55 comments

[ 0.26 ms ] story [ 90.2 ms ] thread
I regularly saw ARJ archives on PC BBSes in the mid 90s, next to LHZ which was most popular. At the same time the Amiga world was (and I think still is) dominated by LZX, which supplanted LHA.
Member ace? I think the entire 90s “scene” switched to it for a entire hot minute then back to split rars
Oh, I loved WinAce back in the days!
This brings back a lot of memories! Do you remember the multi steps decompression phases? The par2 files for repairing
I think this is still used in binary use groups. At least it was a few years ago.
And over in Mac-land you had Stuffit
Which is surprisingly powerful as it does various data analysis and compression testing of files in order to pick the best method, and often coming up with fantastic results. Though I'd be surprised if it has a mentionable user base these days.
And LZX ended up as Microsoft CAB format, since Jonathan Forbes went on to work at Microsoft.
Arj was all the rage when it just came out and for some time after that. Early-to-mid-90s or thereabouts.

It was a major step up from pkzip, because not only it compressed better, it could also split the archive into multiple files (making it spreadable across multiple floppies!), it could extract otherwise undamaged files from corrupted archives and had a lot of other options.

Iirc you could split files with Zip at that time as well, but the major annoyance was that you always needed to insert the last disk of the Zip set at the start. This made it so that compressing and copying between 2 computers by moving one or two disks between the computers was impossible, requiring a full set of disks.
pkzip also wouldn't create the split archives on your hard drive. It had to actually write to the disks and experience the disk full error for itself. No fun, because you'd have to babysit the whole process, which would take longer than normal due to the compression, without knowing how many disks you'd need. You also had to just hope that none of them would turn out to have errors.

With arj you could tell it you wanted each part to be whatever size would fit on a 1.44 MB DOS disk, leave it going, then do the copy-to-disk step afterwards having been forewarned about how many disks you'd need.

You could also do this with RAR. I always used the same command (which I could still recite up to the last character) to create compressed files of the size of a floppy, which I believed to be exactly 1423.5 kilobytes. Then again, I'm not sure where did this number come from. I think I used the first disk I could find as a sample, but maybe that disk had some faulty sectors and the amount in a healthy disk was slightly higher.

I remember, back in the day, trying all the compressors I had at hand to see which one compressed the most. Possibly one of my first truly self-researched bit of computer knowledge... not very impressive, but then again this was the mid-90s and I was like 13. I believe that the answer was something like RAR > WinRAR > ARJ > PKZip > WinZip, with the difference between RAR and WinRAR, and between PKZip and WinZip, being very small, and the other two gaps being higher. I probably botched some settings regarding compression levels, so it's likely that I didn't use the best compression level for some of the compressors.

> Iirc you could split files with Zip at that time as well, but the major annoyance was that you always needed to insert the last disk of the Zip set at the start.

I remember that! Was it some technical detail of the .zip format, or just poor design choice?

The "central directory" of a ZIP archive is at the end of the bitstream; so, yes, it's a consequence of the format.
Because of this, if I'm remembering correctly, you can append a .zip file to, say, a .jpg file and it will work most image viewers and as a zip archive. It used to be the case, anyway, I haven't tried it lately.
This is how some self-extracting .exe files work, too. You modify the unzip command to use its own filename instead of an argument, and then you append a zip to the end of that.
I (and others in demoscene and gamedev) often used this "trick" to avoid having extra resource-files and just distribute a single .exe (that was in reality a small exe plus a huge zip appended just behind it).

Works well because Windows just reads the MZ and then PE headers that says where the EXE data is (up to the original .exe size). Then the .ZIP central directory pointer is easily findable at the end of the file so it works fine after appending.

Also spreadable across multiple downloads (or base64ed Usenet posts...), which was very useful to some sectors back in the days of 28.8kbps connections.
I worked with ARJ archives well into 2015. At one point I studied the format to see if it was possible to change the filenames in the archive without unpacking/repacking. It was. And I did.

To be fair, I had ~230k files I had to "manipulate", so even automating a unpack/repack would have cost too much time and energy.

