My employer completed migration of apps to Linux - Ubuntu

This is a discussion on My employer completed migration of apps to Linux - Ubuntu ; "The Natural Philosopher" wrote in message news:1226092348.22001.1@proxy02.news.clara.net... > dennis@home wrote: >> >> >> "Maxwell Lol" wrote in message >> news:878wrw8ijf.fsf@com.invalid... >>> "dennis@home" writes: >>> >>>> I don't need to see it. >>>> If it does what he states it will ...

+ Reply to Thread
Page 12 of 13 FirstFirst ... 2 10 11 12 13 LastLast
Results 221 to 240 of 250

Thread: My employer completed migration of apps to Linux

  1. Re: My employer completed migration of apps to Linux



    "The Natural Philosopher" wrote in message
    news:1226092348.22001.1@proxy02.news.clara.net...
    > dennis@home wrote:
    >>
    >>
    >> "Maxwell Lol" wrote in message
    >> news:878wrw8ijf.fsf@com.invalid...
    >>> "dennis@home" writes:
    >>>
    >>>> I don't need to see it.
    >>>> If it does what he states it will make the data order somewhat mixed.
    >>>> If he wants it debugged then he will have to post the code and agree
    >>>> the fees.
    >>>
    >>> If he's doing select(), then he can handle several input and output
    >>> files simultaneously.

    >>
    >> He could but that isn't what he said.
    >> He said he could make his application simulate AIO by using several
    >> threads to do whatever the application was doing before. A sure way to
    >> screw up the order.

    >
    > Oh dear. You had better tell Zeus that. Thats what they do and they only
    > manage to be a few orders of magnitude faster and less memory intensive
    > than Apache. Now I know why whenever I went to te server for page A, I got
    > page b instead. How come no one else has ever noticed it?


    Stop being a pita and trying to fit an application into what he said.
    That application doesn't matter what order they are done as individual
    threads are only processing one connection.
    Its when several threads are processing one connection which is what he said
    would emulate AIO.





  2. Re: My employer completed migration of apps to Linux



    "The Natural Philosopher" wrote in message
    news:1226092187.22001.0@proxy02.news.clara.net...

    > Remember that a web server is essentially connectionless, even though it
    > uses TCP/IP. each socket is closed at the transfer of a page.


    The browser opens and closes the connection and most modern browsers will
    keep the connection open knowing that it is a big overhead to open and close
    connections like they used to do in the dark ages.

    > So the web server and the underlying network layer do NOT maintain
    > thousands of sockets for long periods: the number of open sockets is as
    > many as the number of transfers going on.


    Define long, they certainly maintain them for longer than the page transfer
    time.

    If you really wanted to make a fast web server then getting rid of the user-
    kernel space separation would be the best start.
    I have done applications running as STREAMS modules for this very purpose
    and it really boosts the performance.


  3. Re: My employer completed migration of apps to Linux

    After takin' a swig o' grog, Erik Funkenbusch belched out
    this bit o' wisdom:

    > On Fri, 7 Nov 2008 07:34:17 -0500, Chris Ahlstrom wrote:
    >
    >> After takin' a swig o' grog, Erik Funkenbusch belched out
    >> this bit o' wisdom:
    >>
    >>>> One of the best references on this whole general topic is this:
    >>>>
    >>>>
    >>>
    >>> Interesting quote from that article:
    >>>
    >>> "glibc 2.1 and later provide a generic implementation written for standards
    >>> compliance rather than performance."
    >>>
    >>> So even Linux doesn't have performance in mind when it wraps AIO with
    >>> standard functions.

    >>
    >> Another ignorant person who conflates GNU and Linux.

    >
    > Gee, doesn't every call it GNU/Linux?


    Note the separator character between the two. Needed for proper parsing.

    >> In any case, you don't need to use glibc's functions. Try libaio.

    >
    > Then you're not using "portable" code, like Ignoramus insists.


    Sometimes there's a trade-off between efficiency and portability.

    Just ask Microsoft:

    --
    "Portability is for canoes."
    -- Jim McCarthy, former Visual C++ team leader, in "Dynamics of Software
    Development", Microsoft Press

  4. Re: My employer completed migration of apps to Linux

    After takin' a swig o' grog, Larry Page belched out
    this bit o' wisdom:

    > "dennis@home" writes:
    >
    >> "The Natural Philosopher" wrote in message
    >> news:1226092187.22001.0@proxy02.news.clara.net...
    >>
    >>> Remember that a web server is essentially connectionless, even
    >>> though it uses TCP/IP. each socket is closed at the transfer of a
    >>> page.

