One machine can go pretty far if you build things properly

One machine can go pretty far if you build things properly

Okay, so yesterday I posted a picture of an old multi-line bulletin board system which had apparently taken over an entire bedroom with dozens of computers. I mentioned that I intended to come back to the topic to throw in my own two cents, and so here we go.

When you looked at that picture, which way did you react? Were you one of the people who looked at it and thought it was remarkable, and wished you could "drive" such a setup? All of those machines, all needing to be built out and set up, just so your stuff could scale.

Or, were you someone who found it horrific and wished there was a simpler way? Did you wonder if it would be possible to pack all of that into one or maybe two boxes, instead? Did you want a way to negate all of that complexity while still providing the same level of service?

I'll admit: when I first saw that picture in 1993, I was in awe of the scale. It had never occurred to me what kind of layout you would need to run some of these systems at that level of scale. Some small part of me probably wanted to at least experience that, but I never actually went that far.

Now, things are different. I look at that and just sigh at the lost opportunities. The actual amount of work being done on behalf of any of the users who are connected is tiny. If we ignore the fact that the actual BBS software and related accessories (door games...) forced them to run DOS and just focused on the other elements, it becomes possible to see an alternative.

If you wanted to solve for the base problem of having a bunch of people dialing into a machine that would send and receive files, read and write messages, and yes, even playing some games, that did not require a basement full of computers, not even in 1993. Universities, ISPs, and even a couple of forward-thinking high schools were doing exactly that at that time.

There are *large* parallels from those two approaches to what I've been seeing more of as the years have gone by. There are people who go into a situation assuming that they are going to want a whole cluster of machines to do their thing. It doesn't matter that what they are actually doing should be a tiny amount of complexity and it should all fit on one box... or maybe two for some redundancy. They've decided it will have a measurable "footprint" and by god, they're going to make it happen that way.

Meanwhile, someone else who just wants it done comes up with a better approach that actually cares about efficiency, fits it onto the one or two boxes mentioned before, stands it up, and goes on with life.

Part of this has to do with what kind of tools you choose to use. If you are sticking with the 2022 equivalent of MS-DOS and machines that do exactly one thing at a time, then yeah, you are going to have a whole fleet of systems all sitting there busy-waiting on something stupid.

Or, you can make better choices and just have a boring little setup that sits there and crunches through anything you might possibly throw at it. Oh, sure, eventually your business might get big enough to where you have to scale it up, but *IF* (not when) that happens, then you can worry about it.

Don't recreate the basement full of PCs when the problem can actually be solved with a single box sitting in a cabinet somewhere. Put away the urge to check off all of those boxes on your next job search by making your current employer pay for your wasteful shenanigans. Do the right thing and screw around with performance art on your own time.

Epilogue: in the hours after writing yesterday's post, I spotted a couple of things which seemed to hit this same point, and so I will share some links with you.

This is not to say I agree with everything in all of those posts, but I do like the way they remind people that one or two decent machines can probably do the job... assuming you don't just burn money by building things badly.

Finally, I'll also use this opportunity to again share the link to the COST paper.

I love this quote:

You can have a second computer once you've shown you know how to use the first one.

-- Paul Barham, quoted in the COST paper

After all, how can you have any pudding if you don't eat your meat?