Whatever.NET ends up being, Java and Corba had most of it first...that means there are a lot of experienced people out there who could give MS a real good reaming on this architecture.
Unfortunately CORBA had too much of it, most advances in Comp Sci are taking stuff out, not adding stuff in. Java is better than C++ mainly because it took out some really kludgy stuff. C is derrived from B, a simplified version of BCPL which is a simplified version of CPL.
CORBA had several major problems that everyone told the CORBA people about but nobody ever did anything about. Like basing the whole messaging scheme on RPC - subordination is a massively inefficient way to do parallel processing.
The biggest problem with CORBA is you had to buy this object broker and then pretty much rewrite your application to use it. With dotNET the idea is much more like the Web, leave the information servers be and write a thin layer interface to them so you can get to the data from wherever you need to.
I see CORBA as being the SGML to dotNet's XML. It may well be that you can do everything you can do in XML in SGML and some more. But SGML has some real deep seated structural limitations that keep biting you when you try to use it for real stuff. Writting the HTML DTD was pretty miserable.
.net isn't going to use Java cause of MS loosing that lawsuit to sun a wile back, if I remember correctly, MS could use Java in existing products and support it up to 6 years but no more new products.
Microsoft settled the case and would probably argued that it won the points important to it. The big outcome of the case was that the constract between Sun and MSFT was terminated. MSFT are not prohibited from developing a clean room version of Java, e.g. the stuff they bought in from HP and others.
However Microsoft is unlikely to want to lend its support to Java again since if they make extensions that Sun disapprove of Sun can still go and bully third party tools developers that copy them. Sun by its own insistence has the lead in the Java world, so Microsoft has gone to play in its own sandpit.
The big problem with Java in my opinion is that its main selling point - processor independence was something that only Sun really cared about. It has never been a big thing having to recompile code for different processors and most people use Intel in any case. In place of DLL hell I have runtime hell in which I have to make sure I download the right 20gazibabyte run time from Sun to run the program.
I think most people are missing out on the biggest likely impact of C# which in my view is on the Visual Basic side. VB has for years been the worlds most widely used programming language even though everybody agrees that it is something of a mess. Thing is however it has picked up some nice features along the way - particularly insensitivity to early or late binding.
I see C# as being more the union of the VB feature set with the C feature set than a direct competitor to Java. Being able to convert legacy VB applications into a more maintainable base sounds to me like a good thing.
Given that the site has been well and trully slashdotted...
Comparing implementation of the same code in different languages is one thing, measuring their run time is irrelevant unless you also test every compiler out there.
The only point at which computing language design has a big impact on performance is in the ease of optimization. More modern languages tend to be designed to make it easier for the compiler.
If you are using a scripting language and compare compiled code in C against interpreted code.. well duuuhhh! That is not a language issue, in fact one of the party tricks of certain Web servers is to compile server side scripts inline and cache the object code...
Equally comparison of functional languages with declarative is pretty meaningless. If you are doing matrix multiplication tasks and need good performance above all then maybe LISP is not the best choice.
The other issue that comes up is that on simple tasks there can be a massive difference in the ease of coding. Matrix operations can typically be done in one line in APL. But when you get to larger tasks that are beyond the core of any language the differences tend to be much less significant.
Robert Burns once observed that there is no point in settling the precedence between a flea and a louse. I think the apphorism applies here since apart from allowing the slashcrew to rerun Bill-as-Borg I can't see the story.
Kodak seem to be upset only because Microsoft got in the way of their attempt to tie their online prints service to their camera. At least Microsoft gave companies that do prints but not cameras a way to get in on the action.
The idea of charging for browser slots is not new. Netscape made almost all their money from the browser that way. AOL is simply one long infomercial for AOL's "partners".
Personally I get real peeved when companies mess with my file associations, particularly when they don't ask me. If there is a way to stop such meddling in XP I am all for it.
Rather than displaying the page sent back from the server (like a good HTTP-compliant browser), IE provides it's own error page
Bzzzt! Untrue. The HTTP specification explicitly allows error codes to be replaced.
If the error codes had to be reported as generated it would prevent browsers implementing error codes in the users own language. If I go to a french server I don't want to have to use Babel fish to read the error response.
If the intention had been to simply display the error message from the server then we would not have bothered with error codes at all. The purpose of the error codes were to allow the client software to do something more useful to the end user.
You as the server writer might not like the error code handling of the client, but that is explicitly outside your remit. The client controls the user interface, not the server.
dotNet has many parts. Some of which Microsoft themselves would probably write an open source version of if nobody else did. C# for example is a lot better to use than C++. Microsoft wants to break Sun's proprietary lock on Java. To do that they realise that they need to get a C# interface on GCC.
Other parts of dotNET are already being open sourced. SOAP for example is being cloned all over the place. It is pretty clear that someone will have an Apache version.
The part of dotNET that would be hard to copy (and probably not that worthwhile) is the rack of API goodies that ship with windows. But that is not what a clone of dotNET would be useful for.
The main reason I would want a version of dotNET running on another platform would be because the whole big idea is network computing. If you have a legacy system running on a mainframe you don't need to move it onto a different platform just to use it as a Web service. Equally you might want to extend some embedded system and make use of a Web service on another box that is being set up.
Microsoft is trying to sell Windows as a big feature set development platform for building dotNET services. To succeed they need dotNET services to be available on other 'legacy platforms'.
Microsofts complaints about the GPL concern the viral aspect.
