1st place prize: Ubuntu 8.04 Long Term Slop - Linux

This is a discussion on 1st place prize: Ubuntu 8.04 Long Term Slop - Linux ; Hadron wrote: > Kelsey Bjarnason writes: > >> [snips] >> >> On Mon, 19 May 2008 10:35:37 +0800, Ian Hilliard wrote: >> >>> What is necessary is a full risk analysis and then add the expected risk >>> costs into ...

+ Reply to Thread
Page 5 of 5 FirstFirst ... 3 4 5
Results 81 to 97 of 97

Thread: 1st place prize: Ubuntu 8.04 Long Term Slop

  1. Re: 1st place prize: Ubuntu 8.04 Long Term Slop

    Hadron wrote:

    > Kelsey Bjarnason writes:
    >
    >> [snips]
    >>
    >> On Mon, 19 May 2008 10:35:37 +0800, Ian Hilliard wrote:
    >>
    >>> What is necessary is a full risk analysis and then add the expected risk
    >>> costs into the final estimate. There may also be changes made to the
    >>> project, which can reduce the risk. There is however the need to have a
    >>> "Standard Risk" entry, which amounts to 20% of the total estimate before
    >>> adding any other risk values.

    >>
    >> On large-scale projects, this is feasible. On smaller ones, it often
    >> isn't - some risk analysis, sure, but one can only go so far within a
    >> given budget.
    >>
    >> Still, any way you slice it, you're being a blowhard - by suggesting
    >> things be done properly.
    >>

    >
    > Poor big head Kelsey still does not "get it". No one is suggesting
    > things are not done "properly". people are however suggesting that Mr
    > Hilliard is making it sound bigger and more difficult than it really
    > it. We all know to schedule for "unknowns" to a degree but no one that I
    > am aware of has ever managed to do that "efficiently" on any large
    > project. You can add fudge factors, you can reschedule but SW is SW and
    > things can go quicker or slower than planned. And the main crux of the
    > matter is simply that NO ONE delivers a finished product and THEN
    > decides on which platform it will run.
    >


    You still don't understand, do you? The point is that the development should
    be generic. If it then turns out that the preferred platform is not
    suitable for one reason or another, then it is possible to change to a
    platform that is suitable. This adds an extra degree of flexability and
    reduces risk.

    Ian

  2. Re: 1st place prize: Ubuntu 8.04 Long Term Slop


    "Ian Hilliard" wrote in message
    news:1211462917.632894@angel.amnet.net.au...
    > Hadron wrote:
    >
    >> Kelsey Bjarnason writes:
    >>
    >>> [snips]
    >>>
    >>> On Mon, 19 May 2008 10:35:37 +0800, Ian Hilliard wrote:
    >>>
    >>>> What is necessary is a full risk analysis and then add the expected
    >>>> risk
    >>>> costs into the final estimate. There may also be changes made to the
    >>>> project, which can reduce the risk. There is however the need to have a
    >>>> "Standard Risk" entry, which amounts to 20% of the total estimate
    >>>> before
    >>>> adding any other risk values.
    >>>
    >>> On large-scale projects, this is feasible. On smaller ones, it often
    >>> isn't - some risk analysis, sure, but one can only go so far within a
    >>> given budget.
    >>>
    >>> Still, any way you slice it, you're being a blowhard - by suggesting
    >>> things be done properly.
    >>>

    >>
    >> Poor big head Kelsey still does not "get it". No one is suggesting
    >> things are not done "properly". people are however suggesting that Mr
    >> Hilliard is making it sound bigger and more difficult than it really
    >> it. We all know to schedule for "unknowns" to a degree but no one that I
    >> am aware of has ever managed to do that "efficiently" on any large
    >> project. You can add fudge factors, you can reschedule but SW is SW and
    >> things can go quicker or slower than planned. And the main crux of the
    >> matter is simply that NO ONE delivers a finished product and THEN
    >> decides on which platform it will run.
    >>

    >
    > You still don't understand, do you? The point is that the development
    > should
    > be generic. If it then turns out that the preferred platform is not
    > suitable for one reason or another, then it is possible to change to a
    > platform that is suitable. This adds an extra degree of flexability and
    > reduces risk.


