Hacker News new | ask | show | jobs
by bmitc 881 days ago
That's true, but the tools work for their given purpose. If only software tools required basic setup and maintenance.
2 comments

Yeah, it’s not quite the same.

On the other hand the tools are mostly very simple and highly standardized. The simple tools are applied to the problem through a combination of applied experience and constructed jigs. I guess the equivalent in the software world would be classic non-fat Unix as your base tools, and building your own libraries and toolkits as your jigs (no or very limited third party deps), while aggressively avoiding complexity. You’ll get reliability this way; of course, there are serious tradeoffs.

Don’t misunderstand me, though. I agree that the modern software treadmill is horrible and needs to be tamed.

You know, I'd actually like to add to my earlier comment.

> Yeah, it’s not quite the same.

I actually don't know quite why I wrote what I did, because the concession really isn't true. I can only imagine we were discussing some romanticized view of the subject. My own experience, however, shows that reality is otherwise.

Cheap tools usually don't work well or last; you end up constantly complaining about them and constantly replacing them. This is true whether you're talking about power tools or simple hand tools. (How much do you pay for most of the software and libraries that you use?)

For the hand tools --- some things like hammers basically work, but other things, unexpectedly, do not. Handsaws and handplanes are two good examples of simple tools that are fundamentally broken from the manufacturer, and in many cases simply cannot be fixed due to design. Other tools can be salvaged, but don't work out of the box. You can shell out inordinate amounts of money to a company like Lie Nielson, or hope to get lucky on the purchase of an antique, but the typical products you might expect to work, don't.

For power tools, the cheaper tools might technically do a job, but they often function in ways which are detrimental to the work or to personal safety, and don't last long.

You might think the solution is to spend more money for quality tools, but this doesn't always work out either. I'm thinking of a DeWalt biscuit joiner and a Makita pin-nailer; both are well-respected brands for "real contractors" who rely on their tools and expect them to work. The Makita pin-nailer doesn't load its nails right, so I have to futz with it after every nail --- which kind of defeats the purpose of the tool (it didn't do this originally, but it didn't take much use before the problem developed). The biscuit joiner, despite being a well respected model, cuts slots that have so much slop in them that there's no point in using the tool at all.

Now lets talk about machines briefly. My radial arm saw (no longer a common tool, but once quite common both in homes and on jobsites) cannot be calibrated to proper specifications. One of the adjustments cannot be made to spec and it throws off several of the others. The result is a tool that, while useful, is only half-functional. I have theories, but don't know exactly what is causing the issue (I got it second-hand); I would need to completely tear it down and grok the entire design, then rebuild it, to properly fix the issue. By the way, the calibrations it does have need to be periodically checked and corrected (the RAS takes a lot of flak for this, but I find this is true of any machine with adjustments).

This is usually about the time where the internet gets up in arms and tells you to throw away the RAS and use a table saw, so lets talk about my table saw --- the fence can't be calibrated to lock parallel to the miter slots. I don't think the fence is broken by design, but I don't know exactly what's wrong. The calibration process is logical, it just inexplicably doesn't work. I suspect one of the fence rails may be slightly out of alignment. I could fix this by building my own high-quality fence system --- I have the skills --- but this is analogous to throwing out your third party library or tool and writing your own.

In all of the above cases concerning power tools, the answer is maintenance/repair, but in all cases (aside from building a new fence), it is not the work I want to be doing, nor is it work in which I am experienced, nor is the necessary work obvious through casual investigation.

Of course, I have plenty of tools that work great and never give me trouble, other than periodic maintenance, but it's the problem tools that stand out. Software is the same for me; most of it works fine, some of it breaks down due to bit rot (maintenance required), some of it is broken from the manufacturer (out-of-box repairs required), and some of it is just piss-poor in design (see hand tools above).

I still hate the treadmill just as much as you appear to, but romanticizing the life of a tradesman is probably a mistake. Entropy is everywhere, poor design and unfortunate occurrences are everywhere. I don't know if you have any personal insight into the life of a tradesman, but they deal with all kinds of BS, from all angles, that just isn't commonly discussed outside the field in question.

=== Addendum:

I keep mentioning software libraries, but those really aren't tools. They're more analogous to prefab construction components, so in that case we could talk about items such as pre-hung doors and prefab cabinetry that is supposed to save time and require less skill to install. That's great, until you discover that the item was assembled out of alignment at the factory, defeating the entire purpose. I could go on ad nauseam, but I imagine you probably get the point by now...

God, it's all just miserable and depressing when you really start digging into it, isn't it?