The terms were that he must not break the speed-limit. If he did, they would fine him. He broke the speed-limit, and (according to the contract) Acme fined him for it. It's as simple as that
Please save us the libertarian hysterics. It is not as simple as you say. Contract law is not anything like as simple as you appear to believe, otherwise contract law lawyers would be much cheaper to hire.
The enforceability of contract terms is subject at all times to state and federal law. Don't like that liberweenie? well tough. Without the government courts to enforce the contract in the first place there is no contract.
The argument that it is all down in black and white does not move me. I do not believe that in this instance that the car rental company was honestly representing the contract terms. Nor does the dept of consumer safety. Bad faith has been an issue in contract law since the Romans invented the concept.
In the case in question the car company wrote the contract, they are thus on the hook in the case of any ambiguity. In this case there is good reason to doubt that the customer intended to agree to the specific term that he would be fined if the speed of the car went over a certain amount. The contract term in bold can be quite resonably be interpreted as meaning that there is a $150 surcharge if the police issue a speeding ticket.
Above and beyond the contract issues the idea is simply bad business. It is not uncommon to find US companies that believe that dishonest and underhand business techniques are the way to make profit. However even P.T.Barnum later observed that he made very little money when he was dishonest, he did much better when he put on a show that was worth the entry fee.
Look if you didn't notice that the company had its fingers crossed when they wrote the contract then that is your problem, understand your problem I will spell it out for you: Y-O-U-R P-R-O-B-L-E-M.
As a capitalist corporation Netpliance was Fully entitled to remove as much cash from your account as they felt like, as a corporation they are by definition in the right. Look at it this way how many times did you go over the speed limit after you bought the machine, under the Acme rentals argument that would entitle netpliance to fine you $150 per time...
Uhh ohh, the ray gun appears to be wearing off, I think I will go and have a nice little rest and read The Fountainhead or Atlas Shrugged a couple of times.
If there aren't enough paying punters for Ricochet then my guess is that G3 will bomb so baddly that it takes out several major carriers with it.
My strong belief is that there is a connection speed tipping point. The Web was miserable over 9.6Kb modems, barely acceptable at 14.4K and not too unpleasant at 28.8. Now I love my cable modem and 1Mb/sec but I would not wait arround while someone built it if I could get 28.8 somewhere else. So if there isn't enough demand to support Ricochet on its own what chance is there that the demand for G3 will be big enough to support five carriers in the UK or the rest of Europe?
Admittedly Ricochet is less usefull because it is not ubiquitous yet. But I don't think that is the reason there is a problem. CPDP didn't sell too well either. I think that the market for mobile data has been massively overestimated. Or at least the takeup rate has.
Perhaps better handheld devices will help. At the moment the PalmVII is a crippleware Web interface tied to a very slow network (the 9.6K claimed speed is total horse shite to use the technical term, try 0.2K if you are lucky.) The PocketPC series are much better hardwarewise but getting wireless service is pretty difficult and expensive. Plus PocketPCs drain the life out of a Lithium ion battery fast enough without wireless on top.
The Fact is that i-opener were completely within their rights to bill people even if they said they wouldn't. People should read the contract! They didn't, they got billed, so what, so what if the contract didn't mention being billed, they wouldn't have read the contract anyway.
This sort of thing really p****es me off. Its like them goddam environmentalist tree huggers who want to stop the oil companies drilling for oil on the national mall, who cares if there isn't any oil there? Its a free country isn't it? And in any case if the national mall was that good they wouldn't have built a shopping center there now would they?
I think that the Appeals court decision was the least favorable to Microsoft it could possibly have issued. The fact is that Jackson made a whole stream of statements to the press that demonstrated actual bias. If a lawyer had made the same type of statements about the Appeals court they would probably have been disbarred.
Now that the press is reporting that the decision has been reversed the legal niceties of that reversal matter little. The Bush people will settle and Microsoft will give them a nice campaign contribution before the soft money limits kick in.
Rather less of the decision has stood than some Slashdotters appear to think. The case has been remanded back to be heard by a different judge and the Appeals Court has made it clear that the new judge need not take much notice of Jackson's opinions.
The Appeals court has overturned parts of the 'findings of fact' in practically every area. I suspect this is part of a strategy. Under the ridiculous Appeals court rules the idea is that the Appeals court is bound by findings of fact by the lower court - hence Jackson's stated intention to prevent review by issuing the findings of fact as a separate document. If the Appeals Court had vacated the findings of fact directly there would be a possible appeal to the SC. However the Appeals court has essentially remanded the case in such a manner that the new judge is free to reverse as much of Jackson's findings as they see fit.
Specifically missing from the rulling is any endorsement of Jackson's findings, they decline to vacate them, but they don't bind the new judge to them.
I doubt that the case will get as far as a new trial since the Bush people are practically certain to settle. I suspect that the Gore folk would have settled also, although on rather less favorable terms. The terms of the proposed breakup were idiotic, they would have created two monopolies in place of one. I think they had more to do with the size of the lawyers egos than policy sense
You are damn right the NC is the MIS managers wet dream: try to maintain a site with hundreds of windows PCs, have software installed and upgraded on each one of them, do routine maintenance on each PC's hard disk etc etc. Then we can talk.
Unfortunately the MIS managers tended to be as interested in recovering control over the corporate nervous system as control over costs. Back in the 1980s I made money by writing apps to screen scrape data out of corporate mainframes so the middle managers could feed it into Lotus 1-2-3. The MIS managers could have provided the data but they refused point blank to do so. It was all a power-game.
