mwm@ea.UUCP (04/27/84)
#R:ritcv:-102800:ea:800002:000:2097
ea!mwm Apr 27 17:16:00 1984
/***** ea:net.arch / ritcv!kar / 7:56 am Apr 25, 1984 */
Ignore those machines and maybe they'll go away. Architectures that
require aligning multi-byte quantities on particular boundaries sound to me
like they were designed by an engineer who was interested in simplifying his
own task at the expense of the software designers who will use the machine.
Ken Reek, Rochester Institute of Technology
{allegra,seismo}!rochester!ritcv!kar
/* ---------- */
True, but people insist on building machines with silly restrictions on
boundaries - like having to fetch objects that are multiples of 8 bits long
off bit boundaries that are multiples of 8. Given that, I live with what's
available.
If this means a machine that has restrictions on what you can fetch from
where, well, I live with it. If it means one that it costs extra for having
things off alignment (like the VAX), that's great. I don't know of any
machines that don't charge you when you fail to align things, so I'll
decide what that will be when it happens.
I agree that designing machines with such boundary restrictions is
short-sighted, and making off-boundary fetches more expensive isn't
much better off, but there are two things that you have to consider:
1) As you mentioned, machines don't live nearly as long as
software does. Therefore, extra dollars spent on
hardware are harder to recoup from sales. So moving
said dollars from hardware development to software
development may be a good idea.
2) The extra time spent developing the hardware to do what
what you want may result in a considerable delay
in product release. This could well cause your product
to flop completely.
In short, in an industry that changes as rapidly as this one does, being
short-sighted could well be an advantage. After all, if you have the worlds
slickest processor on silicon built with 5 year old technology, you've got
something that won't sell. Conversely, having an antiquated chip design out
the door before anybody else has there slick competitive chip out can make
you lots of money (witness the 8086).
<mike