    >>
    >> The browser opens and closes the connection and most modern browsers
    >> will keep the connection open knowing that it is a big overhead to
    >> open and close connections like they used to do in the dark ages.

    >
    > No they don't. If they did most servers would saturate in no time at
    > all.
    >
    >>
    >>> So the web server and the underlying network layer do NOT maintain
    >>> thousands of sockets for long periods: the number of open sockets is
    >>> as many as the number of transfers going on.

    >>
    >> Define long, they certainly maintain them for longer than the page
    >> transfer time.

    >
    > They most certainly do not.


    Don't confuse MD5 Dennis (aka "dumbass@home") with facts.

    --
    Pryor's Observation:
    How long you live has nothing to do
    with how long you are going to be dead.

  5. Re: My employer completed migration of apps to Linux



    "Larry Page" wrote in message
    news:gf2ft7$vti$3@registered.motzarella.org...
    > "dennis@home" writes:
    >
    >> "The Natural Philosopher" wrote in message
    >> news:1226092187.22001.0@proxy02.news.clara.net...
    >>
    >>> Remember that a web server is essentially connectionless, even
    >>> though it uses TCP/IP. each socket is closed at the transfer of a
    >>> page.

    >>
    >> The browser opens and closes the connection and most modern browsers
    >> will keep the connection open knowing that it is a big overhead to
    >> open and close connections like they used to do in the dark ages.

    >
    > No they don't. If they did most servers would saturate in no time at
    > all.
    >
    >>
    >>> So the web server and the underlying network layer do NOT maintain
    >>> thousands of sockets for long periods: the number of open sockets is
    >>> as many as the number of transfers going on.

    >>
    >> Define long, they certainly maintain them for longer than the page
    >> transfer time.

    >
    > They most certainly do not.


    Open a browser, load a page and then do a netstat.
    Tell me how long it is before the browser connections are open before they
    close then.


  6. Re: My employer completed migration of apps to Linux

    "dennis@home" writes:

    > "Larry Page" wrote in message
    > news:gf2ft7$vti$3@registered.motzarella.org...
    >> "dennis@home" writes:
    >>
    >>> "The Natural Philosopher" wrote in message
    >>> news:1226092187.22001.0@proxy02.news.clara.net...
    >>>
    >>>> Remember that a web server is essentially connectionless, even
    >>>> though it uses TCP/IP. each socket is closed at the transfer of a
    >>>> page.
    >>>
    >>> The browser opens and closes the connection and most modern browsers
    >>> will keep the connection open knowing that it is a big overhead to
    >>> open and close connections like they used to do in the dark ages.

    >>
    >> No they don't. If they did most servers would saturate in no time at
    >> all.
    >>
    >>>
    >>>> So the web server and the underlying network layer do NOT maintain
    >>>> thousands of sockets for long periods: the number of open sockets is
    >>>> as many as the number of transfers going on.
    >>>
    >>> Define long, they certainly maintain them for longer than the page
    >>> transfer time.

    >>
    >> They most certainly do not.

    >
    > Open a browser, load a page and then do a netstat.
    > Tell me how long it is before the browser connections are open before
    > they close then.


    http is a stateless and connectionless protocol.

    State is created using cookies and similar.

    https is a different kettle of fish.

    Most websites are http based. Most do not leave connections open.

    Sorry, but you are wrong I think.



  7. Re: My employer completed migration of apps to Linux



    "Larry Page" wrote in message
    news:gf2hng$lqp$1@registered.motzarella.org...
    > "dennis@home" writes:
    >
    >> "Larry Page" wrote in message
    >> news:gf2ft7$vti$3@registered.motzarella.org...
    >>> "dennis@home" writes:
    >>>
    >>>> "The Natural Philosopher" wrote in message
    >>>> news:1226092187.22001.0@proxy02.news.clara.net...
    >>>>
    >>>>> Remember that a web server is essentially connectionless, even
    >>>>> though it uses TCP/IP. each socket is closed at the transfer of a
    >>>>> page.
    >>>>
    >>>> The browser opens and closes the connection and most modern browsers
    >>>> will keep the connection open knowing that it is a big overhead to
    >>>> open and close connections like they used to do in the dark ages.
    >>>
    >>> No they don't. If they did most servers would saturate in no time at
    >>> all.
    >>>
    >>>>
    >>>>> So the web server and the underlying network layer do NOT maintain
    >>>>> thousands of sockets for long periods: the number of open sockets is
    >>>>> as many as the number of transfers going on.
    >>>>
    >>>> Define long, they certainly maintain them for longer than the page
    >>>> transfer time.
    >>>
    >>> They most certainly do not.