Very nostalgic! My first software was an arj.exe frontend (in French) for Windows 95.

I just launched it! https://i.imgur.com/HfgleBq.png

Of course, it's because ARJ was the standard for warez back in these days...

edit: It was coded in a language called "Visual DialogScript", which amazingly kind of still exists. https://en.wikipedia.org/wiki/Visual_DialogScript

I remember Visual Dialogscript!

I taught myself how to program but couldn’t figure out how to make GUI apps and couldnt afford a Visual Studio license (this is before I discovered pirated sites). I ended up using Visual Dialogscript.

Amazed that this still exists.

In a similar fashion than other people here, my first contact with ARJ was via a copy of Doom compressed and split into several floppies.

Some years ago I remembered out of the blue ARJ and checked their site (the one this submission is about). It was nice to see they were still around. I am the kind of person that love the "About" sections and I could not resist to check their "Our Philosophy" one: http://www.arjsoftware.com/phil.htm:

From the beginning, our family decided that God would be the senior partner in our company and that our business practices would strive to reflect the principles that He has so graciously provided to the world in His Word.

I am not used to people bringing religion into IT related stuff, I was kind of surprised surprised, in a honestly curious way.

(Edited to fix some format)

On the same vein, Zynaddsubfx used to have that on its website: "Please don't use this program to make music that is against God and Jesus Christ. Realize that the only way to the Salvation is Jesus Christ. Please don't lose this chance and don't make others to lose it!"
I did not know the tool (nor the request). Interesting!

Thanks

Excellent synthesizer, totally recommend.
The once ubiquitous UltraEdit text editor for Windows used to have a similar message on its website and still does to this day. I remember it being a little odd back in 1997 or so when I first started using it, especially living in relatively secular Australia.

> This is our story ...to God be the glory!

https://www.ultraedit.com/company/idm-full-story.html

xmodem, ymodem, zmodem and ARJ. Good old times...
Bimodem, baby! Upload AND download simultaneously!
Hydra and Janus! And ZedZap instead of Z-Modem, as ZedZap is much more effective.
I remember ARJ and all the rest of the non-zip compression formats. Back then you could almost be a collector and could certainly be a connoisseur. Fascinating times.
Yeah, I remember being so excited then whenever a new archiver would be released, and running comparisons between them, seeing which ones would compress stuff better than the other. Fun times.
Ha, I was doing exactly the same thing 30 years or so ago; even running comparisons between minor version releases. ARJ was doing very well compression- and performance-wise (and also had open unarj code as far as I remember).
I remember it was fairly typical in DOS days to have a separate directory under utilities just for the different archivers that you might need (ARJ/RAR/ZIP at the minimum, but often also ACE/ARC/ICE/LHA/UHA/ZOO).
At one point, mid 90's maybe, I got the code for the freely available unarj utility and studied it. I remember being blown away by following it and understanding how Huffman compression works. I think that was with the help of a Dr. Dobbs article or something, that was certainly before I had any CS education.
Once in a blue moon I'll feel nostalgic and go see if the web sites for some of my favorite software in the 1990s are still up. ARJ software is one of my regular stops, and so are Semware [1] (I basically lived in QEdit, back in the day) and JP Software [2] (makers of the 4DOS shell).

[1] http://www.semware.com/

[2] https://jpsoft.com/

I haven't thought about either of those in a long time, thanks for the little detour down memory lane!
It’s fun example of vintage software bootstrapping. Love the homemade logo and the gradient fill header showing stripes so it could “save bandwidth” by fitting into a 256-color gif.
In school, back in the early 90s, I was one of the few who'd acquired the eldritch knowledge of making arj compressed files span multiple floppy disks automatically. You see, you needed that if you wanted to copy the pirated games that inevitably got installed on the school's DOS computers. Fun memories.
Ah nostalgia... pirate nostalgia...
Back in the 90s, a good software to use along with ARJ was F-Prot.

MSAV was complete wack, same as its successor Windows Defender.

Off topic: anyone knows a libre RAR5 decompressor? I know PeaZip does it, but it has no CLI.
Back to the nineties... before Linux, I used to love Ms-DoS.