    Actually NO. Assuming that you're developing something that you intend on
    becoming successful then development should target the platform(s) that is
    most likely going to be used by your customers. It should not simply be
    "generic."

    Example: Let's say you plan on writing some sort of plug-in for AutoCad. It
    sure would make sense for your application to run on the same OS that
    AutoCad runs on. Not some generic OS where your app runs great but AutoCad
    isn't available for it.

    Or let's say that you're writing some data analysis package for Oracle
    databases. You would target the most popular platforms that run Oracle and
    not something like OSX.

    And if you're developing some really cool 3D game controller then it makes
    zero sense to develop this controller+software for Solaris because it's not
    a gaming platform. Even if the software did run best on Solaris, it's not
    the platform that you develop for.


    People typically think of "marketing" as being the people that create
    advertisements, commercials and marketing campaigns. This happens
    after-the-fact. The more important aspect of marketing is to identify the
    market for your product and to estimate your potential success in that
    market. It's the job of marketing to identify who your customers are going
    to be, what software+OS is currently used by those potential customers and
    then develop the product based on that.

    It is completely IDIOTIC to develop the software for a bunch of platforms
    first and then afterwards decide that hey... it works best on XYZ so that's
    what we're going to release it on.




    > Ian



    ** Posted from http://www.teranews.com **

  3. Re: 1st place prize: Ubuntu 8.04 Long Term Slop

    Ezekiel wrote:

    > Actually NO. Assuming that you're developing something that you intend on
    > becoming successful then development should target the platform(s) that is
    > most likely going to be used by your customers. It should not simply be
    > "generic."
    >


    Actually, this is the nonsense. Java has proven popular because it permits
    multi-platform. There are also a number of libraries which permit
    multi-platform development. Just look at http://www.boost.org/ or
    http://www.wxwidgets.org/

    Sun and IBM are pushing the idea of generic development. At the moment, the
    only company that is not pushing generic development is Microsoft. This is
    simply because Microsoft has the most to lose if generic code is developed.

    The advantages of generic development are many. The most important in the
    development phase is the wider range of debugging tools that are available.
    Generic code also permits a wider range of customers. This is exactly what
    Microsoft does not want. It wants developers to develop code that can only
    be used on Windows, because it forces people to use Windows, whether they
    want to or not.

    Windows has too much of the market at the moment to be totally ignored, but
    wise developers realize that developing Windows only code is also limiting.
    It is not that hard to develop generic code which can be run on multiple
    platforms. It simply requires clean design and self control. After all, all
    development should require clean design and self control.

    Ian

  4. Re: 1st place prize: Ubuntu 8.04 Long Term Slop


    "Ian Hilliard" wrote in message
    news:1211468654.505492@angel.amnet.net.au...
    > Ezekiel wrote:
    >
    >> Actually NO. Assuming that you're developing something that you intend on
    >> becoming successful then development should target the platform(s) that
    >> is
    >> most likely going to be used by your customers. It should not simply be
    >> "generic."
    >>

    >
    > Actually, this is the nonsense. Java has proven popular because it permits
    > multi-platform.


    Java is mostly popular in the application server market. Can you name a
    couple of "popular applications" that are written in Java? And by popular I
    mean something along the lines of Photoshop, AutoCad, TurboTax and software
    like that.


    > There are also a number of libraries which permit
    > multi-platform development. Just look at http://www.boost.org/ or
    > http://www.wxwidgets.org/


    Cross platform libraries have been around for along time. RogueWave is
    something that comes to mind. The fact is that these cross-platform
    libraries never really caught on. One main reason is that you end up with an
    app that can run anywhere but doesn't look or feel native on any platform.
    So even this "boost" or "wxwidgets" stuff... nice but can you name some
    successful applications that were written using this? I sure can't.

