux comoditizes the UNIX workstation market rather than spend their time engaged starting stupid arguments with Microsoft. As Balmer himself points out lots of companies have lost their way by focussing on Microsoft's business model instead of their own.
The problem with Java is that it is a closed, proprietary language whose primary design criteria has become 'get Microsoft'. In the process Java has been deliberately made less useful to windows programmers, which means the vast majority.
I think that Sun is playing right into Microsoft's hands by balkanizing software development into Java and Windows camps. History suggests that the larger developer pool will win out in the end. A stupid, stupid strategy.
Communicating Sequential Processes. It is an algebra for describing parallel processes that communicate with each other. The key feature of CSP is that when the processes communicate they rendezvous, that is if A is trying to talk to B and B is not ready then A will wait until B is ready, contrawise B will wait if A is not ready.
The main application of CSP in program language design was the inmos Occam programming language.
The advantage of CSP is that it is much easier to write parallel programs because communication between processes is handled in the same way as communication between processors. The type of structures that have to be kludged up out of semaphores and shared memory in UNIX/C or Windows/C can be provided as native language features.
What CSP ammounts to is a start on developing the equivalent of structured programming constructs for network computing applications. In practice you need to add a few extra structures such as FSM support and structures to handle termination and fault handling is a decent fashion.
I am not surprised at the breakdown of the messages, except that the number of messages rejected as 'opinion' (7,000) sounds rather low if anything. The number of form letters (3,000)also sounds like it on the low side.
I doubt that anyone in the administrationis going to treat the messages as 'votes' [what start a lawsuit to stop them being counted? - Ed]. The number of messages on both sides will have been inflated by 'astroturf' (fake grass roots) campaigns by Microsoft, Sun, AOL etc. Fortunately messages of that type tend to be easier to spot than the people who purchase the campaigns think.
The bulk of the messages will simply repeat each other and standard positions fed to people by the media (including slashdot). I suspect that the 48 'substantive' comments are mainly the briefs written by industry lawyers to support one party or another. I strongly suspect however that it is the case that practically every idea expressed in the 22,000 contributions is covered in the 48 'substantive' contributions. Identifying a small number of contributions that put all the important issues well is a tremendous service to people trying to read the materials.
Taking the feedback as email will have helped sorting to an enormous degree. But a structured forum with some form of moderation could have helped the feedback further, collapsing repetative positions down to one instance and such. The moderation need not have been on the slashdot model in which there is a single pool of moderators, there could be twin panels of moderators representing each side. After all posting troll comments and pornography would do nothing for either side unless they wanted to discredit the dabate.
Finally the cost of publication at $400 a page does not seem unreasonable, it is roughly equivalent to the cost of printing and distributing about 1,000 copies. That is not much more than one per senator, congressman, state AG, party affected and news organization.
I don't think anyone should get a Turing for XML, the inventive step took place in SGML about 15 years earlier. XML is simply a cleanup of SGML that removes incompetence, idiocy and illogic from the SGML design. The basic principle is the same though. SGML was too badly executed to give Goldfarb any awards.
Rather than give Dahl and co an award I would much rather people looked at their ideas and acted on them. People slap each other on the back over their use of 'object oriented' languages, but Java and C# both offer only a small part of the power of the OOP model and C++ offers the wrong parts.
The modern OOP languages offer only the data structuring concepts, Simula also had a message passing concept which has largely been lost. It took me ages to work out that the reason the message passing explanation in the C++ book was so hard to follow is that it is actually irrelevant to C++.
A true fine grained parallelism model with message passing would be very useful at this point in time when a lot of programming projects involve networks and GUIs. The pthreads model is too low level. It is a pity that the developers did not take note of the fact that a year after inventing monitors, Hoare invented CSP which is a more powerful model.
To give concrete examples of the benefit of deep grained parallelism and message passing consider how often Windows halts waiting for input in some dialog box in a different window or how often an X-Window will be left unrefreshed during processing. The reason is that doing the job properly requires multithreading and the pthread model is to cumbersome for most developers to deal with.
I think the problem is that people were misreading the numbers both ways. The use of the raw Bugtraq numbers against Windows was always a canard, the use of the raw figures in the reverse direction is a canard.