Besides, you DON'T need an 800MHz PIII, 128MB RAM and a 20Gb hard disk on EVERY PC in your network
But when that setup costs $500 from a computer store the premise that the Network computer would save money by reducing the hardware costs collapsed.
The Network computer failed for many reasons, not least the fact that there was no significant application support for the device. The lack of a hard disk was always going to be a drag on performance though. The idea of eliminating local user storage is not a bad one, end users should probably only have access to a centrally managed file store. The idea of no local long term storage was a complete loser however, as anyone who has used a diskless workstation knows.
It is not suprising that network computers were easy to manage. They could not be used for anything useful. What is more Sun and Oracle appeared to be spending more on idiotic TV ads than on writing aps to make them useful. If Sun had bought Star Office earlier they might have had a point.
Ultimately I see the whole Sun vs. Microsoft thing as denial on the part of Sun. Microsoft is not going to kill Sun, but Intel and Linux will. A high end linux box can be bought for a tenth the cost of the equivalent Sun box.
Also, C# is not an evolution of C++. C# is java under a different name, what microsft thinks java should have been.
Not actually true, If Sun had not behaved as it did then C# might have evolved out of Java. C# is actually built on top of the C++ compiler back end by the C++ team. The feature set is essentially a combination of C, COM and Visual Basic. It is possible to translate from Java into C# but it is also possible to translate from Visual Basic to C# so that hardly makes the languages equivalent. Sun is not the first company ever to suggest a new programming language, nor was Java successful for any new features, even the byte code interpreter has been arround on micros for 25 years (UCSD p system).
Sun's behavior with Java was pretty inexcusable. The idea that we should all be forced to agree to the decisions of one company when it comes to the feature set of a programming language is stupid. Sun deliberately crippled the native language interface for their own purposes. But many of the users of J++ simply wanted a better language than Visual Basic to write one off applications and did not give a monkeys about 'running everywhere'.
The author of the article does not appear to know anything about.NET, authentication services or much else.
Contrary to the discussions on this board.NET has nothing much in common with Sun's failled Network Computer. The Network Computer was nothing more than a new name for an X-Terminal that can run java apps and a browser locally. That Sun tried it as a bet-the-business strategy was the best corporate joke since the Sinclair C5.
The Network computer concept was for desktop apps. It was an MIS manager's wet dream - take away the employees PCs and give them dumb terminals wired to the nice shiny mainframe. Back to the 1960s.
Network Computers failed for the very good reason that MIS managers looking to torture their victims could do the same job much more cheaply with Citrix without buying Scott McNealy and Larry Ellison more fuel for their corporate jets.
.NET is about information service access. It is the way to hook together e-commerce applications. It has zero to do with Suns clueless hardware platform.
The only intersection between.NET and Sun's effort is that Microsoft has rolled out the JIT compiler technology and Java alternative as part of the package. Neither is core to the.NET idea, or for that matter Sun. C# is merely a logical cleanup of C++, there are some points of comparison to Java but all the ideas have been arround long before Sun used them. If Sun hadn't got all proprietary closed and legal Microsoft might not have created their own, but nobody can really blame them for not being beholden to a standard Sun police ownership of with lawyers.
Hailstorm is only one small part of dotNET, getting all wound up about it is to miss the plot entirely. Gates is looking to take Oracle and SAP to the cleaners, Hailstorm is the smokescreen for that agenda.
As for the GPL bashing being to divert attention from anything, the covert agenda there is more likely to ram home to the analysts the fact that Linux is putting Sun and Solaris out of business and the poor performance of Sun the past few quarters is probably reflecting that dynamic rather than an overall slowdown.
I think that the ACME company sound like they are a bunch of slimeballs trying to impose a hidden charge. However I don't think that the idea that trucks should have speeding detectors is a bad one.
In the EU all trucks must by law have a tachometer fitted that makes a continuous record of the speed of the truck. It is used to monitor the drivers hours and stop them from driving when they are too tired. Given that the accidents caused by heavy goods vehicles can kill ten or twenty people at a time this is a pretty good idea.
Anything that discourages HGVs from tailgating has to be a good idea. It would be even better if SUVs also had to have the same gear fitted, after all if they get the benefit of the light trucks emissions limits they should be taxed like trucks and have the same speed limiters.
Sorry, that's not the way contracts work. If you sign, you're liable. The "it wasn't in big, bold, underlined type" excuse won't cut it in a court of law.
Clearly YANAL, neither am I but I have spoken at enough ABA meetings to know something about what I am talking about.
Most consumer contracts contain unenforceable clauses. Every cell phone contract in the US contains a clause that waives the customer's right to being a class action. The clause is almost certainly unenforceable, as indeed is any clause that disallows redress for negligence.
In my business we spend a lot of time considering what contract clauses a court is likely to enforce against a customer - in particular a consumer. It is not remotely as simple as being able to hold the consumer to the writing on the page.
There is in fact more than one contract involved here. If you accept credit or debit cards the card association rules trump anything your insignificant company might try to foist on the customer.
The other 'contract' probably isn't. The mere fact that a consumer scribbles their name on a document does not automatically create an enforceable contract. The list of exceptions is as long as your arm.
In this specific instance it does not appear that ACME made any attempt to bring the unusual contract clause to the customer's attention. That coupled with the fact that enforcement of traffic policy is a matter of public policy probably puts the rental agency in a less than optimum position.
If the notice was less than clear the mere fact that the company is trying to contest the case tends to indicate bad faith, otherwise a more reasonable response would be to put more prominent notices on future contracts.