    I think that at the end of the day Mac users want software that looks, feels
    and runs like Mac software. Not something that looks like it was written for
    Windows and then made cross-platform for the Mac. And the same for any other
    way/direction you want to slice it. And other than trivial/simple apps...
    "cross-platform" toolkits typically limit you to the lowest common
    denominator that's available for all platforms.

    Yes, I'm not a fan of cross platform toolkits.



    > Sun and IBM are pushing the idea of generic development. At the moment,
    > the
    > only company that is not pushing generic development is Microsoft. This is
    > simply because Microsoft has the most to lose if generic code is
    > developed.


    Of course they're pushing the idea of generic development. In theory it
    would be nice for everyone but the world doesn't work that way. It would be
    nice if there was a generic "oil filter" that you could buy that could be
    used with every car in the world. If something as simple as an oil filter
    can't be generic then what are the chances for software.


    > The advantages of generic development are many. The most important in the
    > development phase is the wider range of debugging tools that are
    > available.
    > Generic code also permits a wider range of customers. This is exactly what
    > Microsoft does not want. It wants developers to develop code that can only
    > be used on Windows, because it forces people to use Windows, whether they
    > want to or not.


    Are you forced to use Windows and by who?


    > Windows has too much of the market at the moment to be totally ignored,
    > but
    > wise developers realize that developing Windows only code is also
    > limiting.
    > It is not that hard to develop generic code which can be run on multiple
    > platforms. It simply requires clean design and self control. After all,
    > all
    > development should require clean design and self control.


    The product I work on runs on many platforms and Windows is probably one of
    the less popular platforms for us in terms of sales. (It's enterprise and
    not desktop software.) But we know in advance exactly what OS's we plan on
    supporting and we designed the product from day one to run cross platform.
    It's written entirely in C++ (with a little C and even less assembler) and
    the thought of using something like Java or a cross platform toolkit simply
    isn't realistic. There's no way we could get the performance and the fine
    grain level of control that we need if we used something like that.


    > Ian



    ** Posted from http://www.teranews.com **

  5. Re: 1st place prize: Ubuntu 8.04 Long Term Slop

    Ezekiel wrote some more nonsense:

    The Boost framework massively boosts productivity. It's downside is that it
    produces a "generic" interface. WxWidgits on the other hand is based on the
    MFC paradigm and produces GUIs that match the OS in use. Your lack of
    interest in cross platform libraries probably stems from a lack of
    knowledge about them.

    A great program, written in Java, which is quite popular is Magic Draw
    http://www.nomagic.com/. It may not have the wide usage of Photoshop, but
    this is simply because the market for software development tools is much
    smaller.

    Another great Java program is "freemind", which is a great freeware mind
    mapping tool. This is in wide spread use, but doesn't have to marketing
    budget to get it into the trade press every month.

    > The product I work on runs on many platforms and Windows is probably one
    > of the less popular platforms for us in terms of sales. (It's enterprise
    > and not desktop software.) But we know in advance exactly what OS's we
    > plan on supporting and we designed the product from day one to run cross
    > platform. It's written entirely in C++ (with a little C and even less
    > assembler) and the thought of using something like Java or a cross
    > platform toolkit simply isn't realistic. There's no way we could get the
    > performance and the fine grain level of control that we need if we used
    > something like that.
    >


    I'm sure that you could develop software with a Java front end and a generic
    C++ core, but I am also sure that you haven't tried. I know that it can be
    done, because I have seen it done. It works well.

    The development frameworks that now exist for Java desktop development have
    made it far more productive than any other development platform. Most
    people don't get to hear that because of the loud noises coming from
    Microsoft's marketing department.

