On 10/6/2017 1:54 PM, Jack Harper via cctalk wrote:
Interesting Stuff...
One of my heroes - Ray Kurzweil
https://en.wikipedia.org/wiki/Ray_Kurzweil - says that when he builds a
new product that he always projects what the computer horsepower will be
like when the product is ready to launch.
Say, an 18-month development cycle, then design the product to run on a
computer with twice the processing power as available at the start of
the project.
Obviously, he is not advocating bloatware - far from it.
Go for products that cannot effectively work today because of the
processing requirements, but will work when released - things like
AI/deep learning/advanced DSP/image understanding etc.
Regards to the List from the Rocky Mountains.
Jack
ps - I HATE and LOATHE bloatware - e.g., so much MicroSoft stuff.
I hate the Graphics interface concept like the Web Browsers.
Click on this, click on that! Wait 30 seconds for some banner ad to
download so you can read your page. So Jack does your favorite
developer send a new PC so you run his product?
I suspect the real problem (other than NEW FEATURES you must have
gimmicks) is software is TOO generic, rather useful but reasonable
limitations. Do we need PDF files to scale 109% for viewing, when
say 25,50,75,100,125 seems practical and quicker,
Ben.
PS: Still grumbling over the FORTRAN COMMON FEATURE that software can't
sort local variables and arrays into separate segments to have
a clean cashe, so don't mind me.