Don't fuck with a postgrad rhetorician minoring in lingustics...
Good advice, postgrad rhetoricians minoring in linguistics are well known to be no good at all at any kind of sex and even Madonna stated after the attempt that she would rather have spent the two minutes and five seconds in question doing something more fun like filing her taxes instead.
The US and the rest of the world abandoned the gold standard in two stages. First during the depression the promise to pay that used to underly the dollar bill or pound sterling was 'suspended'. After WWII a bunch of civil servants in the US and UK set up a system in which the value of national currencies was tied to the value of gold reserves. The Ian Flemming novel 'Goldfinger' is partially about the gold running that took place under that system.
Both systems failed for the same basic reason, there simply was not enough metal to back the amount of money required by a modern economy. If you think about it the idea that the optimum amount of money in circulation should be tied to the amount of a shiny metal that has been taken out of the ground is rather odd.
The US federal government still has ownership of something like 50-60% of the total world gold reserves. Most of that is the payment on war loans made by the British following world war I and II which in turn was the booty of Empire. The total quantity of gold bullion has at most doubled since WWII, in the same time the GDP of the US has in real terms expanded at least ten fold.
There simply is not enough yellow metal to go round. Nixon abandoned the gold standard for the simple fact that even under the system of managed exchange rates there was simply not enough yellow metal to support the economic activity. The supply of gold had become the limiting factor for the economy.
The idea that money should be backed by real value has emotional appeal to many. Back at the turn of the century the Deomcratic party was essentially captured by a monomaniac called Willian Bryans Jennings whose sole speech was 'that man should not be crucified on a cross of gold' - monetary reform by moving to a bimetalic standard.
It is not surprising that people trying to invent their own currencies should attempt to base them on gold. But there is a big difference between having a gold ingot in the hand and having an account with a fly by night operator in St Mcru (pop 5 penguins).
It is now 30 odd years since the US was on the gold standard and the number of people expecting a return is rapidly diminishing. At the same time most other countries have abandoned the gold standard and nobody wants to have a managed exchange rate (although some are forced to). I suspect that the diminishing gold price reflects the fact that gold is loosing its traditional role as a safe haven in troubled times.
As to the OP of this thread who mentioned serial numbers, there ARE instances where the game collectors will go nuts over the 00001 machines. Somewhere on the net is a page that was put together when the very first Tempest was discovered by a collector.
I would not be surprised by the first Tempest machine fetching a high price, Tempest was a first generation machine with some pretty exotic hardware.
I would quite happily pay $25K for an Apple I. But I would not pay more than $50 for an Apple II even if it was an ultra rare model.
I have never heard of arcade game collectors being big into serial numbers. The amount of space they take up means most people have two or three at most. If you have to sell one machine to buy a new one, serial numbers are probably not your priority.
Of course some idiot with a fat wallet may read about the auction on Slahdot and bid the box up, but given the vintage the machine is already way over bid. If the instant buy is $25K the guy probablky thinks he will get $10K at least.
There are demonstrably better abstractions for almost all problems that threads can solve. Co-routines, continuations, event models, message queues, sockets, shared memory. "Demonstrably" means they get the job done, but clearly introduce fewer possibilities for error and are easier to debug
There are better models but the ones you have just listed are the ones that Comp Sci moved away from twenty odd years ago finding them complex and likely to introduce bugs.
The principal abstraction in threads is Hoare's monitors (there is some dispute over who invented monitors, Hoare credits Dijkstra, Dijkstra Brinch-Hansen and Brinch-Hansen Hoare). A couple of years after the monitors paper Hoare developed the Communicating Sequential Processes model (CSP) which was later used as the basis for the occam programming language - which remains pretty much the only mainstream programming language with decent support for parallelism.
Unfortunately the threads programming model does not provide decent abstractions for communicating between threads, except for shared memory, which is severely limiting in the multiprocessor context, there are very good reasons why shared memory bus SMP machines have consistenly had a maximum of 16 processors with 4 being the more common limit.
Incidentally, the Amdahls 'law' that kicks off the piece was originaly marketting propaganda to persuade folk that faster processors (like Amdahl made) could not be replaced vector processing add on boxes. The argument is valid, but the framing of the argument is deceptive. Very few problems have the type of fine grained parallelism that can be exploited by vector boxes, however most engineering problems have parallelism at coarser granularities.
People can get into trouble with threads, but they can get themselves into much more trouble with half baked multi-process tweaks. Five years ago the quality of threads implementations generaly was so poor that multiprocess hacks were the only way to go. Today that does not apply.
But with internet ads, you can see EXACTLY how effective an ad is. If your clickthrough is low, the ad is not effective (at least that is the premise). So the revenues go down. It's that simple.
I don't think that the effectiveness can be measured 'exactly'. It is possible to detect occasions on which an advert has lead directly to a sale. But why should that happen very often anyway? A lot of the time advertising is a cumulative process, you see the ad, register the message and then refer back when you actually need (or want) the product.
I don't doubt that Nike get every penny back that they pay Tiger Woods for hawking their sweatshop produced clothing. However I don't think anyone sees Tiger Woods in his swoosh cap and runs out to the mall to buy one like it.
I suspect that a bigger part of the problem is that the Web changes the nature of advertising from paid placement to attracting trafic to the advertiser's web site. If you open up any hobbyist magazine you will see page after page of advertisements that are simply lists of products and prices. Nobody would place that kind of an ad in Salon, they would have a link to their own site and host the catalog there. The other big advertising money spinner is classified ads and EBay has cornered that market.
