Re: normalization of pointers...

Malcolm McLean wrote, On 31/05/07 22:45:

"Mark McIntyre" <markmcintyre@xxxxxxxxxxx> wrote in message news:m0fu53t3lbmk1t187jtjem0n54nf6o3128@xxxxxxxxxx
On Thu, 31 May 2007 07:46:28 +0100, in comp.lang.c , "Malcolm McLean"
<regniztar@xxxxxxxxxxxxxx> wrote:

In most PC programs most of the time is spent on the user interface. You are
writing to an API, but one very closely tied to the hardware.

You may want to rephrase that as "in most PC programs I have worked
with", to cater for those like myself whose experience has been
diametrically opposite.

I've worked on lots of apps that ran on PCs, but mostly they were a
thin layer of platform-specific UI over the top of a huge wodge of
analytical and computational code which was designed to be independent
of hardware. I forget when I last had the luxury of programming solely
for less than three platforms, though the mix has changed from
NT/SunOS/VMS to XP/Solaris/RHEL. Unless you count VBA....

A few years ago the general rule was that the amount of time needed to write a program was determined by the complexity of its user interface.

Things have changed a bit now. Computers can carry out far more intensive computations, and user-interface development technology has improved.
However I think there is still some truth in it.

It always has depended very much on the type of SW you are doing. I used to do SW which was controlling test equipment and vastly more time was spent writing code to control the equipment and analyse the results than on doing the display, this back in the days before having a GUI was common when we still has to display graphs. More recent simpler stuff (running on W2K) I spent all of a couple of hours doing the display and the rest of a week or two on the remainder of the SW.

The stuff I've spent the last few years doing, a small new routine might be dominated by the time doing the screen layout, but on most stuff more than a day the dominating factor is the business logic and database work. So two tasks one taking a day and one taking a few weeks can have close to the same amount of complexity on the user interface.

On other types of SW what you say may well be true. It's just not true on the types of SW I've worked on or that Mark has worked on.
--
Flash Gordon
.

Relevant Pages

  • Re: Macbook Pro und Linux
    ... bekomme, welches einen C2D mit 2-4GB RAM einer 160GB HD einen Display, ... Kenne nur das Display eines ca ein Jahr alten MBP meiner ... die auch seitens des Herstellers unter Linux unterstützt werden. ... Wie in einem anderen Posting schon angedeutet wurde nimmt man Apple-Hardware primär wegen des Betriebssystems Mac OS X, nicht weil die Hardware selbst so qualitativ hochwertig wäre. ...
    (de.comp.sys.mac.misc)
  • Re: What do you take for secure programming?
    ... Neverthless there plenty of them very confident on the result ... I am writing them alone, ... >> Important to know these books is all about about programming, ... > and even having to deal with hardware errors... ...
    (SecProg)
  • Re: Term for kana subtext on kanji?
    ... >> I think, usually, Japanese reader program can display vertical ... >> Japanese novel books should be read. ... >> editors don't support vertical writing while they bothered to ... >> care to support stupid things like Unicode. ...
    (sci.lang.japan)
  • Re: Panic on 5.1-RELEASE with ahc
    ... > All the hardware seems to be detected okay, and I've rebuilt a kernel ... > kernel too ISTR) is that on shutting down, I consistently get a panic, ... > RAM parity error, likely hardware failure. ... > This also causes the Hardware LCD display on the front to display "E13F5 ...
    (freebsd-current)
  • Re: F#
    ... hardware can be spent in paying the developers). ... and you can make N more customers ...
    (comp.lang.functional)