    >>
    >> Open a browser, load a page and then do a netstat.
    >> Tell me how long it is before the browser connections are open before
    >> they close then.

    >
    > http is a stateless and connectionless protocol.
    >
    > State is created using cookies and similar.
    >
    > https is a different kettle of fish.
    >
    > Most websites are http based. Most do not leave connections open.
    >
    > Sorry, but you are wrong I think.
    >
    >


    I think you are just thinking in human time spans while I am thinking in
    machine time spans.
    They don't stay open for minutes at a time but they do stay open a lot
    longer than a transfer requires.
    They are open for about 9 seconds longer than they need to be on my machine.



  8. Re: My employer completed migration of apps to Linux

    "dennis@home" writes:

    > "Larry Page" wrote in message
    > news:gf2hng$lqp$1@registered.motzarella.org...
    >> "dennis@home" writes:
    >>
    >>> "Larry Page" wrote in message
    >>> news:gf2ft7$vti$3@registered.motzarella.org...
    >>>> "dennis@home" writes:
    >>>>
    >>>>> "The Natural Philosopher" wrote in message
    >>>>> news:1226092187.22001.0@proxy02.news.clara.net...
    >>>>>
    >>>>>> Remember that a web server is essentially connectionless, even
    >>>>>> though it uses TCP/IP. each socket is closed at the transfer of a
    >>>>>> page.
    >>>>>
    >>>>> The browser opens and closes the connection and most modern browsers
    >>>>> will keep the connection open knowing that it is a big overhead to
    >>>>> open and close connections like they used to do in the dark ages.
    >>>>
    >>>> No they don't. If they did most servers would saturate in no time at
    >>>> all.
    >>>>
    >>>>>
    >>>>>> So the web server and the underlying network layer do NOT maintain
    >>>>>> thousands of sockets for long periods: the number of open sockets is
    >>>>>> as many as the number of transfers going on.
    >>>>>
    >>>>> Define long, they certainly maintain them for longer than the page
    >>>>> transfer time.
    >>>>
    >>>> They most certainly do not.
    >>>
    >>> Open a browser, load a page and then do a netstat.
    >>> Tell me how long it is before the browser connections are open before
    >>> they close then.

    >>
    >> http is a stateless and connectionless protocol.
    >>
    >> State is created using cookies and similar.
    >>
    >> https is a different kettle of fish.
    >>
    >> Most websites are http based. Most do not leave connections open.
    >>
    >> Sorry, but you are wrong I think.
    >>
    >>

    >
    > I think you are just thinking in human time spans while I am thinking
    > in machine time spans.
    > They don't stay open for minutes at a time but they do stay open a lot
    > longer than a transfer requires.


    No they don't. The server closes the connection after each
    transfer. If it didn't then the server of a busy web site would saturate
    in no time at all.

    Anyone? This has always been my understanding.

    > They are open for about 9 seconds longer than they need to be on my
    > machine.


    That might be so.

    But define "they" and "need to be".


  9. Re: My employer completed migration of apps to Linux



    "Larry Page" wrote in message
    news:gf2io9$uf4$1@registered.motzarella.org...
    > "dennis@home" writes:
    >
    >> "Larry Page" wrote in message
    >> news:gf2hng$lqp$1@registered.motzarella.org...
    >>> "dennis@home" writes:
    >>>
    >>>> "Larry Page" wrote in message
    >>>> news:gf2ft7$vti$3@registered.motzarella.org...
    >>>>> "dennis@home" writes:
    >>>>>
    >>>>>> "The Natural Philosopher" wrote in message
    >>>>>> news:1226092187.22001.0@proxy02.news.clara.net...
    >>>>>>
    >>>>>>> Remember that a web server is essentially connectionless, even
    >>>>>>> though it uses TCP/IP. each socket is closed at the transfer of a
    >>>>>>> page.
    >>>>>>
    >>>>>> The browser opens and closes the connection and most modern browsers
    >>>>>> will keep the connection open knowing that it is a big overhead to
    >>>>>> open and close connections like they used to do in the dark ages.
    >>>>>
    >>>>> No they don't. If they did most servers would saturate in no time at
    >>>>> all.
    >>>>>
    >>>>>>
    >>>>>>> So the web server and the underlying network layer do NOT maintain
    >>>>>>> thousands of sockets for long periods: the number of open sockets is
    >>>>>>> as many as the number of transfers going on.
    >>>>>>
    >>>>>> Define long, they certainly maintain them for longer than the page
    >>>>>> transfer time.
    >>>>>
    >>>>> They most certainly do not.
    >>>>
    >>>> Open a browser, load a page and then do a netstat.
    >>>> Tell me how long it is before the browser connections are open before
    >>>> they close then.
    >>>
    >>> http is a stateless and connectionless protocol.
    >>>
    >>> State is created using cookies and similar.
    >>>
    >>> https is a different kettle of fish.
    >>>
    >>> Most websites are http based. Most do not leave connections open.
    >>>
    >>> Sorry, but you are wrong I think.
    >>>
    >>>