"New Media" is the term that people (dare I say like yourself) came up with to describe the Internet. They did this because they were initially unable to comprehend the aspects of it that were not directly analagous to their world.
Untrue on both counts. The term 'New Media' was coined to get people consulting gigs. The people who actually invented the Internet and the Web knew what they were doing all along. The people who had zero clue were the analysts and journalists who spent their time interviewing each other and bilging out puff pieces about 'internet time'. Nine years on and the Web is still a work in progress, so what was that 'internet time' they were talking about?
Most people talking about new media were talking about two things, interactivity and a different cost structure to print or TV. Interactivity is what brings people to Slashdot. Newspapers have always had letters pages, but online forums take the concept much further.
The difference in the cost structure online vs print is dramatic. If you don't have to pay for the content, publishing becomes close to free. If you are a government you can probably save money by putting documents online rather than printing them.
Where people's expectations failed was when they fooled themselves into thinking that new media would lead to new media empires. I don't believe that was ever going to happen and if it did what does it benefit anyone if an old media conglomerate like Time-Warner is replaced by a new media conglomerate like AOL?
We always thought that online new media would be small scale mom 'n pop type stuff with a few medium sized outfits (which it is mainly, look at the prOn sites). When the new media companies started to employ staffs of 100+ the writing was on the wall.
It is sad to see so many titles go down, however the good news is that the advertising business model is viable, the problem is that the business plans that believed the advertizing base would grow fast enough were not.
The Web will still change the media world, it will just take fifteen or twenty years rather than the six months the 'Internet time' cretins blathered about.
Changing peoples way of life takes time, live with it.
However, they had absolutely no powers to do anything about this...
Oh really, tell that to IBM. After the Reagan administration dropped the anti-trust case against them (large campaign contributions) the EU went ahead and fined them over a billion dollars - the largest corporate fine in history at the time.
The Commission can bring proceedings against the studios in the European court, the judgement can be enforced in any EU member state.
Slashdot Mirror
Unfortunately CORBA had too much of it, most advances in Comp Sci are taking stuff out, not adding stuff in. Java is better than C++ mainly because it took out some really kludgy stuff. C is derrived from B, a simplified version of BCPL which is a simplified version of CPL.
CORBA had several major problems that everyone told the CORBA people about but nobody ever did anything about. Like basing the whole messaging scheme on RPC - subordination is a massively inefficient way to do parallel processing.
The biggest problem with CORBA is you had to buy this object broker and then pretty much rewrite your application to use it. With dotNET the idea is much more like the Web, leave the information servers be and write a thin layer interface to them so you can get to the data from wherever you need to.
I see CORBA as being the SGML to dotNet's XML. It may well be that you can do everything you can do in XML in SGML and some more. But SGML has some real deep seated structural limitations that keep biting you when you try to use it for real stuff. Writting the HTML DTD was pretty miserable.
Microsoft settled the case and would probably argued that it won the points important to it. The big outcome of the case was that the constract between Sun and MSFT was terminated. MSFT are not prohibited from developing a clean room version of Java, e.g. the stuff they bought in from HP and others.
However Microsoft is unlikely to want to lend its support to Java again since if they make extensions that Sun disapprove of Sun can still go and bully third party tools developers that copy them. Sun by its own insistence has the lead in the Java world, so Microsoft has gone to play in its own sandpit.
The big problem with Java in my opinion is that its main selling point - processor independence was something that only Sun really cared about. It has never been a big thing having to recompile code for different processors and most people use Intel in any case. In place of DLL hell I have runtime hell in which I have to make sure I download the right 20gazibabyte run time from Sun to run the program.
I think most people are missing out on the biggest likely impact of C# which in my view is on the Visual Basic side. VB has for years been the worlds most widely used programming language even though everybody agrees that it is something of a mess. Thing is however it has picked up some nice features along the way - particularly insensitivity to early or late binding.
I see C# as being more the union of the VB feature set with the C feature set than a direct competitor to Java. Being able to convert legacy VB applications into a more maintainable base sounds to me like a good thing.
Comparing implementation of the same code in different languages is one thing, measuring their run time is irrelevant unless you also test every compiler out there.
The only point at which computing language design has a big impact on performance is in the ease of optimization. More modern languages tend to be designed to make it easier for the compiler.
If you are using a scripting language and compare compiled code in C against interpreted code.. well duuuhhh! That is not a language issue, in fact one of the party tricks of certain Web servers is to compile server side scripts inline and cache the object code...
Equally comparison of functional languages with declarative is pretty meaningless. If you are doing matrix multiplication tasks and need good performance above all then maybe LISP is not the best choice.
The other issue that comes up is that on simple tasks there can be a massive difference in the ease of coding. Matrix operations can typically be done in one line in APL. But when you get to larger tasks that are beyond the core of any language the differences tend to be much less significant.
Kodak seem to be upset only because Microsoft got in the way of their attempt to tie their online prints service to their camera. At least Microsoft gave companies that do prints but not cameras a way to get in on the action.
The idea of charging for browser slots is not new. Netscape made almost all their money from the browser that way. AOL is simply one long infomercial for AOL's "partners".
Personally I get real peeved when companies mess with my file associations, particularly when they don't ask me. If there is a way to stop such meddling in XP I am all for it.
Bzzzt! Untrue. The HTTP specification explicitly allows error codes to be replaced.
