[e-lang] Pipelining figure
markm at cs.jhu.edu
Tue Jun 28 18:50:52 EDT 2005
Ka-Ping Yee wrote:
> (The messages are in transit between the vats.)
Looks good. I prefer to have the b() message farther to the left, so that
their right-to-left horizontal order parallel the order in which they're sent.
This helps suggest that the inter-vat stream is also a FIFO-order preserving
Given that b() is to the left of a(), perhaps X should be shown to the left of Y?
Vertical space is precious. The gray bar at the bottom of this figure is
unnecessary. (Yes, I know it's a placeholder for the next diagram's queue, but
it's still unnecessary here.)
> (The messages are queued in the second vat, and the result
> pointer is shown.)
The c(r2) message is shown in the wrong position. It is queued within the r1
promise. It is *not* queued in VatB's queue until this promise has been
resolved. (And even then, only if it's resolved to an object on VatB.
See p52 of <https://www.cypherpunks.to/erights/talks/promises/promises.pdf>
while imagining all the distractions removed.
(Better yet if you have Powerpoint for Windows 2002 or later, please watch the
> This figure is bigger and smaller at the same time. By this, i mean
> the figure itself has more detail, and if it were shown at the same
> scale as the previous version it would be larger. But it has a wider
> aspect ratio, so if you scale it to the text width, it will actually
> take up less space in the paper.
So long as something fits within the horizontal margins, vertical space is all
that matters. But rather than scaling the .eps files with a LaTeX directive,
could we revise these .svg files so that they simply fit? The advantage is
that we'd then have simpler control over what font size we're using.
> All of my previously posted files related to this paper have been
> moved into the http://zesty.ca/promises/ directory.
Grabbing 'em now. Thanks!
Text by me above is hereby placed in the public domain
More information about the e-lang