    >>
    >> I think you are just thinking in human time spans while I am thinking
    >> in machine time spans.
    >> They don't stay open for minutes at a time but they do stay open a lot
    >> longer than a transfer requires.

    >
    > No they don't. The server closes the connection after each
    > transfer. If it didn't then the server of a busy web site would saturate
    > in no time at all.
    >
    > Anyone? This has always been my understanding.
    >
    >> They are open for about 9 seconds longer than they need to be on my
    >> machine.

    >
    > That might be so.
    >
    > But define "they" and "need to be".


    Well a page like news.bbc.co.uk loads in about 1 second.

    If I get around to it I will put an analyzer on the network and see what
    packets are sent.
    It won't be today though as I have other things to do.
    You can do it if you want.
    >


  10. Re: My employer completed migration of apps to Linux

    On Fri, 07 Nov 2008 15:46:14 +0000, Christopher Hunter
    wrote:

    >Bruce in Bangkok wrote:
    >
    >> My point is that there are many, many more of these people than there
    >> are knowledgable people and it is these unskilled people who make up the
    >> bulk of the computer buyers. And they are keeping MS happy.

    >
    >Funny you should say that. I've just been in Japan, and the biggest selling
    >machines right now are the Linux sub-notebooks. Their "average" computer
    >user isn't at all interested in which OS they use, as long as it works
    >properly. Both Linux and XP versions of these little computers are
    >available, but the XP ones just gather dust - they cost almost twice as much
    >as the Linux ones, and their performance is appalling.
    >
    >The "average" Japanese computer user is finding Linux easy enough to use.
    >Perhaps they're just more intelligent than the average American...
    >
    >I just received a one month old Dell 1525 with Vista on it from a friend. A
    >quick look through its hard drive (it won't boot any more) showed that it was
    >riddled with innumerable pieces of malware of all sorts, despite having
    >endless anti-this and anti-that software. As far as its owner was concerned,
    >it was broken beyond repair. I recovered all their data and burned it to a
    >couple of CDs so they shouldn't lose anything they'd saved. Installation of
    >Ubuntu took just over 15 minutes and everything "Just Worked", but the owner
    >of the machine had gone to a local computer shop and bought another one with
    >Vista on it... Some people never learn!
    >
    >C.


    You are just full of these odd little snippet's of information. I talk
    about the world wide market and you select relatively small corner of
    the world and I suspect faulty data.

    But read what the industry press had to say about these little
    computers. It certainly doesn't sound like Asus is betting the
    business on Linux..

    From INFORMATION WEEK - 14 MAR 2008
    http://www.informationweek.com/blog/..._eee_gets.html

    At the CeBIT convention in Germany the other week, Asus unveiled a new
    edition of its flash-based Eee sub-notebook PC, nominally running a
    custom version of Xandros Linux. New models, Asus said, will run
    Linux ... and now Windows XP.

    According to Asus' press release, Microsoft (NSDQ: MSFT) Works and
    many of the Windows Live applications will come preloaded with the
    system. It claims it has sold 350,000 units so far and is on track to
    selling anywhere between 3 and 5 million by the end of the year.

    So if they're selling so well as it is, why Windows? From the sound
    of it, customer demand. People want it, plain and simple, even if it
    means the base cost of the unit goes up by a certain amount -- and I'm
    betting Asus, like most other PC OEMs, has worked out some kind of
    deal to get a price break on XP in bulk.

    Windows is still king of the roost as far as desktop OSes go. It's
    what people know and are comfortable with -- even if most of their
    existing documents can be opened and used on Linux, there are still
    many people who simply prefer to stick with what they know, even if it
    means (and in some cases especially if it means) dealing with familiar
    hassles.

    I've seen parallels to this sort of thing in my own experience. One
    friend was suffering from the usual trauma of a spyware-infested
    computer. I mentioned a couple of Linux distributions, and his
    response was on the order of "It sounds nice, but I don't want to have
    to learn everything all over again." It's a case of the devil you
    know vs. the devil you don't know, and if anything went wrong at least
    he could always throw in his restore disc and start over.

