pejacoby@mmm.serc.3m.com (Paul E. Jacoby) (05/14/91)
In article <ZACCONE.91May13183825@jasper.bucknell.edu> zaccone@jasper.bucknell.edu (zaccone - 1393) writes: >I just tried the following experiment. I used Apple's DiskCopy to >make an .image file from an 800K disk. I compressed the resulting >file with both StuffIt Deluxe and Compact Pro. Here are the results: > >StuffIt Deluxe - reduced the file from 838,843 bytes to 18,977 in 60 >minutes (yes, one hour). > >Compact Pro - reduced the file from 838,843 bytes to 1,731 in 10 >seconds! First off, which version of StuffIt Deluxe are you using? There are supposed to be speed improvements in 2.0 (2.01 is the latest); I have yet to find them... :-( Second, are you stuffing with "Best" mode? If so, TURN IT OFF! Use "Best Guess" only. The "Best" mode will try EACH of the compression algorithms and keep only the file which is smallest. As you might guess, it can take a HUGE amount of time to try all four options. I'd guess that is what you are seeing. Note that I saw this problem too (a 2.7 megabyte 4th Dimension database made StuffIt 2.01 appear to _hang_ my machine for nearly as long), and Aladdin let me know about the "Best" mode situation. Didn't spot _that_ useful tip in the docs. -- | Paul E. Jacoby, 3M Company | | | Maplewood, MN 55144-1000 | Parachuting? Why jump out of a | | => pejacoby@3m.com | perfectly good airplane? | | (612) 737-3211 | |
folta@tove.cs.umd.edu (Wayne Folta) (05/15/91)
>Second, are you stuffing with "Best" mode? If so, TURN IT OFF! Use >"Best Guess" only. The "Best" mode will try EACH of the compression >algorithms and keep only the file which is smallest. As you might >guess, it can take a HUGE amount of time to try all four options. I'd >guess that is what you are seeing. > Note that I saw this problem too (a 2.7 megabyte 4th Dimension >database made StuffIt 2.01 appear to _hang_ my machine for nearly as >long), and Aladdin let me know about the "Best" mode situation. Didn't >spot _that_ useful tip in the docs. I really like StuffIt Deluxe--it seems to be a very elegant program. *Except* for the slowness of its compression. I believe that there is a bug in the compression, since certain files cause it to act radically different. Example: I compressed two Nisus files, two TIFFs, two APPLs and a QXP file. The smaller TIFF (about 50Kb) took as long to compress as the other 700Kb of files combined! This was using Best Guess (twice as fast as Better, same compression). I've got to believe that certain binary combinations cause its compression code trees (if it is using such things) to be changed repeatedly, or something like that... -- Wayne Folta (folta@cs.umd.edu 128.8.128.8)
zaccone@jasper.bucknell.edu (zaccone - 1393) (05/15/91)
In article <1991May14.125138.8162@mmm.serc.3m.com> pejacoby@mmm.serc.3m.com (Paul E. Jacoby) writes:
First off, which version of StuffIt Deluxe are you using? There are
supposed to be speed improvements in 2.0 (2.01 is the latest); I have
yet to find them... :-(
Second, are you stuffing with "Best" mode? If so, TURN IT OFF! Use
"Best Guess" only. The "Best" mode will try EACH of the compression
algorithms and keep only the file which is smallest. As you might
guess, it can take a HUGE amount of time to try all four options. I'd
guess that is what you are seeing.
Try it yourself. Use DiskCopy to make a .image file of an 800K
*blank* disk. Use StuffIt Deluxe version 2.0.1, best guess, and don't
allow background processes. (Aladdin says this combo gives the best
performance.) Compess the file with both StuffIt Deluxe and Compact
Pro. On a Mac IIci, Compact Pro ran 360 times faster, and the file it
produced was 11 times smaller!
I'm not claiming that this result is characteristic. I've seen
instances where StuffIt Deluxe performs better than Compact Pro, but
it's always by a small amount. There are many more cases where
Compact Pro performs dramatically better.
Rick Zaccone
--
zaccone@bucknell.edurang@cs.wisc.edu (Anton Rang) (05/16/91)
In article <34437@mimsy.umd.edu> folta@tove.cs.umd.edu (Wayne Folta) writes: >I really like StuffIt Deluxe--it seems to be a very elegant program. *Except* >for the slowness of its compression. I believe that there is a bug in the >compression, since certain files cause it to act radically different. I'm not sure if it's a bug, but it could be. Large strings of 0 bytes make it *really* slow. I haven't reversed engineered it to find out why (violation of license agreement, and anyway more importantly I haven't got a working debugger 'till I buy the newest Macsbug :-). If somebody's got practice with data compression, they might try writing a custom compressor. Aladdin has docs on that which they kindly sent me via e-mail a while back (no, I can't redistribute them, but you should be able to get them just by asking). Maybe someone out on the net could implement one? (I'm planning on trying the LZ-based algorithm from Australia which is s'posed to be really fast, when I have free time this summer.) Anton +---------------------------+------------------+-------------+----------------+ | Anton Rang (grad student) | rang@cs.wisc.edu | UW--Madison | "VMS Forever!" | +---------------------------+------------------+-------------+----------------+