    Ian

  6. Re: 1st place prize: Ubuntu 8.04 Long Term Slop

    Ian Hilliard wrote:

    > Windows has too much of the market at the moment to be totally
    > ignored, but wise developers realize that developing Windows only
    > code is also limiting.


    Limited to virtually 100% of the corporate desktops, and at least 90% of the
    home desktops. That's a real limited target platform, huh?

    That's why most, if not all, of your products are developed for Windows.



  7. Re: 1st place prize: Ubuntu 8.04 Long Term Slop

    DFS wrote:

    > Ian Hilliard wrote:
    >
    >> Windows has too much of the market at the moment to be totally
    >> ignored, but wise developers realize that developing Windows only
    >> code is also limiting.

    >
    > Limited to virtually 100% of the corporate desktops, and at least 90% of
    > the
    > home desktops. That's a real limited target platform, huh?
    >


    It may be news to you that the majority of the world's population DO NOT
    live in the United States. Outside of the USA alternatives to Windows are
    becoming more and more popular.

    > That's why most, if not all, of your products are developed for Windows.


    Those companies that are able to support those people using alternatives to
    Windows will most likely become the market leaders. Remember, "You don't
    become a leader by staying with the pack."

    Ian


  8. Re: 1st place prize: Ubuntu 8.04 Long Term Slop

    Ian Hilliard wrote:
    > DFS wrote:
    >
    >> Ian Hilliard wrote:
    >>
    >>> Windows has too much of the market at the moment to be totally
    >>> ignored, but wise developers realize that developing Windows only
    >>> code is also limiting.

    >>
    >> Limited to virtually 100% of the corporate desktops, and at least
    >> 90% of the
    >> home desktops. That's a real limited target platform, huh?
    >>

    >
    > It may be news to you that the majority of the world's population DO
    > NOT live in the United States. Outside of the USA alternatives to
    > Windows are becoming more and more popular.


    Why are you even pretending your customers aren't Windows users in North
    America and Europe? You really need to grow some honesty.

    So, do you pick your sales target market after developing the software, the
    same way you pick your OS after developing the code? You seem to have it
    all bass-ackwards.



    >> That's why most, if not all, of your products are developed for
    >> Windows.

    >
    > Those companies that are able to support those people using
    > alternatives to Windows will most likely become the market leaders.


    Yeah, that'll happen right about the time that MS open sources Windows and
    voluntarily ceases operations...



    > Remember, "You don't become a leader by staying with the pack."


    I'll bite: which development house whose products don't run on Windows is a
    leader in a market in which MS\Windows is also a contender? Oracle database
    on Unix, maybe, but Oracle was in business long before Windows took over.




  9. Re: 1st place prize: Ubuntu 8.04 Long Term Slop

    Ian Hilliard writes:

    > Ezekiel wrote some more nonsense:
    >
    > The Boost framework massively boosts productivity. It's downside is that it
    > produces a "generic" interface. WxWidgits on the other hand is based on the
    > MFC paradigm and produces GUIs that match the OS in use. Your lack of
    > interest in cross platform libraries probably stems from a lack of
    > knowledge about them.
    >
    > A great program, written in Java, which is quite popular is Magic Draw
    > http://www.nomagic.com/. It may not have the wide usage of Photoshop, but
    > this is simply because the market for software development tools is much
    > smaller.


    Amazing. Kelsey copycat Ian Hilliard uses a program most of us have
    never heard of to back up his ludicrous claims.

    I have never seen a successfull Java desktop application. They are slow
    and resource hungry. And for years the supposed "cross platform" UIs
    have created ugly garbage. And you know it. And it is well documented.


  10. Re: 1st place prize: Ubuntu 8.04 Long Term Slop

    On Fri, 23 May 2008 17:43:39 +0800, Ian Hilliard wrote:

    > DFS wrote:
    >
    >> Ian Hilliard wrote:
    >>
    >>> Windows has too much of the market at the moment to be totally ignored,
    >>> but wise developers realize that developing Windows only code is also
    >>> limiting.