    So how will the presence of XP affect Asus' plans to ship Linux
    editions of its product? So far, nothing's changed; you can still get
    the Linux version side-by-side. But if Asus finds it has more of a
    market with XP on its machine than Linux -- and there's little to
    suggest they won't! -- then I fear it will see no reason to continue
    offering something that doesn't sell as aggressively as it hoped.
    Bruce-in-Bangkok
    (correct Address is bpaige125atgmaildotcom)

  11. Re: My employer completed migration of apps to Linux

    On Fri, 07 Nov 2008 16:03:08 +0000, Christopher Hunter
    wrote:

    >Bruce in Bangkok wrote:
    >
    >> I'm not several years behind times although I may be several weeks
    >> behind times.

    >
    >You are if you think Linux is more difficult to install than Windoze!
    >
    >> I recently installed OpenSuse 11.0 on a computer. At the end on the
    >> installation everything worked except for the printer and a USB Wi-Fi
    >> adapter I use when on the boat. So... I contacted Canon, "got a Linux
    >> driver for your printer?" Nope, have got this source code thingy that
    >> might work." Downloaded thingy. "make thingy". computer says "no have
    >> compiler and other dependencies".

    >
    >Your USB wireless-thingy /should/ be easily resolved, and (I don't mean to be
    >rude but) only a fool would buy /any/ Canon product. All Canon stuff is
    >overpriced junk, and even /they/ are beginning to recognise that they're
    >losing sales to manufacturers who /do/ provide Linux drivers (like HP and
    >Epson).
    >
    >> Installed copy of Windows XP. At the end of the installation
    >> everything worked except for the printer and the wi-fi adapter. Went
    >> to the printer box, stuck Canon supplied CD in drive, drive starts and
    >> loads drivers. Went to Wi-Fi box and stuck vendor supplied CD in
    >> drive, drive starts and loads drivers. Everything works.

    >
    >You had *no* applications installed at this point, and first connection to
    >the 'net (to get the innumerable Windoze and anti-virus updates) would
    >guarantee infection with the latest malware nasty, rendering the installation
    >effectively useless...
    >
    >> Now don't bother to give me any BS about "well, you ought to know how
    >> to fix it" as I do know how to fix it, but just answer me honestly -
    >> which installation went the smoothest. Which was the easiest?

    >
    >No comparison. If you had a proper supported printer, you'd have had no
    >problem, /and/ a comprehensive suite of useful programmes if you installed
    >Linux.


    That is absolutely the weakest argument I have ever heard. In fact it
    reminds me of the old mother watching the soldiers march by. Oh look
    she cried, "Every one is out of step, except my Johnnie".

    You are a real piece of work aren't you?

    I think that this weak-kneed retort shall end the discussion. After
    all as Friedrich von Schiller said: With stupidity the gods
    themselves contend in vain.

    Bruce-in-Bangkok
    (correct Address is bpaige125atgmaildotcom)

  12. Re: My employer completed migration of apps to Linux

    "John F. Morse" wrote:
    >

    .... snip ...
    >
    > I didn't go to China, but I saw the BSOD on the Bird Nest Stadium
    > from my living room chair in Kansas.
    >
    > So did the rest of the world I would guess.


    I think that just indicates that RCA hasn't woken up yet.

    --
    [mail]: Chuck F (cbfalconer at maineline dot net)
    [page]:
    Try the download section.

  13. Re: My employer completed migration of apps to Linux

    On 2008-11-07, dennis@home wrote:
    >
    >
    > "Joe" wrote in message
    > news:slrngh89ig.ps0.joe@barada.griffincs.local...
    >
    >> That is not "Installing". That is running a recovery image. Dell
    >> ships these for Ubuntu, as well.
    >>
    >> For a ground-up installation, though, in most cases, Ubuntu is much
    >> quicker and easier to install that Windows. A full Windows install
    >> includes over an hour of waiting for the initial install, then waiting
    >> while hardware is detected, swapping disks several times (should you
    >> have all of them) to get the drivers installed, hitting the internet
    >> to add the current drivers for the devices you don't have the disks
    >> for, another chunk of time to run Windows update, then the install of
    >> your antivirus and anti-spyware software. Now you are at a point
    >> where you can undertake the tedious task of installing all of the apps
    >> you use, one at a time, with 20 user dialog boxes during the install.

    >
    > That isn't how to install windows.
    > You boot vista, select install, answer four or five questions, wait 35 mins.
    > *if* it hasn't recognised you network card, install the drivers.
    > Let it run windows update to get the remaining drivers.
    > Then add AV (avg or avast).


