In article
,
(Robert Myers) wrote:

In all this "I knew all along" talk about Itanium, I've heard a few
insightful comments indicating that people actually understood
something of importance about the actual architecture, and not what
they've heard from others.

Indeed, at first, it looked convincing. My intuition still tells me that
it ought to be extremely fast, but when it clashes with repeatable
measurements, I have to believe the measurements.

If anyone *really* understood what went wrong with Itanium, it
would make the case study of all case studies for business
schools interested in the development and management of technology.
As it is, I don't think anyone really knows.

Here's two non-technical things that contributed to people getting fed
up with Itanium:

* Penny-pinching in the wrong places. The original Itanium I development
systems were not fast, but they were solid and workmanlike. Most of
them were shipped with early-stepping processors. There were some
upgrades, but if you wanted to upgrade from a B-series with some
significant errata to a fully working C-series, it cost you about
US$1000 per processor (and a lot of them were duals). It appears MS
found themselves facing a bill for a million or so dollars to get
proper hardware. That's the kind of thing that makes people ask
"why are we doing this?"

* Differences in objectives between Intel and HP. HP just wanted it as a
PA-RISC replacement, Intel wanted it as an everything replacement.
Notably, Intel allowed HP to have a monopoly in Itanium II systems
with 4 or less processors, like the ones used for most software
development. HP gouged on prices for these, and Intel abandoned
support on all the Itanium I systems they had given out - even
after you'd bought new processors for them - so you had no choice
but to buy new ones from HP.

I have very little sympathy for the concerns of software developers.

Funny, I always thought software was what made computers useful?

We'd be much better off with longer software development cycles so we
had less bad software.

Yes, we would. But the current economics mean that people perceive more
advantage in first-to-market than best-quality. If you would like to
start a political movement to change that, go right ahead.

--
John Dallman
"C++ - the FORTRAN of the early 21st century."