The article's argument is sufficient to demonstrate that the 'Linux is more secure' argument is false, but insufficient to prove 'Windows is more secure'. As you point out few bugs are found on the MacOs, that is not surprising since these days they hold MacWorld in a telephone booth and in any case just how many security holes did Edison have in his desk lamp?
The problem is that security really is complex, it certainly is not a linear issue and it is completely determined by your operating environment. An O/S configuration to secure server will almost certainly prevent a user getting useful work done.
As a security professional who is pretty well known in the field, I can tell you that both the Linux is more secure and Windows more secure religious arguments are wrong.
Windows cannot currently compete in the real high end security configuration where we strip down the O/S to run only the services that are absolutely essential. However Microsoft make no secret of the fact they are working on a platform of that type. If I could find a way to audit that work I would rather buy a secure kernel in than have to spend seven figure sums doing the strip down in house.
The multiple eyes argument in favor of Open software only works up to a point. The problem is that you rely on the defenders being more vigilant than the attackers, that is not always the case. Although these days the trend in hacking has been to go for the binaries rather than muck arround with source code.
The biggest problem with Windows is the predeliction for supporting active code in email messages. But Microsoft is not the only company that does not understand the importance of code/data separation. Sun and Netscape have both been guilty of equally eggregious abominations.
That is a consequence of the C2 standard which was written by the military for the US govt in the days before networking.
C2 was obsolete before the Web existed. Back in 1993 when I was asked to do a security audit of the Web standards against the Orange book I concluded that the standard was no help at all.
The other reason that C2 is not very useful is that the main concern in Orange book is partitioning multiple users data on the same machine. These days each user has their own machine, a one person computer that does not meet C2 mandatory access control requirements can be perfectly secure - look at a Palm or Pocket PC or a smartcard.
The main reason that Mosaic was the first mega successful Web browser was that it was the first to use a GUI toolkit that did not look like crap. It was not actually the first Motif browser but it was the first with Motif look and feel. The other browsers looked like science projects.
Ten years later it is quite possible that Motif's time has come and gone. The Motif look is somewhat dated and the OSF licensing model is certainly dated. Unless it was released as open source sometime I didn't notice you still have to pay for Motif which pretty much rules it out in the Linux world. I don't think that the chances of survival for the non-open source Unix world are very good these days.
What puzzles me is that these toolkits still need a rack of twenty manuals each of which is six inches thick. Its only a goddam menu system!
There is a very determined group of people who want to project their particular version of events as 'the truth'. So someone at GOP spin control thinks the issue important enough to tell their people to log into slashdot, see ifthey have mod points and mod people down. And no, I am not making this up, they really do do that sorta thing these days, pretty pathetic eh?
Kinda makes the pissing contest between Kleinrock et. al. over who was the sole inventor of packet switching sound small beer. I mean who actually cares who did what? None of them invented the whole thing, the fact that there are three possible contenders points to the fact that however necessary the design step was, it would have been discovered somehow at that time. It is not like Hawking radiation or Einstein's relativity which could easily have waited several decades longer but for a flash of inspiration.
It is a shame that the GOP have to pay dweebs to sit in front of computers to repeat lies. Gore never claimed to invent the Internet.
Gore correctly and truthfully took credit for getting us the money to develop the Internet. He was also very helpful in the development of the Web. The endorsement of the Web by the Whitehouse had a massive effect on commercial use. It was also the final nail in the coffin for 'Interactive TV'.
Of course if a lie is repeated often enough people will eventually mistake it for the truth. This particular lie was invented because the GOP was frightened of the comparison between Gore who had achieve a lot and their empty suit of whom the best they could say was he would not interfere with his advisers.
The cost differential between RAM and disk has been eroding for some time, particularly if you compare RAM with SCSI disk. While the price of IDE had dropped, SCSI is still premium priced for the business market, even though there is no reason why a SCSI controller should cost a cent more than IDE.
A 80Gb SCSI-160 drive costs $800, RAM costs $150 for a 512Mb DIMM. So Disk costs $10 per Gb compared to $300.
