Re: [NTLK] OT: Wasted CPU cycles

From: Steven Scotten <splicer_at_paroxysm.com>
Date: Sun Mar 02 2008 - 18:34:55 EST

On Mar 2, 2008, at 12:18 PM, Martin Joseph wrote:
>
> This bloat is generally unfelt by the end user, as the speeds and
> disk spaces involved in the newer systems make this bloat moot.

There are two things to remember about code bloat.

First thing is that people want their upgrades yesterday. If a
developer takes the time and resources to write truly well-optimized
code, another developer will get to market with the same product first
and promise to fix the bugs later. The people buying software don't
seem to mind, or at least don't mind enough to stop buying buggy and
slow software.

Second thing is to remember the value of different forms of
optimization. A programmer can optimize for speed, or for readability
and maintainability, or for resource usage. The fastest, tightest,
most efficient code can conceivably introduce bugs down the line when
some other programmer *thinks* they know how it works but
misunderstands, just as an example.

Computing is perhaps less efficient than it was 20 or 30 years ago,
but it's much more complicated and programs are more likely to be
interdependent, making it unnecessary for developers to reinvent all
the wheels. A programmer 30 years ago would be writing the tightest,
fastest routines for moving a cursor from one place to another on the
screen. Today's application developers don't have to spend time on
such things.

That said, it would be nice if processor speeds would advance faster
than developers make things slower. I don't like that my 2GHz computer
takes longer to boot than my old 4.077MHz 8088.

Steve

====================================================================
The NewtonTalk Mailing List - http://www.newtontalk.net/
The Official Newton FAQ - http://www.splorp.com/newton/faq/
The Newton Glossary - http://www.splorp.com/newton/glossary/
WikiWikiNewt - http://tools.unna.org/wikiwikinewt/
====================================================================
Received on Sun Mar 2 18:34:58 2008

This archive was generated by hypermail 2.1.8 : Sun Mar 02 2008 - 22:30:00 EST