If the error codes had to be reported as generated it would prevent browsers implementing error codes in the users own language. If I go to a french server I don't want to have to use Babel fish to read the error response.
If the intention had been to simply display the error message from the server then we would not have bothered with error codes at all. The purpose of the error codes were to allow the client software to do something more useful to the end user.
You as the server writer might not like the error code handling of the client, but that is explicitly outside your remit. The client controls the user interface, not the server.
Other parts of dotNET are already being open sourced. SOAP for example is being cloned all over the place. It is pretty clear that someone will have an Apache version.
The part of dotNET that would be hard to copy (and probably not that worthwhile) is the rack of API goodies that ship with windows. But that is not what a clone of dotNET would be useful for.
The main reason I would want a version of dotNET running on another platform would be because the whole big idea is network computing. If you have a legacy system running on a mainframe you don't need to move it onto a different platform just to use it as a Web service. Equally you might want to extend some embedded system and make use of a Web service on another box that is being set up.
Microsoft is trying to sell Windows as a big feature set development platform for building dotNET services. To succeed they need dotNET services to be available on other 'legacy platforms'.
Microsofts complaints about the GPL concern the viral aspect.
Please save us the libertarian hysterics. It is not as simple as you say. Contract law is not anything like as simple as you appear to believe, otherwise contract law lawyers would be much cheaper to hire.
The enforceability of contract terms is subject at all times to state and federal law. Don't like that liberweenie? well tough. Without the government courts to enforce the contract in the first place there is no contract.
The argument that it is all down in black and white does not move me. I do not believe that in this instance that the car rental company was honestly representing the contract terms. Nor does the dept of consumer safety. Bad faith has been an issue in contract law since the Romans invented the concept.
In the case in question the car company wrote the contract, they are thus on the hook in the case of any ambiguity. In this case there is good reason to doubt that the customer intended to agree to the specific term that he would be fined if the speed of the car went over a certain amount. The contract term in bold can be quite resonably be interpreted as meaning that there is a $150 surcharge if the police issue a speeding ticket.
Above and beyond the contract issues the idea is simply bad business. It is not uncommon to find US companies that believe that dishonest and underhand business techniques are the way to make profit. However even P.T.Barnum later observed that he made very little money when he was dishonest, he did much better when he put on a show that was worth the entry fee.
As a capitalist corporation Netpliance was Fully entitled to remove as much cash from your account as they felt like, as a corporation they are by definition in the right. Look at it this way how many times did you go over the speed limit after you bought the machine, under the Acme rentals argument that would entitle netpliance to fine you $150 per time...
Uhh ohh, the ray gun appears to be wearing off, I think I will go and have a nice little rest and read The Fountainhead or Atlas Shrugged a couple of times.
My strong belief is that there is a connection speed tipping point. The Web was miserable over 9.6Kb modems, barely acceptable at 14.4K and not too unpleasant at 28.8. Now I love my cable modem and 1Mb/sec but I would not wait arround while someone built it if I could get 28.8 somewhere else. So if there isn't enough demand to support Ricochet on its own what chance is there that the demand for G3 will be big enough to support five carriers in the UK or the rest of Europe?
Admittedly Ricochet is less usefull because it is not ubiquitous yet. But I don't think that is the reason there is a problem. CPDP didn't sell too well either. I think that the market for mobile data has been massively overestimated. Or at least the takeup rate has.
Perhaps better handheld devices will help. At the moment the PalmVII is a crippleware Web interface tied to a very slow network (the 9.6K claimed speed is total horse shite to use the technical term, try 0.2K if you are lucky.) The PocketPC series are much better hardwarewise but getting wireless service is pretty difficult and expensive. Plus PocketPCs drain the life out of a Lithium ion battery fast enough without wireless on top.
This sort of thing really p****es me off. Its like them goddam environmentalist tree huggers who want to stop the oil companies drilling for oil on the national mall, who cares if there isn't any oil there? Its a free country isn't it? And in any case if the national mall was that good they wouldn't have built a shopping center there now would they?
Now that the press is reporting that the decision has been reversed the legal niceties of that reversal matter little. The Bush people will settle and Microsoft will give them a nice campaign contribution before the soft money limits kick in.
Rather less of the decision has stood than some Slashdotters appear to think. The case has been remanded back to be heard by a different judge and the Appeals Court has made it clear that the new judge need not take much notice of Jackson's opinions.
The Appeals court has overturned parts of the 'findings of fact' in practically every area. I suspect this is part of a strategy. Under the ridiculous Appeals court rules the idea is that the Appeals court is bound by findings of fact by the lower court - hence Jackson's stated intention to prevent review by issuing the findings of fact as a separate document. If the Appeals Court had vacated the findings of fact directly there would be a possible appeal to the SC. However the Appeals court has essentially remanded the case in such a manner that the new judge is free to reverse as much of Jackson's findings as they see fit.
Specifically missing from the rulling is any endorsement of Jackson's findings, they decline to vacate them, but they don't bind the new judge to them.
I doubt that the case will get as far as a new trial since the Bush people are practically certain to settle. I suspect that the Gore folk would have settled also, although on rather less favorable terms. The terms of the proposed breakup were idiotic, they would have created two monopolies in place of one. I think they had more to do with the size of the lawyers egos than policy sense
Unfortunately the MIS managers tended to be as interested in recovering control over the corporate nervous system as control over costs. Back in the 1980s I made money by writing apps to screen scrape data out of corporate mainframes so the middle managers could feed it into Lotus 1-2-3. The MIS managers could have provided the data but they refused point blank to do so. It was all a power-game.
