bright@Data-IO.COM (Walter Bright) (08/13/88)
In article <6084@xanth.cs.odu.edu> rlb@cs.odu.edu (Robert Lee Bailey) writes: >Suppose that SEAs attitude had been prevalent at the turn >of the century? Can you imagine the early automobile manufacturers >suing each other because their competitors product also happened to >have 4 wheels, an engine, and a steering wheel? Don't laugh! This happened. All the automobile manufacturers agreed to pay royalties to this one outfit (I forgot the name), except Henry Ford. Ford took them to court and beat them. It also happened in the airplane business. The Wright bros patented the concept of varying the curvature of the wing to effect lateral control, and also the concept of linking this and the rudder movement to one combined control. This basically gave them a hammerlock on the aircraft industry, and there were many attempts to design an airplane that didn't infringe (the Wright bros won all those cases). Lawsuits abounded, and the Wrights lost much popularity about this. This continued until WW1, when the government forced a patent pool to be created so that war production wouldn't be stalled by patent fights. By the way, for those people who are designing new arc formats, here's things I'd like to see that nobody ever talks about: 1. If one byte is bad in an ARC file, the rest of the files are not recoverable. Arc files should have the directory at the start, and if a byte is lost in the arc file, only the file that the byte is in is lost. 2. Arc files should use a fat scheme similar to how DOS stores files on a disk, this is so that files can be deleted or updated without rewriting the entire arc file, which is a major reason why it's slow. Also, the BBS I run will be converted to using ZOO.
deanr@lakesys.UUCP (Dean Roth) (08/31/88)
SEA strikes again, taking PKWare to court Sept. 9. Seems SEA does not like PKWare's PKPAK/PKUNPAK 3.61 reading and writing *.arc files. I think the complaint is along the lines of "contempt of court". As I understand the agreement, PKWare can continue reading/writing .arc files until 1/1/89, or have I missed something? I will publish details when I get them. (I am not associated with either PKWare or SEA. I just think the whole thing stinks. Gee, I can read and write *.arc files using DEBUG. Maybe SEA should sue me, IBM and Microsoft too.) Dean A. Roth deanr@lakesys.UUCP {rutgers, uwvax} uwmcsd1!lakesys!deanr
campbell@hpbsla.HP.COM (gary campbell) (09/14/88)
>By the way, for those people who are designing new arc formats, here's things >I'd like to see that nobody ever talks about: I would like an archiver with compression that can handle multi-floppy archives, primarily as a means of system backup. I have never heard of a freeware offering addressing system backup. This surprises me a bit, with all of the horror stories about backup/ restore floating around. (Is this because noone wants to trust their system backup to a freeware product :-)?) Zoo, with its data compression and its ability to store subdirectories looks like a possibility, but there is the problem of handling multiple floppy backups. The PD Tar just distributed will handle multi-floppy backups, but doesn't compress, at least not on the PC. First, do you know of any existing freeware solution to this problem? It seems like it shouldn't be too hard to add an option to Zoo, either to sense the available room on a floppy, or to supply a maximum archive section size. I don't know the format of a Zoo file, but I assume that there is or could be a header identifier to identify continuations of an archive. I haven't decided whether such a thing should be able to break a file between disks, allowing processing of a file which compresses to a size that is larger than a floppy, or whether to force a file to fit on a floppy. What does Backup do? I think it should be possible to list and extract files from any volume of the archive without having to read the whole archive, and it would be nice to be able to get a complete listing from the first volume along with volume number. This latter might be better done with a log file. Someone on a local BBS made the following comments about compression in a backup utility: "Also, you would have to wonder about the integrity of the backed-up files. If the archive dropped one bit -- it could destroy multiple files within the one archive.". I think that Zoo compresses each file separately, so the part about losing many files wouldn't be true, but is his other concern valid? I am interested in any comments or suggestions you may have on this subject. -- Gary Campbell {decvax,fortune}hplabs!hpbsla!campbell
wheels@mks.UUCP (Gerry Wheeler) (09/20/88)
In article <360004@hpbsla.HP.COM>, campbell@hpbsla.HP.COM (gary campbell) writes: > I would like an archiver with compression that can handle multi-floppy > archives, primarily as a means of system backup. Something which fits that description is the cpio utility in the MKS Toolkit. It is not, however, freeware. On Unix, one would use a command like "find /usr | cpio -ocv | compress >/dev/xxx". On DOS, of course, because there are no real pipes, the entire output of cpio would be saved in a temp file before compress was run. This doesn't work well if you are trying to back up 10 megs of data. :-) So, we added an option to cpio to have it do the compression on the fly. The results are the same as above, but no pipe is involved. Our cpio will also assume that a write error indicates a full disk, and asks if you want to continue on another volume. This allows the cpio file occupy as many floppies as necessary. > Someone on a local BBS made the following comments about compression in > a backup utility: "Also, you would have to wonder about the integrity > of the backed-up files. If the archive dropped one bit -- it could > destroy multiple files within the one archive.". True. You don't want to use compression if your floppies are flakey. (But who would use flakey floppies for backup?) If you're going to start from scratch and create a backup program with compression, you probably want to compress each file separately and try to arrange some way to re-synchronize to the headers within the file in case there is some corruption. Possibly the use of a separate file with pointers into the backup file would work. Or, perhaps the indexes could be stored at the end of the file in a fixed size area. (Not too handy on a multi-volume backup, though.) -- Gerry Wheeler Phone: (519)884-2251 Mortice Kern Systems Inc. UUCP: uunet!watmath!mks!wheels 35 King St. North BIX: join mks Waterloo, Ontario N2J 2W9 CompuServe: 73260,1043