Saving Software
Tuesday, April 1, 1997 by Dave Winer.
Can you tell if it's an April Fools DaveNet or not? I've never admitted to writing one. It's a matter of pride with me. On a good day I mislead just by speaking. I do the best I can, but I can change my mind. It happens all the time! So why go out of my way to lie? Good question.
Anyway, I have an evangelistic plea today to follow my enthusiastic Fat Pages With Wheels piece yesterday. I'm slowing down a little, taking a few days to think and absorb. Learning a lot.
Anyway...
Fred Davis 
Fred Davis, fred@weblust.com, is a computer journalist, author of the Windows Bible, a former consultant to Bill Ziff at Ziff-Davis, and now a one-man think tank for CMP. I've known Fred since our Apple II days when he was founding editor of A+.
Fred's latest project is the San Francisco Computer Museum. He's got ambitious plans, I've been hearing about them periodically, it sounds cool. But last week he told me a chilling story that's been bothering me ever since.
Old movies
The film industry is setting up a system so that prints of movies will be viewable a hundred years from now. There have been disasters, classic movies that have been lost or badly damaged. They're getting systematic about doing backups so their work has historic value and as an investment in the future.
Software is different
Software is different from film. Even slight damage to a piece of software can render it unusable, where a film can sustain a lot of damage and still be viewable.
In film there are just a few formats, a few different kinds of players. Software is more complex. My story is an example, in the late 70s I did development on a Cromemco Z2D, with 8-inch floppies, running the UCSD P-System on top of CP/M. The machine is long gone, but I still have the floppies containing the code I was working on. Does anyone else have a machine enough like this one to read its floppies? I doubt it.
Then I switched to an Apple II running the P-System, then to an IBM PC running Microsoft Pascal on MS-DOS. Then a big discontinuity, a new source code base on a Macintosh using Think C (originally called Lightspeed C). Now we use Metrowerks and Microsoft's development environments on Macs and Win32.
The point -- I don't have a trail of working computers behind me that can read all those disks. I have no idea if any of it is readable or usable. Even though I have the disks -- there's no way of being sure that the formats of the disks are readable by any computer I could put my hands on.
Perspective
We leave things behind. Our industry is always in panic mode. We never have time to save our work, to be sure that we leave a sure trail behind us. There's no backup. Things get lost.
And there have been disasters, in major companies, that haven't been reported in the press. Has Apple ever explained to its investors how it protects its software assets? Has Microsoft? Netscape? Has the company you work for or own stock in?
Even with systems in place there can be disasters. But most software isn't systematically archived.
I think this is bigger than the Year 2000 problem.
What Fred's doing
Back to earth. Fred is the only person I know who's doing something about this problem. What's he doing?
He's collecting computers and getting them running. An example, he believes he has the only bootable Xerox Star on earth. I hope he's wrong! He's also collecting software.
I'm glad that Fred decided to take this on. Let's help him if we can.
http://www.fog.com/sfcm/sfcm.html
Dave Winer
PS: I started archiving my stuff this weekend. It's going to be a big job.