    >>
    >> Limited to virtually 100% of the corporate desktops, and at least 90% of
    >> the
    >> home desktops. That's a real limited target platform, huh?
    >>
    >>

    > It may be news to you that the majority of the world's population DO NOT
    > live in the United States. Outside of the USA alternatives to Windows are
    > becoming more and more popular.


    Oh, dear, dear, dear! Doofu$ hasn't caught up with April 2008 LinuxFormat
    (LX104) "Linux in America" which says: Linux is being taken more seriously
    across a wider demographic.

    Companies like Office Depot, Warner Music Group, Comcast, AIG,
    Ticketmaster, McKesson, Metropolitan Bank Group (Chicago) are now using
    Linux. Education establishments, like San Diego United Schools District
    (130,000 students using Linux computers), Oakland University (30,000
    students using Linux computers). The US Army who manages the records of
    1.2 million GIs with RHEL.

    It seems DooFu$ is upset, because his world is shrinking & he won't adapt.

    >> That's why most, if not all, of your products are developed for Windows.

    >
    > Those companies that are able to support those people using alternatives
    > to Windows will most likely become the market leaders. Remember, "You
    > don't become a leader by staying with the pack."
    >
    > Ian


    --
    Mandriva 2008.1 64-bit.
    This message was sent from a
    computer which is guaranteed
    100% free of the M$ Windoze virus.

  11. Re: 1st place prize: Ubuntu 8.04 Long Term Slop

    William Poaster wrote:

    > Oh, dear, dear, dear! Doofu$ hasn't caught up with April 2008
    > LinuxFormat (LX104) "Linux in America" which says: Linux is being
    > taken more seriously across a wider demographic.


    Now if Linux could only get the locking/hanging/freezing problems fixed, and
    get a decent office suite, and some games, and a few thousand good apps, and
    gags on the loud mouths of the "advocates", and some buyers, and some good
    businessmen/marketers...

    You getting the picture yet, Dumb Willie?



    > Companies like Office Depot, Warner Music Group, Comcast, AIG,
    > Ticketmaster, McKesson, Metropolitan Bank Group (Chicago) are now
    > using Linux.


    A lot more - and bigger - companies than those are using Linux in one way or
    another. Does that make you proud?



    > Education establishments, like San Diego United Schools
    > District (130,000 students using Linux computers), Oakland University
    > (30,000 students using Linux computers). The US Army who manages the
    > records of 1.2 million GIs with RHEL.


    Funny! For years various cola idiots scoffed that the popularity of an OS
    says nothing about its quality, but all of a sudden cola bozos are starting
    to quote the biggest numbers they can find to support Linux.

    cola weenie "advocates" have mastered the art of talking from both sides of
    their mouth.



    > It seems DooFu$ is upset, because his world is shrinking & he won't
    > adapt.


    Windows has approx. 90% worldwide installed share. Virtually 100% of
    corporate desktops are Windows (this is my 'world'), and they aren't going
    anywhere.

    Who needs to adapt to your non-existent fantasy?




  12. Re: 1st place prize: Ubuntu 8.04 Long Term Slop


    "Ian Hilliard" wrote in message
    news:1211507707.650468@angel.amnet.net.au...
    > Ezekiel wrote some more nonsense:
    >
    > The Boost framework massively boosts productivity.


    I've used frameworks before and so have other people where I work. For
    various reasons these cross-platform "productivity boosting" frameworks
    always seem to promise more than they actually deliver and finding/fixing
    bugs in them is problematic. We have our own cross-platform framework that
    we've developed here which we can control and maintain. For us this is a
    better solution but other shops may feel differently.


    > It's downside is that it produces a "generic" interface.


    Yuck. This is essentially the 'kiss of death' for a commercial application.


    > WxWidgits on the other hand is based on the
    > MFC paradigm and produces GUIs that match the OS in use. Your lack of
    > interest in cross platform libraries probably stems from a lack of
    > knowledge about them.