Besides, you DON'T need an 800MHz PIII, 128MB RAM and a 20Gb hard disk on EVERY PC in your network
But when that setup costs $500 from a computer store the premise that the Network computer would save money by reducing the hardware costs collapsed.
The Network computer failed for many reasons, not least the fact that there was no significant application support for the device. The lack of a hard disk was always going to be a drag on performance though. The idea of eliminating local user storage is not a bad one, end users should probably only have access to a centrally managed file store. The idea of no local long term storage was a complete loser however, as anyone who has used a diskless workstation knows.
It is not suprising that network computers were easy to manage. They could not be used for anything useful. What is more Sun and Oracle appeared to be spending more on idiotic TV ads than on writing aps to make them useful. If Sun had bought Star Office earlier they might have had a point.
Ultimately I see the whole Sun vs. Microsoft thing as denial on the part of Sun. Microsoft is not going to kill Sun, but Intel and Linux will. A high end linux box can be bought for a tenth the cost of the equivalent Sun box.
Also, C# is not an evolution of C++. C# is java under a different name, what microsft thinks java should have been.
Not actually true, If Sun had not behaved as it did then C# might have evolved out of Java. C# is actually built on top of the C++ compiler back end by the C++ team. The feature set is essentially a combination of C, COM and Visual Basic. It is possible to translate from Java into C# but it is also possible to translate from Visual Basic to C# so that hardly makes the languages equivalent. Sun is not the first company ever to suggest a new programming language, nor was Java successful for any new features, even the byte code interpreter has been arround on micros for 25 years (UCSD p system).
Sun's behavior with Java was pretty inexcusable. The idea that we should all be forced to agree to the decisions of one company when it comes to the feature set of a programming language is stupid. Sun deliberately crippled the native language interface for their own purposes. But many of the users of J++ simply wanted a better language than Visual Basic to write one off applications and did not give a monkeys about 'running everywhere'.
Contrary to the discussions on this board .NET has nothing much in common with Sun's failled Network Computer. The Network Computer was nothing more than a new name for an X-Terminal that can run java apps and a browser locally. That Sun tried it as a bet-the-business strategy was the best corporate joke since the Sinclair C5.
The Network computer concept was for desktop apps. It was an MIS manager's wet dream - take away the employees PCs and give them dumb terminals wired to the nice shiny mainframe. Back to the 1960s.
Network Computers failed for the very good reason that MIS managers looking to torture their victims could do the same job much more cheaply with Citrix without buying Scott McNealy and Larry Ellison more fuel for their corporate jets.
.NET is about information service access. It is the way to hook together e-commerce applications. It has zero to do with Suns clueless hardware platform.
The only intersection between .NET and Sun's effort is that Microsoft has rolled out the JIT compiler technology and Java alternative as part of the package. Neither is core to the .NET idea, or for that matter Sun. C# is merely a logical cleanup of C++, there are some points of comparison to Java but all the ideas have been arround long before Sun used them. If Sun hadn't got all proprietary closed and legal Microsoft might not have created their own, but nobody can really blame them for not being beholden to a standard Sun police ownership of with lawyers.
Hailstorm is only one small part of dotNET, getting all wound up about it is to miss the plot entirely. Gates is looking to take Oracle and SAP to the cleaners, Hailstorm is the smokescreen for that agenda.
As for the GPL bashing being to divert attention from anything, the covert agenda there is more likely to ram home to the analysts the fact that Linux is putting Sun and Solaris out of business and the poor performance of Sun the past few quarters is probably reflecting that dynamic rather than an overall slowdown.
In the EU all trucks must by law have a tachometer fitted that makes a continuous record of the speed of the truck. It is used to monitor the drivers hours and stop them from driving when they are too tired. Given that the accidents caused by heavy goods vehicles can kill ten or twenty people at a time this is a pretty good idea.
Anything that discourages HGVs from tailgating has to be a good idea. It would be even better if SUVs also had to have the same gear fitted, after all if they get the benefit of the light trucks emissions limits they should be taxed like trucks and have the same speed limiters.
Clearly YANAL, neither am I but I have spoken at enough ABA meetings to know something about what I am talking about.
Most consumer contracts contain unenforceable clauses. Every cell phone contract in the US contains a clause that waives the customer's right to being a class action. The clause is almost certainly unenforceable, as indeed is any clause that disallows redress for negligence.
In my business we spend a lot of time considering what contract clauses a court is likely to enforce against a customer - in particular a consumer. It is not remotely as simple as being able to hold the consumer to the writing on the page.
The other 'contract' probably isn't. The mere fact that a consumer scribbles their name on a document does not automatically create an enforceable contract. The list of exceptions is as long as your arm.
In this specific instance it does not appear that ACME made any attempt to bring the unusual contract clause to the customer's attention. That coupled with the fact that enforcement of traffic policy is a matter of public policy probably puts the rental agency in a less than optimum position.
If the notice was less than clear the mere fact that the company is trying to contest the case tends to indicate bad faith, otherwise a more reasonable response would be to put more prominent notices on future contracts.
Good advice, postgrad rhetoricians minoring in linguistics are well known to be no good at all at any kind of sex and even Madonna stated after the attempt that she would rather have spent the two minutes and five seconds in question doing something more fun like filing her taxes instead.
Both systems failed for the same basic reason, there simply was not enough metal to back the amount of money required by a modern economy. If you think about it the idea that the optimum amount of money in circulation should be tied to the amount of a shiny metal that has been taken out of the ground is rather odd.