    Bull****. I installed Vista Ultimate on a PC at my parent's house.
    Intel CoreDuo, 2 GB of RAM. The inital install took over an hour, and
    Windows update did not have the video drivers (nvidia), the system
    board drivers, or the Sound Card drivers. Needed to download each
    one, as the disks we had were for XP.

    >
    >>
    >> Now, after several hours, you have a working box. Then you get to put
    >> it on the internet to find out that your virus defs from this morning
    >> do not cover that new virus you just got, and you get to mess with
    >> that.

    >
    > Odd that soem of use have never had a virus even though we are on the
    > Internet all the time.
    > Some users must be far more susceptible to virus infections than others,
    > just as they are likely to answer phishing emails.


    Older folks are very succeptable to viruses. All they use the PC for
    is e-mail and web browsing, and the e-mails that they share with each
    other are almost always the crap that's been forwarded 30 times and
    has all sorts of fun attachments. Considering what I do for a living,
    like most IT people, I end up supporting the PC usage of everyone in
    my family.

    >
    >>
    >> The biggest problem in an Ubuntu install is getting some newer
    >> hardware to work due to lack of sufficient drivers. This has gotten
    >> much better, but it is still an issue, especially with a new laptop.
    >> It can be frustrating, but in most cases a few minutes of research
    >> will yield positive results, and you're on your way.

    >
    > Yes, but it takes knowledge of what you actually have to find the answers,
    > lots of people don't know what a NIC is or if they have an nvidia 8400M
    > graphics chip.
    > They don't need to with windows, they do with Ubuntu.


    They don't need to know which graphics chip with Linux, either. You
    MAY need to know the specific card model, though, with either OS. You
    go to the nvidia or ATI web site, and you browse by model.

    For Linux, though, there is EnvyNG, which does everything for you. It
    detects the chipset, downloads and installs the correct driver, then
    recommends a reboot because of the changes to the kernel image.
    Pretty simple...

    >
    >>
    >> My desktop came with Vista installed. It took 30 minutes, from
    >> beginning to end, to remedy that. Same with the kids' desktop. My
    >> laptop took a little longer, because the Atheros 5006 Wireless
    >> adapter wasn't supported by madwifi at the time, and I had to hunt
    >> down XP drivers to get it to work under ndiswrapper. Total time:
    >> about 45 minutes.

    >
    > Try that with someone that doesn't know what Atheros is.
    >


    They don't deserve to own the laptop then, since it is on a large
    sticker on the lower left of the keyboard wrist rest.

    And while it came pre-loaded with Windows, the Atheros driver is NOT
    automatically installed by a stock Windows install, either. You
    either need the CD or a downloaded driver.

    >>
    >>
    >>>
    >>>>
    >>>>> Windows, on the other hand is, on the surface, a much more simple
    >>>>> system
    >>>>> for
    >>>>> the average user to comprehend, and that is important.
    >>>>
    >>>> Entirely wrong! Most users *never* install an operating system. A
    >>>> fairly
    >>>> substantial proportion of them will even /buy/ /a/ /new/ /computer/ when
    >>>> Windows fouls up (as it invariably does), because they assume that the
    >>>> computer is "faulty". You wouldn't believe the number of "nearly new"
    >>>> machines that can be bought for (effectively) nothing because of this.
    >>>>
    >>>
    >>> Well when you are selling to the masses there are bound to be some who
    >>> can't
    >>> use windows, what makes you think linux would fair any better in those
    >>> hands?
    >>> At least with windows you can take it into a shop and they will restore
    >>> it
    >>> for about £30.
    >>> Or even free if you know which shop or know one of the other 99% of
    >>> windows
    >>> users that understand how to do it.
    >>>

    >>
    >> I am not one of those Linux purists. Windows has done plenty of
    >> things OK. The UI on XP was decent, and XP was fairly reliable, so
    >> long as you kept it clean of viruses. Vista, on the other hand, is a
    >> pain in the ass. I have it on one machine, and I've gotten to the
    >> point that I have it auto log on and run Windows Media Center at boot.
    >> My only interaction with it is through a remote control, except for
    >> the annoyance of needing to reboot it every couple of days when Media
    >> Center locks and won't start back up right without a boot.

    >
    > So what's wrong with it then?


    Not a damned thing is wrong with it. Running Mint, it ran forever and
    a day. Installed Vista, and it started acting up. Likely a
    problematic device driver, as most lockups are...