    More likely is that I don't do any GUI work. Our product barely has a GUI.
    At a previous job when we tried to use a framework (Roguewave) for heavy
    duty computational work it was more trouble then it was worth. For us at
    least.



    > A great program, written in Java, which is quite popular is Magic Draw
    > http://www.nomagic.com/. It may not have the wide usage of Photoshop, but
    > this is simply because the market for software development tools is much
    > smaller.


    I've never heard of it but that's just me. It's probably a decent niche
    application but it's not what I would classify as being mainstream. If these
    cross-platform frameworks were as good as they claim then you'd think that
    somebody like Adobe or Microsoft of AutoCad or IBM or Oracle would be using
    them for application development.



    > Another great Java program is "freemind", which is a great freeware mind
    > mapping tool. This is in wide spread use, but doesn't have to marketing
    > budget to get it into the trade press every month.


    I've never heard of this one either. Java is great and popular for
    application servers. But Java never really made any significant inroads to
    desktop user applications. The "write once run anywhere" claim works well
    for backend processing but not so well for GUI's. Running a Java desktop
    application *feels* like you're running a Java app and not like a native
    app.


    >> The product I work on runs on many platforms and Windows is probably one
    >> of the less popular platforms for us in terms of sales. (It's enterprise
    >> and not desktop software.) But we know in advance exactly what OS's we
    >> plan on supporting and we designed the product from day one to run cross
    >> platform. It's written entirely in C++ (with a little C and even less
    >> assembler) and the thought of using something like Java or a cross
    >> platform toolkit simply isn't realistic. There's no way we could get the
    >> performance and the fine grain level of control that we need if we used
    >> something like that.
    >>

    >
    > I'm sure that you could develop software with a Java front end and a
    > generic
    > C++ core, but I am also sure that you haven't tried. I know that it can be
    > done, because I have seen it done. It works well.


    If by Java front-end you mean Java GUI then ehhhhhh. I've run these sorts of
    apps (the Adaptec Storage Manager front-end for example) and it's fine for
    little utilities and stuff like that. But if your app is heavily GUI based
    (by this I mean that the GUI is essentially the application) then I don't
    buy it. Something like Photoshop needs a responsive and native GUI. People
    will not put up with something that feels different and generic which is
    what you'd get with a Java based GUI.



    > The development frameworks that now exist for Java desktop development
    > have
    > made it far more productive than any other development platform. Most
    > people don't get to hear that because of the loud noises coming from
    > Microsoft's marketing department.


    Java has been out for well over a decade. It's a nice language (I've used it
    before) and for over a decade Java has managed to find it's way into various
    parts of the software industry. I'm not claiming that there are no/zero Java
    based GUI apps but developing full-featured native GUI apps is not one of
    Java's strengths.


    ** Posted from http://www.teranews.com **

  13. Re: 1st place prize: Ubuntu 8.04 Long Term Slop

    On Fri, 23 May 2008 13:33:53 +0200, Hadron wrote:


    > Amazing. Kelsey copycat Ian Hilliard uses a program most of us have
    > never heard of to back up his ludicrous claims.
    >
    > I have never seen a successfull Java desktop application. They are slow
    > and resource hungry. And for years the supposed "cross platform" UIs
    > have created ugly garbage. And you know it. And it is well documented.


    A consultant friend of mine has to use training materials from various
    companies and quite often the courses are written in Java.
    I think he has about 15 different versions of Java on his system because of
    all the incompatibilities.




