|
In lugnet.cad.dev.mac, Andrew Allan wrote:
> Therefore if anybody else wants to try my crude experiment, please send me the
> results.
>
> Andrew...
Thanks for the MBC update!
PowerBook G4, 400MHz, 512MB RAM, IBM TravelStar 40GN 40GB HD (4200 RPM), OS X
10.3.3 (build 7F44), 8MB Rage128.
I don't have the .mpd extension mapped to MBC, so I dropped the Star Destroyer
MPD on the MBC app. Timed with menu bar clock.
-55 seconds from drop to first display (doesn't seem to matter much whether or
not MBC is running).
-12 seconds to redraw orthographic view.
From your description, I figure the slower spindle speed of the notebook HD
bears some responsibility for the slowdown compared to your 400MHz Yikes.
One other thing, if you rotate a piece 90 degrees (or an integer multiple n
thereof), some coordinates of the transformation matrix that should be zero are
instead +/- n*4.37114E-08. From a look at the source code (2.2, where this also
occurs), it looked like there were some places where the digits of pi had been
truncated to 3.1415, which I think may be the right magnitude of error for this.
Obviously, the error itself is trivial, but it makes the model text hard to
read, particularly if the display is taking up a lot of the width of the screen
and the model text drawer is on the side.
Tom
|
|
|
In lugnet.cad.dev.mac, Tom Bozzo wrote:
> One other thing, if you rotate a piece 90 degrees (or an integer multiple n
> thereof), some coordinates of the transformation matrix that should be zero are
> instead +/- n*4.37114E-08. From a look at the source code (2.2, where this also
> occurs), it looked like there were some places where the digits of pi had been
> truncated to 3.1415, which I think may be the right magnitude of error for this.
> Obviously, the error itself is trivial, but it makes the model text hard to
> read, particularly if the display is taking up a lot of the width of the screen
> and the model text drawer is on the side.
>
> Tom
I'm aware of this and you'll find that when you save and re-open the file, MBC
will have truncated these very small numbers to 0, this code should however one
day be incorporated into the rotation routines.
Andrew...
|
|
|