This archive contains an updated and improved version of ARC 5.21 called for identification purposes ARC 5.21C. ARC 5.21C features: o larger buffers for file buffering (28,672 bytes as opposed to 512) for greater speed. o new data compression algorithms for greatly improved speed AND compression. o allowing invocation of 12, 13, 14, 15, or 16 bit file crunching. To select bit size, insert 12-16 in the command line (e.g. "a16 sample game.prg", "c14 sample", etc.). Each larger bit size requires more free memory. The resultant compression ratio grows more dramatic the larger the file. o on some files, 12 bit crunching (the default) results in a smaller compressed file than ARC 5.12 produces, although it is completely compatible with the original ARC 5.12. o a bug in all version of ARC that causes a crash whenever a verbose list of a file with an invalid month (less than 1 or greater than 12) has been fixed. I've run into this on STart disks. o the bug in ARC 5.21 that allows an archive to be built that is larger than the sum of its included files has been fixed. o the original wild card expansion algorithm has been reinstated ('*.*' means ALL files again, not just all files with extensions). The speed improvements are sometimes dramatic over earlier versions of ARC. Below are the results of running a suite of functions between each version of the ARC program currently circulating. To conduct the tests, two double sided disk drives were used. All temporary files were written to a floppy disk. Some randomly selected files of various types were read from disk B:\ to an archive created in disk A:\. After each archive was finished, the archive was tested, which actually extracts the file without writing it out to disk. Then a "verbose" listing was made of the archive. The exact elapsed time for these operations was then recorded. Then a disk directory was recorded that shows the file length of the completed archive. Each version of the ARC program started with the same disk setups. Each test was done from a script without any human intervention. The time and size differences reflect only the performance of the ARC programs themselves. ========================================================================= arc512.ttp Name Length Stowage SF Size now Date Time CRC ============ ======== ======== ==== ======== ========= ====== ==== EXAMPLE.PRG 116939 Crunched 36% 75184 1 Dec 89 9:53p 99A0 PICTURE.NEO 32128 Squeezed 51% 15816 1 Feb 88 9:24a 6B0F RLE.DOC 505 Crunched 92% 44 1 Dec 89 10:03p 5008 SAMPLE.ARC 7808 Packed 1% 7758 1 Dec 89 10:35p CD5A SAMPLE.DOC 189180 Crunched 48% 98922 2 Dec 89 9:24a 4ADC SUMTHIN.SEQ 96572 Crunched 58% 40789 1 Dec 89 10:27p FE71 TEST.ARC 127693 -- 0% 127693 1 Dec 89 11:30p 5452 ==== ======== ==== ======== Total 7 570825 36% 366206 elapsed time: 00:28:16.655 directory listing: -rw-- 366411 Sat Dec 2 09:24 test.arc ========================================================================= arc521.ttp Name Length Stowage SF Size now Date Time CRC ============ ======== ======== ==== ======== ========= ====== ==== EXAMPLE.PRG 116939 Crunched 42% 67922 1 Dec 89 9:53p 99A0 PICTURE.NEO 32128 Squeezed 51% 15816 1 Feb 88 9:24a 6B0F RLE.DOC 505 Crunched 92% 44 1 Dec 89 10:03p 5008 SAMPLE.ARC 7808 Packed 1% 7758 1 Dec 89 10:35p CD5A SAMPLE.DOC 189180 Crunched 47% 102118 2 Dec 89 9:24a 4ADC SUMTHIN.SEQ 96572 Crunched 58% 41019 1 Dec 89 10:27p FE71 TEST.ARC 127693 -- 0% 127693 1 Dec 89 11:30p 5452 ==== ======== ==== ======== Total 7 570825 37% 362370 elapsed time: 00:20:54.135 directory listing: -rw-- 405406 Sat Dec 2 09:24 test.arc ========================================================================= arc521b.ttp Name Length Stowage SF Size now Date Time CRC ============ ======== ======== ==== ======== ========= ====== ==== EXAMPLE.PRG 116939 Crunched 42% 67922 1 Dec 89 9:53p 99A0 PICTURE.NEO 32128 Squeezed 51% 15816 1 Feb 88 9:24a 6B0F RLE.DOC 505 Crunched 92% 44 1 Dec 89 10:03p 5008 SAMPLE.ARC 7808 Packed 1% 7758 1 Dec 89 10:35p CD5A SAMPLE.DOC 189180 Crunched 47% 102118 2 Dec 89 9:24a 4ADC SUMTHIN.SEQ 96572 Crunched 58% 41019 1 Dec 89 10:27p FE71 TEST.ARC 127693 -- 0% 127693 1 Dec 89 11:30p 5452 ==== ======== ==== ======== Total 7 570825 37% 362370 elapsed time: 00:15:46.390 directory listing: -rw-- 362575 Sat Dec 2 09:24 test.arc ========================================================================= arc.ttp (version 5.21C) Name Length Stowage SF Size now Date Time CRC ============ ======== ======== ==== ======== ========= ====== ==== EXAMPLE.PRG 116939 Crunch12 44% 65763 1 Dec 89 9:53p 99A0 PICTURE.NEO 32128 Squeezed 51% 15816 1 Feb 88 9:24a 6B0F RLE.DOC 505 Crunch12 92% 44 1 Dec 89 10:03p 5008 SAMPLE.ARC 7808 Packed 1% 7758 1 Dec 89 10:35p CD5A SAMPLE.DOC 189180 Crunch12 49% 97696 2 Dec 89 9:24a 4ADC SUMTHIN.SEQ 96572 Crunch12 58% 40762 1 Dec 89 10:27p FE71 TEST.ARC 127693 -- 0% 127693 1 Dec 89 11:30p 5452 ==== ======== ==== ======== Total 7 570825 38% 355532 elapsed time: 00:09:52.300 directory listing: -rw-- 355737 Sat Dec 2 09:24 test.arc ========================================================================= It's interesting to note that ARC 5.21C performed the functions 3 times faster than the original 5.12, and that's without using a ramdisk or hard disk for temporary files! Also note that while EXAMPLE.PRG is smaller in 5.21C, the original ARC 5.12 has no trouble whatsoever with extracting it. Notice that both EXAMPLE.PRG and SAMPLE.DOC compress to a larger size in the ARC521 archive. And look at the final size of the archive, 405,406 bytes! Even ARC521B pales with both compressed file size and elapsed time, although at least it doesn't add garbage like ARC521. Enjoy! Patrick Dell'Era Compuserve 73157,2623 GEnie DELL-ERA Delphi DELL_ERA