    > My vista MCE machine has been running for months and never locks up.
    > I can trust it to record TV without any bother.
    > If yours is so unreliable there is something wrong with it and it needs
    > fixing.


    Sure does. It needs MythBuntu instead of Windows.

    > Mine is an ancient Shuttle with a Hauppauge pci and a Hauppauge USB tuner
    > providing the two channels.
    > Vista is running in 1G RAM with two 250g drives and a 8G stick to provide
    > readyboost, which reduces paging.


    Proud of you. Mine is newer than that, more HD space, but no
    readyboost.
    >
    >>
    >> As soon as I build a new machine for the Entertainment system, with
    >> hardware fully compatible with MythBuntu, Vista will be eliminated
    >> from my house.

    >
    > Why not build new hardware to fix your vista machine?


    There's nothing wrong with the hardware that's in there now. Like I
    said, probably drivers, but Vista sure doesn't make things any better.
    Hell, if I had a copy of XP MCE sitting around, I'd wipe it and switch
    to that.

    > And which newbie is going to install and get MythTV to work?
    > Remember most windows users are newbies.


    I don't care about them when it comes to MY machine...

    Also, distribution-based Myth isn't that hard to set up and configure.
    It may take someone that is inexperienced a bit more time, but it
    isn't that rough. If you take the defaults in the MythBuntu setup,
    and you have fairly standard equipment, it just goes.

    >
    >>
    >> And I've yet to go to any business with more than 10 users that has
    >> ANY copies of Vista running. They are all sticking with XP.

    >
    > That is sensible.
    > They already have systems in place and there is no point in replacing them.
    > M$ know this and allow people to use XP on the Vista license as they allowed
    > 200 on their XP license and 3.1 on 95 in the past.
    >


    You are missing the point, though. When they talk about how many
    copies of Vista they have sold, they count all of these machines that
    have NEVER run Vista.

    We are replacing all of our lower-end PIII's and P4's slowly, but
    surely. at least a few hundred machines per month. They all have an
    XP image put on them, but each of the machines has a Vista image.
    That's where the bogus sales figures come from.

    Vista is a piece of junk.


    --
    Joe - Linux User #449481/Ubuntu User #19733
    joe at hits - buffalo dot com
    "Hate is baggage, life is too short to go around pissed off all the
    time..." - Danny, American History X

  14. Re: My employer completed migration of apps to Linux

    On Sat, 08 Nov 2008 00:07:10 +0100, Larry Page wrote:

    > http is a stateless and connectionless protocol.


    Half right. It's stateless, but not connectionless.

    > Most websites are http based. Most do not leave connections open.


    HTTP 1.1 provides for "keep-alives"

    > Sorry, but you are wrong I think.


    He's not.

  15. Re: My employer completed migration of apps to Linux

    On Sat, 08 Nov 2008 00:24:40 +0100, Larry Page wrote:

    > No they don't. The server closes the connection after each
    > transfer. If it didn't then the server of a busy web site would saturate
    > in no time at all.
    >
    > Anyone? This has always been my understanding.


    http://www.io.com/~maus/HttpKeepAlive.html

  16. Re: My employer completed migration of apps to Linux

    After takin' a swig o' grog, Erik Funkenbusch belched out
    this bit o' wisdom:

    > On Sat, 08 Nov 2008 00:07:10 +0100, Larry Page wrote:
    >
    >> http is a stateless and connectionless protocol.

    >
    > Half right. It's stateless, but not connectionless.
    >
    >> Most websites are http based. Most do not leave connections open.

    >
    > HTTP 1.1 provides for "keep-alives"
    >
    >> Sorry, but you are wrong I think.

    >
    > He's not.


    Sure he's wrong. HTTP connections are normally transitory. He claimed they
    were all permanent.

    --
    You will stop at nothing to reach your objective, but only because your
    brakes are defective.

  17. Re: My employer completed migration of apps to Linux

    After takin' a swig o' grog, Erik Funkenbusch belched out
    this bit o' wisdom:

    > On Sat, 08 Nov 2008 00:24:40 +0100, Larry Page wrote:
    >
    >> No they don't. The server closes the connection after each
    >> transfer. If it didn't then the server of a busy web site would saturate
    >> in no time at all.
    >>
    >> Anyone? This has always been my understanding.

    >
    > http://www.io.com/~maus/HttpKeepAlive.html


    Why don't you just run netstat for yourself?

    On Linux, try

    $ netstat -c --inet

    and watch it while you access various web sites.