The US federal government still has ownership of something like 50-60% of the total world gold reserves. Most of that is the payment on war loans made by the British following world war I and II which in turn was the booty of Empire. The total quantity of gold bullion has at most doubled since WWII, in the same time the GDP of the US has in real terms expanded at least ten fold.
There simply is not enough yellow metal to go round. Nixon abandoned the gold standard for the simple fact that even under the system of managed exchange rates there was simply not enough yellow metal to support the economic activity. The supply of gold had become the limiting factor for the economy.
The idea that money should be backed by real value has emotional appeal to many. Back at the turn of the century the Deomcratic party was essentially captured by a monomaniac called Willian Bryans Jennings whose sole speech was 'that man should not be crucified on a cross of gold' - monetary reform by moving to a bimetalic standard.
It is not surprising that people trying to invent their own currencies should attempt to base them on gold. But there is a big difference between having a gold ingot in the hand and having an account with a fly by night operator in St Mcru (pop 5 penguins).
It is now 30 odd years since the US was on the gold standard and the number of people expecting a return is rapidly diminishing. At the same time most other countries have abandoned the gold standard and nobody wants to have a managed exchange rate (although some are forced to). I suspect that the diminishing gold price reflects the fact that gold is loosing its traditional role as a safe haven in troubled times.
I would not be surprised by the first Tempest machine fetching a high price, Tempest was a first generation machine with some pretty exotic hardware.
I would quite happily pay $25K for an Apple I. But I would not pay more than $50 for an Apple II even if it was an ultra rare model.
Of course some idiot with a fat wallet may read about the auction on Slahdot and bid the box up, but given the vintage the machine is already way over bid. If the instant buy is $25K the guy probablky thinks he will get $10K at least.
There are better models but the ones you have just listed are the ones that Comp Sci moved away from twenty odd years ago finding them complex and likely to introduce bugs.
The principal abstraction in threads is Hoare's monitors (there is some dispute over who invented monitors, Hoare credits Dijkstra, Dijkstra Brinch-Hansen and Brinch-Hansen Hoare). A couple of years after the monitors paper Hoare developed the Communicating Sequential Processes model (CSP) which was later used as the basis for the occam programming language - which remains pretty much the only mainstream programming language with decent support for parallelism.
Unfortunately the threads programming model does not provide decent abstractions for communicating between threads, except for shared memory, which is severely limiting in the multiprocessor context, there are very good reasons why shared memory bus SMP machines have consistenly had a maximum of 16 processors with 4 being the more common limit.
Incidentally, the Amdahls 'law' that kicks off the piece was originaly marketting propaganda to persuade folk that faster processors (like Amdahl made) could not be replaced vector processing add on boxes. The argument is valid, but the framing of the argument is deceptive. Very few problems have the type of fine grained parallelism that can be exploited by vector boxes, however most engineering problems have parallelism at coarser granularities.
People can get into trouble with threads, but they can get themselves into much more trouble with half baked multi-process tweaks. Five years ago the quality of threads implementations generaly was so poor that multiprocess hacks were the only way to go. Today that does not apply.
I don't think that the effectiveness can be measured 'exactly'. It is possible to detect occasions on which an advert has lead directly to a sale. But why should that happen very often anyway? A lot of the time advertising is a cumulative process, you see the ad, register the message and then refer back when you actually need (or want) the product.
I don't doubt that Nike get every penny back that they pay Tiger Woods for hawking their sweatshop produced clothing. However I don't think anyone sees Tiger Woods in his swoosh cap and runs out to the mall to buy one like it.
I suspect that a bigger part of the problem is that the Web changes the nature of advertising from paid placement to attracting trafic to the advertiser's web site. If you open up any hobbyist magazine you will see page after page of advertisements that are simply lists of products and prices. Nobody would place that kind of an ad in Salon, they would have a link to their own site and host the catalog there. The other big advertising money spinner is classified ads and EBay has cornered that market.
Untrue on both counts. The term 'New Media' was coined to get people consulting gigs. The people who actually invented the Internet and the Web knew what they were doing all along. The people who had zero clue were the analysts and journalists who spent their time interviewing each other and bilging out puff pieces about 'internet time'. Nine years on and the Web is still a work in progress, so what was that 'internet time' they were talking about?
Most people talking about new media were talking about two things, interactivity and a different cost structure to print or TV. Interactivity is what brings people to Slashdot. Newspapers have always had letters pages, but online forums take the concept much further.
The difference in the cost structure online vs print is dramatic. If you don't have to pay for the content, publishing becomes close to free. If you are a government you can probably save money by putting documents online rather than printing them.
Where people's expectations failed was when they fooled themselves into thinking that new media would lead to new media empires. I don't believe that was ever going to happen and if it did what does it benefit anyone if an old media conglomerate like Time-Warner is replaced by a new media conglomerate like AOL?
We always thought that online new media would be small scale mom 'n pop type stuff with a few medium sized outfits (which it is mainly, look at the prOn sites). When the new media companies started to employ staffs of 100+ the writing was on the wall.
The Web will still change the media world, it will just take fifteen or twenty years rather than the six months the 'Internet time' cretins blathered about.
Changing peoples way of life takes time, live with it.
Oh really, tell that to IBM. After the Reagan administration dropped the anti-trust case against them (large campaign contributions) the EU went ahead and fined them over a billion dollars - the largest corporate fine in history at the time.
The Commission can bring proceedings against the studios in the European court, the judgement can be enforced in any EU member state.