Re: 65020 CPU? Look this...

["Michael J. Mahon" <mjmahon@aol.com>]

Ranando King wrote:
> We do agree on 1 thing. Programmers have gotten very, very sloppy with their
> larger RAM and faster CPUs. Look at Windows for instance. Why the hell does
> it take a new PC to run the next version of Windows at the same speed as the
> previous version on hardware 1/10 the power? Yeesh!

It is an amazing and wierd situation to one with an engineering
mindset.  However, it makes perfect sense to a marketing person.  ;-)

As recently as 1998, Microsoft's VP of technology explained to an
auditorium of ACM conference-goers that their job was to keep on
consuming storage and processing resources at an exponential rate
so that there would be a market for the new products of Moore's
Law.  (He also noted the similarity of a plot of Microsoft's stock
price to Moore's Law.  ;-)  Now it seems that both have stalled out.

As a member of the industry, I pointed out that the industry dynamics
were intimately related to maintaining the exponential growth predicted
by Moore's Law, and that any reduction in the rate of improvement would
cause countless disruptions and decreased sales.  In short, the whole
industry had become "drunk" on a rate of growth that must certainly
moderate, with disasterous "hangover" effects.

My prediction was that Moore's law would begin to signifcantly
decelerate in this decade, for reasons dictated by physics if not
for technology issues.  It looks like I was optimistic, since the
rate of improvement of processor speed has all but stopped, as
Intel discovers (as most did much earlier) that longer and longer
pipelines reach a point of diminishing returns.

> I don't think this trend is the result of the changes in hardware though. I
> think the new generation of programmers are being innundated by the "just
> get it done" mindset. They don't care much for efficiency and creativity in
> trying to make their code as small and clean as possible like we used to do
> with our Apples. Today's programmers are simply comparatively either lazy or
> sloppy. I don't know how many times I've cringed at the memory requirements
> to run an application I've written. I look at a memory manager and watch it
> take up 8 megs trying to run a 256K application that only allocates 512K of
> memory for itself. The rest of the memory is being chewed up by libraries I
> needed and Windows management code. Ridiculous!

Absolutely.  But it is the hardware advances that have enabled this
laziness to proliferate and even prosper.  This mindset would have
been completely bankrupt in the world of scarce memory and processor.

> What makes things even worse is that hardware designers are following the
> same sloppy trends with their component designs. Yeesh!

And for many of the same reasons.  The cost to design a system
is often huge compared to the cost of manufacturing it, so cutting
engineering costs and getting to market sooner is the name of the
game.  If innovation and efficiency take a back seat, so be it.  ;-(

> I still think that there's a way. If the simplicity of an Apple were to be
> applied to a 32 bit machine and developed by people who want to squeeze the
> most out of the system, we could see some highly efficient and creative
> programs appearing for it. Such programmers still exist, even among newer
> developers. Look at game developers. They have to do exactly that on a daily
> basis. So the possibility still exists.

Agreed, but even game developers have been seduced by "more is better".
Today, more time and money is spent on motion capture studios and
graphic design than on playability and fun.  "Giving great demo" is
more important to making a sale than remaining interesting and
enjoyable after the purchase.

> "Michael J. Mahon" <mjmahon@aol.com> wrote in message
> 9_-dnZ6HxJg1VrnfRVn-2A@comcast.com">news:9_-dnZ6HxJg1VrnfRVn-2A@comcast.com...
>
> > I understand where you are coming from, and I respect your opinion.
> >
> > My point is that the finely crafted code you and I both so admire
> > using, reading, and writing, becomes a "waste of programmer's time"
> > on a machine that has just an order of magnitude more speed and linear
> > RAM.
> >
> > When you don't have much speed and space, you always wish that you
> > had more; but if you get it, all hell breaks loose--in the form
> > of lost discipline.
> >
> > In the area of programming craft, we have seen that high speed and
> > large space result in the widespread proliferation of:
> >
> > -sloppy algorithms,
> > -sloppy compilers,
> > -sloppy libraries,
> > -sloppy interfaces, and, ultimately,
> > -sloppy programmers.
> >
> > I still adhere to the principles of economy and synergy that guided
> > Woz and so many other great designers.  I find that there is such
> > a thing as _too much_ memory and speed for one's programming integrity.

-michael

New Applesoft BASIC interface for NadaNet networking!
Home page:  http://members.aol.com/MJMahon/