    --
    Even more amazing was the realization that God has Internet access. I
    wonder if He has a full newsfeed?
    -- Matt Welsh

  18. Re: My employer completed migration of apps to Linux

    On 2008-11-08, Chris Ahlstrom wrote:
    > After takin' a swig o' grog, Erik Funkenbusch belched out
    > this bit o' wisdom:
    >
    >> On Sat, 08 Nov 2008 00:07:10 +0100, Larry Page wrote:
    >>
    >>> http is a stateless and connectionless protocol.

    >>
    >> Half right. It's stateless, but not connectionless.
    >>
    >>> Most websites are http based. Most do not leave connections open.

    >>
    >> HTTP 1.1 provides for "keep-alives"
    >>
    >>> Sorry, but you are wrong I think.

    >>
    >> He's not.

    >
    > Sure he's wrong. HTTP connections are normally transitory. He claimed they
    > were all permanent.
    >


    They could be semi-permanent, though it does not lead to great
    efficiencies.
    --
    Due to extreme spam originating from Google Groups, and their inattention
    to spammers, I and many others block all articles originating
    from Google Groups. If you want your postings to be seen by
    more readers you will need to find a different means of
    posting on Usenet.
    http://improve-usenet.org/

  19. Re: My employer completed migration of apps to Linux

    dennis@home wrote:
    >
    >
    > "The Natural Philosopher" wrote in message
    > news:1226092187.22001.0@proxy02.news.clara.net...
    >
    >> Remember that a web server is essentially connectionless, even though
    >> it uses TCP/IP. each socket is closed at the transfer of a page.

    >
    > The browser opens and closes the connection and most modern browsers
    > will keep the connection open knowing that it is a big overhead to open
    > and close connections like they used to do in the dark ages.
    >
    >> So the web server and the underlying network layer do NOT maintain
    >> thousands of sockets for long periods: the number of open sockets is
    >> as many as the number of transfers going on.

    >
    > Define long, they certainly maintain them for longer than the page
    > transfer time.
    >


    They don't. Unless the browser at the other end is badly configured and
    aborts the transaction. Once its recieved its page it shuts the connection.


    > If you really wanted to make a fast web server then getting rid of the
    > user- kernel space separation would be the best start.
    > I have done applications running as STREAMS modules for this very
    > purpose and it really boosts the performance.


    No, I would buy one that already has done the work,. Its called Zeus.

  20. Re: My employer completed migration of apps to Linux

    dennis@home wrote:
    >
    >
    > "Larry Page" wrote in message
    > news:gf2hng$lqp$1@registered.motzarella.org...
    >> "dennis@home" writes:
    >>
    >>> "Larry Page" wrote in message
    >>> news:gf2ft7$vti$3@registered.motzarella.org...
    >>>> "dennis@home" writes:
    >>>>
    >>>>> "The Natural Philosopher" wrote in message
    >>>>> news:1226092187.22001.0@proxy02.news.clara.net...
    >>>>>
    >>>>>> Remember that a web server is essentially connectionless, even
    >>>>>> though it uses TCP/IP. each socket is closed at the transfer of a
    >>>>>> page.
    >>>>>
    >>>>> The browser opens and closes the connection and most modern browsers
    >>>>> will keep the connection open knowing that it is a big overhead to
    >>>>> open and close connections like they used to do in the dark ages.
    >>>>
    >>>> No they don't. If they did most servers would saturate in no time at
    >>>> all.
    >>>>
    >>>>>
    >>>>>> So the web server and the underlying network layer do NOT maintain
    >>>>>> thousands of sockets for long periods: the number of open sockets is
    >>>>>> as many as the number of transfers going on.
    >>>>>
    >>>>> Define long, they certainly maintain them for longer than the page
    >>>>> transfer time.
    >>>>
    >>>> They most certainly do not.
    >>>
    >>> Open a browser, load a page and then do a netstat.
    >>> Tell me how long it is before the browser connections are open before
    >>> they close then.

    >>
    >> http is a stateless and connectionless protocol.
    >>
    >> State is created using cookies and similar.
    >>
    >> https is a different kettle of fish.
    >>
    >> Most websites are http based. Most do not leave connections open.
    >>
    >> Sorry, but you are wrong I think.
    >>
    >>

    >
    > I think you are just thinking in human time spans while I am thinking in
    > machine time spans.
    > They don't stay open for minutes at a time but they do stay open a lot
    > longer than a transfer requires.
    > They are open for about 9 seconds longer than they need to be on my
    > machine.
    >
    >

    That more likely a function of the way netstat interacts with the kernel.

+ Reply to Thread
Page 12 of 13 FirstFirst ... 2 10 11 12 13 LastLast