Slashdot Mirror


User: jeremyp

jeremyp's activity in the archive.

Stories
0
Comments
2,700
First seen
Last seen
Profile
(view on slashdot.org)

Comments · 2,700

  1. Re:Genetic Algorithms, Rat Bags and Cheetahs. on Breeding Race Cars With Genetic Algorithms · · Score: 1
    Of course, we already know that this stuff works; as a quick trip to the zoo will show you what evolution has done to optimize the cheetah.
    After several million years, the best that nature has come up with can do about 70 mph for short periods. Human rally cars can sustain higher speeds for longer over the same terrain.

    There are many cases were evolution has led to sub-optimal "designs" (the connection of the retina to the optic nerve in the human eye being one that springs to mind).

  2. Re:None, really... on iTunes 4.6, DRM, and Hymn · · Score: 1

    Apple has another motive. The iTunes store doesn't make much money for them. It's a loss leader to generate revenue from iPod sales. If it was easy to alter the songs without loss of quality such that they could be played on other cheaper MP3 players, fewer people would be buying iPods.

  3. Re:Am I the only one... on The Mythical Man-Month Revisited · · Score: 2, Interesting

    Why would it be obvious that adding more resource to a task makes it slower? There are plenty of people about (some who are in my company, some who have read MMM!) who think that works even for software dev.

    The answer is yes: read it. It's a classic of the IT World and contains some important ideas (as well as being an interesting view of the IT World 30 years ago).

  4. Re:Background article on The Technology Behind Formula One · · Score: 2, Informative

    The real context here:

    Michael Andretti competed in 13 races. He retired from the first four races following three collisions and a spin. The collisions may or may not have been his fault. His other three retirements were another spin, another collision and a throttle failure. So that's one mechanical failure and at least two driver errors.

    In that year Mclaren had five wins and nine podiums and finished second in the Constructors championship. That doesn't suggest uncompetitive to me except in respect of Williams who were miles ahead of everybody that year and the next three.

  5. Re:The language does matter on Java Faster Than C++? · · Score: 1

    This is not garbage collecting. Garbage collecting is automatically freeing memory when it is no longer in use. The virtual memory system has no way of telling when a process has finished with its memory until the process exits.

    When you malloc some memory and then use it, the OS has to allocate some physical RAM. If you then don't use it for a bit, the OS can say "OK he's not using this memory any more, I'll get rid of it" but it can't throw the memory away so it saves it in the swap space. If you use enough virtual memory, eventually both the physical RAM and swap space wil become full. At that point your server keels over in a heap (no pun intended).

  6. Re:Sorry, no. on Java Faster Than C++? · · Score: 1

    "Thus "clean" code in c++ will be copy-intensive"

    In C++ use a reference to a constant e.g.

    int foo (const MyClass& bar)
    { ...
    }

  7. Re:Matrices on Java Faster Than C++? · · Score: 1

    Point of information:

    The computer only has a one dimensional array of memory locations to work with. If in C you define a 2D array e.g.

    int foo [5][5] ;

    it will allocate a single dimensional array and do the exact same calculations you would do.

  8. Re:BBC on Interesting Tech-Related Online Talk Radio? · · Score: 1

    When they set the site up, Realplayer was apparently the most widely used media player. They wanted to get to the widest audience.

  9. Re:Biased on Searching for the Best Scripting Language · · Score: 1

    Any language that depends on white space as part of its syntactic structure is evil (except as a delimiter for tokens). Languages that use the line ending as a statement separator are just about acceptable although they encourage people to write stupidly long lines of code.

    If I can't trust a programmer to indent properly on a voluntary basis, how can I trust them to always check their return codes etc?

  10. Re:The Office of the Information Commissioner on UK Anti-Spam Laws Criticised · · Score: 3, Informative
    The UK government uses the electoral register to sell your data (regardless of whether you "opt out") to third party marketing companies to get revenue.

    This is a lie. It's not even maintained at the national level, but at the district council level. By law, anybody can see a copy of the register (under supervised conditions) and certain companies are allowed to use the register for restricted purposes. For example, credit agencies can use it to verify your address. It is illegal for them to pass the data on to anyone else or use it for any but a set of restricted purposes.

    The editted version of the register may be bought by anybody, but you can opt out of that. See here for more details.

  11. Re:And Who Pays For This? on 19th Century News Coming Online · · Score: 1

    What law? Give a reference to a substantiated news story. Also note that it is not necessary to have a TV licence to listen to the radio.

    Personnally I do not begrudge a single penny of my licence fee. When I look at the BBC and at the alternatives, I pray that the current state of affairs continues for as long as possible.

  12. Re:far more important than optimization on Why Learning Assembly Language Is Still Good · · Score: 1

    But if your game comes out three months earlier, you have the advantage.

  13. Re:This is a Good Thing... on RFID License Plates in the UK · · Score: 1

    I don't think his tractor was actually in London. People fake number plates to avoid the congestion charge thus landing some poor innocent person with the bill.

  14. Re:A bit different view on Marking 50 Years Since Alan Turing's Death · · Score: 3, Insightful

    At the time, homosexuality was illegal in the United Kingdom. That would have made Turing a target for blackmail in the eyes of the security organisations. In terms of the moral and legal climate of the time, they probably felt themselves to be totally justified in what they did.

    One of the reasons why he was treated so badly by the legal organisations (i.e. those that arrested and condemned him for breaking what is now considered an abhorrent law) is that they didn't know what he had done for his country during the war. It was still classified then.

    The tradgedy not that Turing brought this on himslef, but that people didn't know any better then.

