Expandability - does it add value?

Ben prompted a question:

We often find that a home computer offering is expandable or upgradeable to some degree. But, how often do we actually use that capability? Is it worth something, in terms of extra price? What if the expandability somehow compromises the initial product?

Thoughts and experiences? (I’ll add my own a bit later)

From Old Computers, the TI-99/4A, as an example:


The big issue is whether or not expansions are supported. For my C64, the 1541 floppy drive was extremely well supported, and my printer was supported by everything that reasonably would support it.

I don’t really remember my 3rd party 300 baud modem being supported by anything other than its modem software, but that was good enough for me.

My first experiences with 8-bitters were the Apple II line, and I think their expandability was pretty key – particularly compared to the original ][ or even the ][+. In particular, an 80 column card, extra RAM, and a Disk ][ made a much more capable machine! The first computer I had at home was a //c, which was essentially not internally expandable, but already had the essential expansions for a //e (80 cols, RAM, disk, Super Serial for a printer or modem) built in.

Many of the other 8-bit console-style computers suffered from not having actual internal slots (as in the TI-99/4A pictured above!), or only dedicated-function internal slots (like the Amiga 500 RAM header), which made expansion a bit more painful.

The IBM PC was in many ways similar in this respect to the Apple ][; out of the box, it just didn’t have enough Stuff (and in particular it had no display at all!), so a couple of expansion cards made it a much more usable and useful computer.

I have at various points on my own journey used relatively niche expansion cards or products for various reasons. For example, when I was in college I used an ISA television capture card to save the space of a television in my dorm room without giving up television access. That said, in many cases, a relatively capable external peripheral bus (such as SCSI in the 90s) would probably have served me just fine.

1 Like

It depends on what you need.

Thinking about my first PC - Sperry HT, an IBM-XT clone. 256K of RAM. Monochrome. 2 floppy drives.

It was well worth the money to go to 640K (so I could play NetHack :slight_smile: ) and do some basic testing of my code before copying it to the college’s mainframe. I did upgrade to a Hercules monochrome card for reasons.

But I never got an 8087-2 for it because my work wouldn’t really use one. It didn’t get a hard drive until it got food-chained to my brother when I got my 386.

But I think that IsaacKuo has the right idea. A particular expansion is not really relevant since it’s so dependent on the person. But the ability to expand is very important.

It’s more important in an immature or speciality market, less important in a mature or commodity market. Maturity of a marketplace is more about the mindset of the customer base and less about the product offerings. In a mature marketplace the expectations of customers are well-established and understood (small cars, home appliances) so they can be targeted accurately in the base product, making expandability less important.

I think, it depends – on the era.
Generally, I think it’s a great idea and it worked well for a certain time, and probably best in the more professional segments (just think of the PDP-11 and VAX expansion boards). But I also think that it became more problematic by the late 1980s, as the general speed of chip innovation started to take off.

E.g., something like the Apple Jonathan project (1985) looks great on paper: just keep expanding and it also seems like a good insurance, in terms of future proofing, since you can exchange components as they age and become eventually obsolete.

However, in practice, it’s much more about bus standards, bandwidth and supported clock speeds, and the entire allotted assembly will become obsolete, anyway. (E.g., a Jonathan system designed with 8086 compatibility in mind, would have probably been ridiculously obsolete with the arrival of the i386, regardless of any modularity and expandability.) So, at about the mid 1990s, you were probably best off with a complete, generational machine.

(Personally, I had to change most of the peripheral equipment at least at every second iteration, as we went from SCSI to USB to Firewire to USB3 to USB C, etc. And it’s not just about the cables, which will end the quest for cables for all time, each time, it’s also about communication standards and drivers and/or supporting software, which renders the equipment unsupported, soon after the next standard arrives. The only somewhat future proof peripherals were those connecting via Ethernet. So Ethernet may be – quite unexpectedly – the most viable expansion standard.)


Is this what is referred to as “sideloading”?

Most companies had more or less peripherals. Maybe it was a point when selling, but the reality was then that one would rather buy a new next gen computer.
Of course different on the PC or rack computers.

Maybe it was a point when selling, but the reality was then that one would rather buy a new next gen computer.

There was an ACM article a long time ago that basically showed that the computer you need today would cost (for discussion purposes) $2000. The computer you wanted today cost $4000.

But by the time you needed the computer you wanted today, it would cost $2000 at that time.

So you should always buy the computer you needed when you needed it. Then you end up with computers you can “food chain” to other members of the family.

In this case, upgradability was not really relevant since you simply purchased a new computer every few years when you outgrew the one that you have.

1 Like

Yes and hopefully you got one as present.

Oops, I did say I’d add my own thoughts.

I think where I was initially coming from was the vague recollection that I might be in the habit of buying things which are upgradeable, but in practice never upgrading them.

But in reality, I did

  • buy a BBC Micro and later add a floppy disk interface
  • buy an Amiga 500 and later buy a trapdoor memory upgrade
  • buy an Amiga 1200 and later buy a trapdoor CPU upgrade
  • buy a no-name P75 PC and later add RAM and replace with a faster motherboard and CPU (I think… if not this one then a later one)

Acorn’s Electron is perhaps a good example of a cost-minimised product with expansions architected in: well-packaged expansions literally bolted to the back of the machine, where the edge connector has almost all the signals you might need. Lionel Smith’s Electron Pages shows third-party expansions:

A fully-expanded Electron is not very far short of a Beeb.

But if the main CPU bus is to be brought out and daisy-chained devices attached, one might think that would have some speed and reliability implications.


I do like the symmetric form factor of the Electron, and appreciate rearward expansion more than sidecars.

I admire the Electron, as well – but, where do you put the monitor?

(Was an “expanded Elk neck syndrome” a common complaint? :wink: )

Possibly on a stand, but most likely simply behind the computer.

This is less of a disadvantage compared to other systems than you think, because other systems annoyingly placed cables sticking out the back. So, unless you raised the monitor you still had to place the monitor pretty far back behind the computer to accommodate the cables.

I’m a bit skeptical about a stand, because of the top-loading expansion boxes – which really raised my concerns.
I guess, these may be the most serious real-world problems with expandability: desk space on the one hand (sideways) and ergonomics (backwards and stacking) on the other hand…

(The only concept, I know of, that would have solved this somewhat graciously, is the Mac Jonathan project, as long as the monitor size doesn’t exceed a certain limit.)

How many top loading expansion boxes would you have, though? You could make it the first in line.

In this photo, it’s 2 out of 3 – and the third one (front-most) seems to have a button on top…

(There isn’t that much space on the sides, back and front are already reserved for daisychaining, and any cables and connectors make side space even more scarce or difficult to access, as cables may spread from various boxes to both sides. Which really only leaves the top… So this could be a systematic issue.)

Use the TV on the other side of the room dummy, :slight_smile:
A monitor was another add on feature. Some people like myself
could only afford one piece at a time.
Ed had it right, Trapdoor (under the PC) upgrades. :slight_smile:

I really don’t think a typical user would be swapping RAM and ROM chips very often. So it could go at the back.

I don’t really see what the big deal is, anyway. The way the Electron is, you won’t have any cables sticking out the back. As such, you could simply pull the entire assembly toward you to get at the “top” of expansions.

You couldn’t do this with, say, a Commodore 64 because you’d likely pull out the video cable and/or serial cable, with no easy access to reach the unplugged cable(s) underneath the stand.

A lot of games came as a rom, so your game slot needs to handy.