[Oberon] Ceres documents scanned

[eas lab]

curl -I ftp://ftp.inf.ethz.ch/pub/publications/dissertations/th8431.pdf\
 >> ShowPdfSize

cat ShowPdfSize  ==
Last-Modified: Fri, 07 Dec 2012 12:50:05 GMT
Content-Length: 5'834'230  <-- file-size
Accept-ranges: bytes

OK, I don't like BIG pdf, but the abstarct tells all:---
> The analysis of processor-memory communication is motivated by the
> observed small benefit in performance gained by replacing the NS32016
> CPU with the NS32032 CPU and thus doubling the memory bus bandwidth.
> [Measurements] show that the bus capacity of the NS32032-based Ceres is
> only used to a small degree. Therefore, the additional costs can hardly
> be justified.  This contrasts with the frequently heard claims of the
> superiority of 32-bit computers.

In the previous thread, where the writer asked "why don't you use
XYZ virtual-machine; it's GOT EVERY THING?". He missed the fact that
Wirth's key is to have LESS to burden/confuse the mind.