I have archives of almost all my old programming and hardware designs going back 20 years or so. I was poking through some of it, written for DOS, 80 x 24 screens. No GUI. I have to detect monochrome monitors, Hercules Graphics, ISA backplane boards, fixed I/O ports, swapping overlays in and out of the huge 256 KB memory to make large (for that time) programs possible.

What a nightmare. But the program I was looking at (a DSP chip debugger) was about 400 KB. Source is about 200,000 lines of C. It is hard to find a big program of merely 400 KB these days.

Another one was an audio disk recorder using SCSI disk drives attached to a PC (sort of a precursor to what Avid's Digidesign became) circa 1988. The whole thing was 270 KB, including a custom handwritten DOS for the SCSI drives. With the slow processors of those days (8 MHz was a screamer IIRC), I had to mix in assembly language with the C code to get it all to work in real-time. I learned to write ASM on Z80s and DSP chips. Wonderful stuff. No debuggers.

Yes, I had no life in those days; married but no children yet.

All the bloatware that passes for software these days is just a big conspiracy to sell more RAM. But I appreciate the 4 GB RAM, the Quad Xeons, the six monitors, the PCI-E bus, the SATA2 and 320 MB/second 15K RPM SCSI drives on my current PC even though it bogs down when the market is running too fast. Sigh - can't win.

Tim Richmond adds:

Bloatware occurs because 1) RAM is relatively inexpensive, 2) clock speeds continue to increase, and 3) Moore's Law continues to hold true. Software developers will not be compelled to create more efficient, streamlined code when semiconductor physics offers them such an unconstrained playing field.

Alex West remarks:

I started programming on Turbo Pascal 5.5, IBM PC XT, 640 RAM. All I needed was stored on 3" disk: turbo.exe, help file and library file. It was indeed a very good time.





Speak your mind


Resources & Links