The problem with the raw comparison is that you still need a lot of RAM to service a large disk, caching etc. There is also a limit on the amount of disk data one CPU can effectively manage. From experience I can asure folk that that limit is certainly less than 80Gb if the lokups are frequent!
So when you add the cost of a CPU and box into the equation the RAM solution is gong to look much better. I doubt that a single CPU could effectively manage more than 4Gb of disk data, but 4Gb of RAM data is quite viable. And you probably need at least 1Gb of RAM to support the disk data in any case so the all RAM solution looks good.
For most database applications RAM wins hands down. On top of the cost of the disk you have to count on
The main problem for the RAM route is getting persistence on transactions. So you need some secondary storage in case of power failure or disaster. This could be tape, but ironically disk is cheaper to run these days than almost all tape systems. A 40Gb cartridge for a tape drive can easily cost $150, which is more than an IDE disk drive that outperforms on practically every level (probably even longevity).
The key is that you use your secondary storage to write out the transaction log, you don't attempt to maintain the data structure on disk like SQL databases do. For high reliability you use a complete duplicate of the system to provide your first level backup with disaster recovery at a remote site.
Hey if Dufus could not be bothered to show up at his base to fly really kewl military fighter jets what makes you think he would be interested in going into space?
Besides they recently added a pretzel test to the medical. If you have a dodgy ticker the last thing you should be doing is going in for unnecessary G-forces.
I have extensive dealings with that industry. I don't do the actual business negotiations but I am in a position where they would make their requirements known.
I don't think that they are likely to accept the type of usage fees suggested. If I were in their situation I would never accept them if I had the option of using MPEG2 instead, even if the choice cost me money in the short term.
It is a visibility issue as much as anything. I would not allow anyone else to know how much I was streaming or allow them to put their monitoring infrastructure in my racks.
The majors distributors must realise that they can get better terms by simply waiting to deploy.
Since then I have been looking for an excuse for a second display. Until recently however the thought of paying for dual 18" LCD displays was just too much and now the model I have is no longer made so if I bought a second one it would not match. Like what is the point in having kewl stuff if it looks crappie? Also the demise of 3DFX means that I would also have to get a new PCI monitor card to drive the thing.
I agree with the other posters about not really wanting my PC built into my display. My computer system lives out of sight about 5 ft from my desk and is connected to the desk by 2 cables, the monitor cable and the USB cable. I have a USB keyboard, mouse and CDROM drive on the desk
Idealy I would move the computer into another room altogether 'cos the fan is pretty loud.
I think that before I start spending more money on decorating the office that the NASDAQ needs to go up above 3000 or so.
What, do you mean Civ2 or Civ1?
I have known many students who spent many hours on Civ, I never heard of any getting course credit for it, still less majoring in it. Cool!
The last guy to go is one of those crypto dudes who wears all black. He holds out his hand and taps his palm a few times. Then after a brief pause he starts speaking to someone as if on the phone, which it turned out he was, this dude had a cell phone implanted into his palm and skull!
Anyway we continued drinking for some time (it was IETF after all) and the dude asked us to watch his laptop for a while while he went to the little boys room. We had some more drinks and were about to leave when someone pointed out that the dude had not returned yet. So I went off to the bathroom to find him.
I find the dude bent over the can with his legs stretched out and a bog roll stuck up his ass. Immediately I think the dude has been mugged. "Hey whats up, you OK?" I ask. "No I'm fine", the dude replies "I'm just waiting for a fax".
Bzzt! The law is 17 years from issue or 20 years from filing, whichever is longer. The filing date on 4,558,302 is June 20, 1983, the issue date is December 10, 1985.
So the patent expires on the later of Dec 10, 2002 or June 20 2003.
What I don't quite understand is why anyone would use MPEG 4 under the proposed license instead of MPEG 2. Chances are that devices will have to support MPEG2 far into the future. It is not very likely that MPEG4 will offer such a devastating improvement in performance that many will be paying 2 cents an hour.
The problem with razor and blades type business models is that they are only good for the seller. Given the choice the customer will strategise to avoid razor and blades type models. Polaroid has gone chapter 11 because people prefer to pay $300 for a good digital camera that costs $0.00 per shot rather than pay $30 for a Polaroid camera and $1 for every shot they take.