    --
    Moshe Goldfarb
    Collector of soaps from around the globe.
    Please visit The Hall of Linux Idiots:
    http://linuxidiots.blogspot.com/

  14. Re: 1st place prize: Ubuntu 8.04 Long Term Slop

    Moshe Goldfarb (flatfish) in real life Gary Stewart

    http://colatrolls.blogspot.com/2008/...arb-troll.html
    http://colatrolls.blogspot.com/2007/...ish-troll.html

    Traits:

    * Nym shifting (see below)
    * Self confessed thief and proud of it
    * Homophobic
    * Racist
    * Habitual liar
    * Frequently cross posts replies to other non-Linux related newsgroups
    * Frequently cross posts articles originally not posted to COLA

    --
    ! Don Zeigler
    ** Posted from http://www.teranews.com **

  15. Re: 1st place prize: Ubuntu 8.04 Long Term Slop

    [snips]

    On Thu, 22 May 2008 09:47:57 -0400, Ezekiel wrote:

    >> You still don't understand, do you? The point is that the development
    >> should
    >> be generic. If it then turns out that the preferred platform is not
    >> suitable for one reason or another, then it is possible to change to a
    >> platform that is suitable. This adds an extra degree of flexability and
    >> reduces risk.

    >
    > Actually NO. Assuming that you're developing something that you intend
    > on becoming successful then development should target the platform(s)
    > that is most likely going to be used by your customers. It should not
    > simply be "generic."
    >
    > Example: Let's say you plan on writing some sort of plug-in for AutoCad.
    > It sure would make sense for your application to run on the same OS that
    > AutoCad runs on. Not some generic OS where your app runs great but
    > AutoCad isn't available for it.


    If you're developing _for AutoCAD_, you are restricted to developing on
    the platform(s) which AutoCAD runs on. If it won't run on a C64, there's
    **** all point in developing the code to run on a C64.

    Now if you would plug your grey matter in a moment, you'd realize he was
    obviously talking about other notions, applications which are *not* tied
    to particular platforms by being tied to a particular application or
    technology only available on those platforms.

    So take, as an example, a database system not of the scope your local
    company or school might use, but of the scope that, say, Statistics
    Canada might use.

    Such a system presumably should be widely distributed, somewhat
    redundant, resistant to error (eg drive or tape failure), easy to set up
    a new server and join it to the mix, use efficient communications,
    efficient storage and retrieval mechanisms, possibly involve local
    caching of some portions of the data... and this all without even looking
    at the front end, which might be an X terminal, or a Windows box, or even
    a web interface.

    Such a system can readily be developed to run under Windows or Linux or
    OSX or a dozen other systems. Which is the most effective will be
    determined, in large part, by the exact scope and nature and distribution
    and other requirements of the specific client. For one, a collection of
    Windows servers might be the ideal choice, for another, it might be the
    worst possible option.

    If your system is designed flexibly, portably, it can be deployed on
    whichever system meets the client's requirements best. If your system is
    developed *only* to work in Linux, anyone for whom a Windows server back
    end is the best choice won't use your code.

    Developing to be flexible maximizes the benefit to everyone: it allows
    the developers to deploy on a maximal number of possible platforms, it
    allows the client to choose the platform which suits them best... it even
    allows the client to change their mind later and switch to a different
    platform without having to reengineer the entire back end.

    It is maximally beneficial to code for flexibility. Coding for a
    specific platform is, in almost every case, a bad idea, except in those
    few instances where there is simply no point in coding generically... and
    even *that* is based on the assumption that just because AutoCAD is only
    available for Windows today, it won't be available for OSX tomorrow. If
    your extension is written to be too Windows-specific, you lose the option
    to release an OSX version, without considerable cost and effort. If
    you'd written it flexibly, you could produce the OSX version with minimal
    headaches.

    > Or let's say that you're writing some data analysis package for Oracle
    > databases. You would target the most popular platforms that run Oracle
    > and not something like OSX.


    If Oracle is available for OSX, then by writing the code flexibly, you
    can now support Oracle on both - or all - platforms, instead of just
    one. As a developer, why would I want to go out of my way to *limit* my
    potential user base, especially if I'm developing commercially?

    > It is completely IDIOTIC to develop the software for a bunch of
    > platforms first and then afterwards decide that hey... it works best on
    > XYZ so that's what we're going to release it on.