  15. Re:Excuse me, but ... on Send A Message To An LED Sign · · Score: 1

    Yes. Is it News? Does it Matter? Woo hoo, I've got a web server where you can type stuff in and have it displayed. That's never been done before.

    My theory is the editors thought he was an idiot and decided to teach him a lesson. Unfortunately, they have probably toasted the internet connections of everybody else on the same bit of cable. Can you say "contention"?

  16. Re:That's 2 words. on Making Operating Systems Faster · · Score: 1

    And me, except for dev work.

    Let's test my new killer app. Oh two bombs. That's the last three hours down the tubes.

    Also I used to play "Hack" from a floppy.

    Down to level 21. RAM disk full. Two bombs. That's the last three hours down the tubes.

  17. Re:/.'ed already. on The Mathematics of Futurama · · Score: 1

    Schroedinger's Cat was a thought experiment designed to illustrate the philosophical problems of the Copenhagen interpretation of quantume mechanics i.e. the act of observation collapses the wave function. What is an observer etc.

    The Uncertainty Principle states that the product of the uncertainty in position and momentum can never be smaller than some value (h/2pi I think). It's important to note that this does not just mean that we cannot measure a particle's position and velocity exactly, but that it doesn't "know" its position and velocity exactly.

  18. Re:Who's the retard? on First IA64 Windows Virus Released · · Score: 1

    Viruses looks and sounds fucking stupid.

    Been thinking about that. I've never heard anybody use the word "virii" in spoken language. It's a matter of personal preference, but I actually find it harder to get my mouth around v-eye-r-eye than viruses and v-eye-r-eye doesn't sound nice to my ear.


    The other good reason for sticking with "viruses" is that the English language is already screwed up with irregular plurals and verbs and spelling etc etc. I see no reason for deliberately making it worse.

  19. Re:Windows on HPC? on In The Works: Windows For Supercomputers · · Score: 3, Insightful

    You make one or two good points, but there's also a lot of FUD here. OK, let's look at this rationally:

    Microsoft OS's core was designed to run on a single user system with that user operating with administrative privilages.Networking was straped on top of that. Security was added after that. Copmatibility with older MS systems and API's were such a priority, that the things needed to make the system secure can't be added.

    FUD. I doubt if they are suggesting using Win98. Win2K and XP work acceptably in a multi-user way. The users have admin privileges issue is more to do with the culture where people expect to be able to do anything they want without logging in as a different user. The security model (with proper ACLs) is much better than that of Unix.

    The issues you see with Microsoft's products tend to be with things like IIS and Internet Explorer. Do you say Linux is insecure because of a sendmail, apache or mozilla exploit?

    Every time Microsoft tightens security, it seems 10 more holes are found.

    More FUD. Are you saying that every M$ security patch causes 10 new faults? Do you have any real evidence for that other than your own prejudices.

    Support, how large of team does Microsoft has with experience running clusted computer systems based on a Microsoft OS? Do you want to pay for them to come up to speed?

    M$ has quite a reasonable clustering product already. Admittedly it's a high availability clustering solution rather than a supercomputer solution, but there will always be a few suckers to help them with the learning curve on the super computing side. Just wait long enough not to be one of them (if you are determined to go with them).

    CPU license requirements and all the BS needed to keep them current.

    A good point. M$ clustering is not going to be cheap.

    Vendor lock in
    Software designed to take advantage of a clustered environment. Where is all the experience at, writing it for *nix or Windows?

    Lots of software has been ported to Windows from *nix before. It's only a small hurdle but with the pricing issue might be a show stopper.

    The system registry, a bad idea no matter how many processors you run.

    There are many Unix features that are a bad idea no matter how many processors you run.

    Needing to run a cluster with AntiVirus software

    Yep, agree with this point. That's more licences to buy.

    People that know how to operate clusterd environments have *nix expereince and are used to using *nix tools, not looking for the registry tweeks to fix system problems.

    This is a good point too up to the FUD about registry tweaks.

  20. Re:Spam And Viruses on University Capitulates, Switches Off Spam Filters · · Score: 2, Insightful

    Slightly pedantic: the bounce message will go to the address in the forged reply-to header, or from header, or envelope sender.

    I would question quietly deleting such mails. Most of the worm/virus ridden mails that I get come from people who have infected systems and where I am in their address book. They need to know they have an infected system.

  21. Re:Spam And Viruses on University Capitulates, Switches Off Spam Filters · · Score: 1

    I've read the RFC. Tell me what features of it make it unreliable. If all e-mail servers were correctly implemented and correctly configured, the only way to lose mail would be a hardware failure.

  22. Re:An argument for distributed version control on Security Holes in CVS and Subversion Found · · Score: 1
    These vulnerabilities are a consequence of an architectural security flaw in both CVS and Subversion: they require an active server that talks a complex protocol to an unauthenticated client.

    Actually, you are not required to run anonymous access to either Subversion or CVS.

    Anyway, allowing anonymous access to untrusted clients is a feature of FTP, SMTP, HTTP, SMB and many other protocols. It's always a weak point which must be programmed for. The logical extreme is not to allow any access whatsoever - then it might be secure.

  23. Re:Which? on The Physics of Baseball · · Score: 2, Funny

    The link to howstuffworks for seeing was particularly useful for me. I never realised what the round slippery things in the front of my head were for.

  24. Re:HFS+ on Measuring Fragmentation in HFS+ · · Score: 1

    XP on my PC and laptop don't seem to. Or the doiks get fragmented in spite of it according to the defrag tool.

  25. Re:Who? What? on SBC CWA Strike Imminent · · Score: 1

    SBC?

    CWA?

    WTF is that all about?

    Not import IMHO, but IANAA