The Tool Rant

This is a Work in progress

by Perette Barella

Working in the computing industry for a while now, I've found that companies are often unwilling to invest in proper tools for their employees. Under the guise of "cost cutting", companies make more work for their employees by refusing to give them the proper tools. I find this incredibly frustrating, as I am very willing to buy tools when I need them. In this document, I compare and contrast my purchasing of tools with that of the companies I have worked for. I will draw parallels between the tools I have in my basement, and the tools I should be given at work. I will show that not only are these cost methods ineffective, but that they usually waste money and result in lower quality products.

The right tool makes a job easy.

Using the wrong tool makes a job more difficult, and wastes money. You could frame a building using a 5-pound hammer, but I certainly wouldn't want to. The amount of time saved by investing in a good framing hammer would easily pay for itself, and would make the job easier.

In the software development world, companies often provide workers with a cheap solution rather than the proper one. I had the joy of trying to develop under Windows 95 for 2 months. It sucked ass. When there was a bug, it usually took Windows down. The debugger rarely worked, and did not prevent the system crashes. Windows 95 is not the proper tool, and the amount of my time I wasted could have easily paid for a Windows NT license. After 2 months we moved to NT, thus spending the money anyway. And, had I been working on Windows NT from the beginning, I'd have been much less frustrated for 2 months of my life. Had the company not moved to NT, they would have continued to pay for lost time lost for crashes, as well as reduced productivity due to frequent interruption of my thought sequence.

The wrong tool reduces the quality of work. You can put a connector on the end of RG-6 (cable TV wire) without the proper stripper and crimper. It can be done. It usually results in a bad connection (if any) with the shielding. The connector usually falls off easily, or is crimped on in a mangled way such that you cut your fingers if you aren't careful with it it.

Using the proper stripper allows you to remove the insulation with minimal damage to the shielding. With a razor blade or regular strippers, you usually cut much of the braided shielding off with the outer insulation. The proper crimper (the $20 one, not the $3 one) crimps the whole connector at once, nice and evenly without a feat of strength. The crappy crimper beats vice grips, in that it doesn't crush the connector flat and leave sharp edges, but it requires a massive effort to crimp anything because of the low MA (mechanical advantage). It just isn't as good as the Right Tool.

At Nortel Networks, we used a page layout program, Framemaker, for our design documents. Adding a graphic to a document was simple: I just placed the cursor at the point the graphic was desired, and selected "insert anchored frame." Recently, I've tried to use Microsoft Word for my design documents. It is not adequate for my needs. If I want to generate a graphic to indicate how various components will fit together, there is no reasonable way to do it. The graphic tools are limited, and there is no way to create a region in a document that is reserved for a graphic. Instead, the graphics mysteriously float above the text, without forcing the text out of the way. Using a generic word processor instead of a page layout program results in my design documents not having any graphics, making them more difficult to understand and, in my opinion, simply inadequate.

A cheap, low-quality version of a tool is often inadequate, resulting in a lower quality product and wasted money. I recently had my bike chain break, so I used my chain tool to fix it. Within a few weeks, it broke again. Concluding that the chain was at fault, I replaced the chain. After installing the new chain, I inspected it and determined that one of the links had been mangled. I determined that the reason it was mangled was that I have the cheap chain tool, which isn't made out of a nice, hard metal. Therefore, after a few uses parts of it had been bent, resulting in the tool mangling a link and weakening the chain. The cheap tool wasted my time and money.

One of my friends is a system administrator. Back in January, the company he worked for went against his recommendations and bought the cheap router. The cheap one had reduced configurability, which impacted its ability to act as a firewall. It took 2 weeks to set it up as good as it got, and it still didn't do firewalling properly. It isn't expandable, so when capacity is reached, it needs to be replaced rather than have a new card added. The company could have just bought the better router, which would have paid for itself in reduced time to configure, and it would have saved money in the future by allowing upgrades rather than requiring replacement. Lastly, if my friend didn't keep getting screwed around with the cheap equipment that made his life difficult, he might not work there only in the past tense.

Doing a low-quality job comes back to bite you in the ass. A week later, my bike chain story got even better. I figured the inadequate job done by the cheap chain tool would be okay until I could buy a good one, and worst case I'd just have to reattach my chain. Therefore, I didn't make going back and doing it right a priority. Well, that came back to bite me in the ass. The poor link partially failed, jamming the rear deraileur and bending it toward the tire. The deraileur caught in the spokes, bending the deraileur and the point at which it attaches to the frame. I knew I should have fixed the poor craftmanship I had done with the cheap chain tool, but since it "worked", I didn't get to it fast enough, resulting in a much larger problem than I had in the first place.

In the software development world, there is commonly a delusion that a development team can build a "quick hack" to statisfy customer needs, and then go back and do a proper, well-designed, clean implementation. In reality, this never happens except as a result of a problem, and even then it usually doesn't happen. People might wish it did, but, when you think about it, it can't. The reason for building the quick hack in the first place is that a deadline needs to be met. If there isn't enough time to do it right in the first place, how can there be time to do a quick hack as well as the real version? In reality, if the quick hack works, then it will be deemed to be adequate after it is completed because another new piece of development will take prescedence. The only way a quick hack will be revisited is if it fails, after it has already had a chance to show the customer the lack of quality of the software. Furthermore, there is a good chance that, rather than fixing it right, the quick hack will get a quick fix. After all, the code is written and just needs to be debugged.

One tool does not fit all scenarios, and trying to make a tool fit all scenarios reduces quality. Vice grips are a great generic tool, and very handy. They can hold just about any object. However, they tend to destroy the ends on hex bolts. While a set of vice grips should be in everyone's tool kit, they are not a replacement for a good set of open-ended spanners.

It is always worth investing in a tool. A few years back, I wanted to get the freewheel (the gear cluster on the rear wheel) off my bike. I bought a freewheel socket from a bike shop, and tried to pull it off with a 8" rachet. No luck. I tried a 10". No luck. Finally, I took my wheel to an automotive shop, and told them that I would buy whatever wrench was necessary for the job. They brought out a little 10", and I told them it was inadequate. So, they went back and brought out the 25" chrome-plated 1" socket driver, which I have since named my BFW. They got out a 1" to 1/2" adapter. After a bit of muscling, the two of us got my freewheel loose. I bought the $50 worth of wrench and adapters.

In the 2 years I've had it, I've used the BFW a handful of times. Each time, the per-use cost of that wrench drops. I figure it's down to about $12 now. At that rate, in 2 years it will cost $6 per use. In 4 years, $4. Given that it would be at least a half hour of time to take my wheel down to the bike shop, and a couple of bucks to have them pull the freewheel off, sooner or later the tool will pay for itself.

Update: I got to use it again, so it's down to $10 per-use now.

In some cases, different people need different tools. My dad is a taper. He's the guy who comes in and hangs dry wall on wall studs, then puts drywall compound over the cracks and in the corners so the walls seem nice and smooth. He uses a 10" drywall knife and trowel. While doing some home repairs, I've tried using similar tools but found that they just don't work for me. They are too wide, and I can't manage them well. Instead, I like to use a small putty knife and a 5" painting knife. I work much more quickly with the smaller tools, and the quality of my work is better than when trying to use the larger ones. As an added benefit, my wrists don't get as sore using the smaller tools.

Some people like linux, some people like windoze.