    Actually, it's a perfectly sensible approach, *for at least some
    applications*. You persist in looking only at one class of applications,
    those for which there is already a compelling reason to focus on a single
    platform, rather than looking at all apps, which *include* apps for which
    choice of platform is best left an open option.


  16. Re: 1st place prize: Ubuntu 8.04 Long Term Slop

    On Fri, 23 May 2008 17:43:39 +0800, Ian Hilliard wrote:

    > DFS wrote:
    >
    >> Ian Hilliard wrote:
    >>
    >>> Windows has too much of the market at the moment to be totally
    >>> ignored, but wise developers realize that developing Windows only code
    >>> is also limiting.

    >>
    >> Limited to virtually 100% of the corporate desktops, and at least 90%
    >> of the
    >> home desktops. That's a real limited target platform, huh?
    >>
    >>

    > It may be news to you that the majority of the world's population DO NOT
    > live in the United States. Outside of the USA alternatives to Windows
    > are becoming more and more popular.


    Nor has he clued in - yet - that however popular _desktops_ may be, they
    pale in comparison to computing as a whole.

    He persists in focusing on the big fish in a very, very small pond, while
    ignoring the oceans around him, as if this somehow makes a point.


  17. Re: 1st place prize: Ubuntu 8.04 Long Term Slop


    "Kelsey Bjarnason" wrote in message
    news:k9tog5-l3f.ln1@spankywork.localhost.net...
    > [snips]
    >
    > On Thu, 22 May 2008 09:47:57 -0400, Ezekiel wrote:
    >
    >>> You still don't understand, do you? The point is that the development
    >>> should
    >>> be generic. If it then turns out that the preferred platform is not
    >>> suitable for one reason or another, then it is possible to change to a
    >>> platform that is suitable. This adds an extra degree of flexability and
    >>> reduces risk.

    >>
    >> Actually NO. Assuming that you're developing something that you intend
    >> on becoming successful then development should target the platform(s)
    >> that is most likely going to be used by your customers. It should not
    >> simply be "generic."
    >>
    >> Example: Let's say you plan on writing some sort of plug-in for AutoCad.
    >> It sure would make sense for your application to run on the same OS that
    >> AutoCad runs on. Not some generic OS where your app runs great but
    >> AutoCad isn't available for it.

    >
    > If you're developing _for AutoCAD_, you are restricted to developing on
    > the platform(s) which AutoCAD runs on. If it won't run on a C64, there's
    > **** all point in developing the code to run on a C64.
    >
    > Now if you would plug your grey matter in a moment, you'd realize he was
    > obviously talking about other notions, applications which are *not* tied
    > to particular platforms by being tied to a particular application or
    > technology only available on those platforms.


    No. He was talking nonsense where the claim was that the right way to do
    software development is to develop for many different platforms, then when
    development is finished test the application then release it on the
    platform where it runs best. No mention was made of what platform is
    actually best suited or most widely used by the target customer.



    > Actually, it's a perfectly sensible approach, *for at least some
    > applications*. You persist in looking only at one class of applications,
    > those for which there is already a compelling reason to focus on a single
    > platform, rather than looking at all apps, which *include* apps for which
    > choice of platform is best left an open option.


    Actually it's a ridiculous approach. Do give a single example where it
    makes any sense to develop for a variety of platforms and once, and then do
    internal testing and release the product on the platform where the tests
    run the best. Damn what customers actually use. The only criteria that's
    used to determine what platform the software will be released on is the OS
    where the tests run the best.

    Perhaps this make sense in looney-linux land where you write software in
    your mom's basement but it's a joke of a business model. Any software
    company stupid enough to use this approach to determine what OS to support
    will be bankrupt and out of business in no time flat.





    ** Posted from http://www.teranews.com **

+ Reply to Thread
Page 5 of 5 FirstFirst ... 3 4 5