It's an educational exercise to take something like Joel's essay, try to understand it in the way raganwald advocates here, and then go and edit it such that all the caveats, simplifications, etc are all explicitly enumerated. After you're done doubling the length, ask yourself if anybody would actually have read the end result.
Having gone through a phase where I tried to write like that, I can tell you two things: First, "no". Second, it's a complete waste of time. Add in the caveat that you don't literally mean that there are 0 software packages sold in that price range, and you will still get flamed for the claim you just disclaimed. A lot of people are sloppy readers and putting in that level of detail is like putting fine detail in your painting for a farsighted person without their glasses.
I've come to view this sort of misunderstanding as an unavoidable tax for trying to communicate with humans. I say "humans" because it's not "English's" fault. Natural language is fuzzy mostly because human ideation is fuzzy. Noticing when English creates fuzziness is a noteworthy event precisely because it is an anomaly in the vast sea of fundamentally fuzzy ideas being expressed in English.
This is why academic writing - even academic writing about cool technical topics - is often "dry." We have to write like that. And it is tiresome to read. But it's necessary for two reasons. First, it would never get past peer review otherwise. But even if we could eke it past, it's still not a good idea because when someone else goes to replicate or build on what we did, all those details that drag down the writing become mightily important.
I've often thought that people who publish papers should also have a blog, and where they should have a more digestible explanation of their work. Of course, I have not done this. (Yet.)
I remember reading an advice piece on how to read scientific papers. The core of it was that you need to read them breadth first (i.e. skim many times, each time trying to go deeper into the details.)
What you have written explains why it is such a good piece of advice - if you read a paer depth first you'll get bogged in the details without a chance to get the whole picture.
This is so true. And the comment about academics is well taken as well.
Having had many, many experiences on Usenet back in the day where I would post something, someone would pick out some quibbling points, and I'd respond with more clarity, and they would pick at that, and I'd respond again. Until we had like 4 or 6 postings which ended in "Oh, well if you put it like that I guess." and knowing I had basically said the same thing half a dozen times, lead me to the same conclusion as jerf here.
What ever you write, some fraction of the audience won't get it. And some fraction of that group may be inclined to respond. Outside of actual trolling, it is the nature of communication. Especially words only communication, that it is imperfect.
Once I realized that I tend to be more like raganwald in that I read something, and if it doesn't make sense to me I try to read it again and see if I can figure it out, like a puzzle.
I'm past the point to get annoyed with those sort of replies from anonymous Internet users; but I still do get annoyed at times by similar replies from smart people that I know.
I continually have to be on my guard against letting my smug, egocentric engineer (not saying all engineers are smug and egocentric, but the one in me sure is) take over. Raganwald is correct that as soon as he does, I learn nothing. I like learning too much to allow that.
It is also obnoxious to be in discussions with others who do that. Whether it's the mentioned "I stopped reading when" or the particularly pernicious "/s/your/you're/g" type of comments (quick, somebody point out if I was supposed to escape the apostrophe. I'm not a sed/vi wiz), it takes a discussion away from ideas and focuses them on something less.
Great minds discuss ideas, average minds discuss events, small minds discuss people.
I think whomever said this (apparently there is some amount of debate about whether it was Eleanor Roosevelt or not[1]) was right. When we start talking about sentence structure or apostrophes, you know we aren't talking about ideas.
I come to HN because there are a lot of really smart people here. I find it particularly grating, then, when a thread devolves (or worse, starts) below the level of ideas. I don't get to learn anything then.
There's plenty of software that falls into the 1k to 75k range. Off the top of my head: any professional embedded C compiler or IDE license, MATLAB, MathCad, Adobe CS Suite, various CAD packages. These are all used in engineering environments where maybe a handful of licenses are purchased, at least in the environment I'm currently working in. We don't get salesmen calling us and pushing new releases; we maybe purchase every second or third new version.
Joel writes "You need purchasing managers and CEO approval and competitive bids and paperwork."
I think this is a bit over-reaching. Sure, you need manager approval, but that's been pretty easy to obtain in any engineering environment I've been in. Most of the time the software vendors come up with a new file format that's incompatible with the past, so you're practically forced to upgrade. Never have I seen bids or CEO/CTO approval on the software I'm discussing.
I think perhaps a more relevant range is 5k-75k or even 10k-75k.
(Sorry if I'm doing it wrong, this is my first post here.)
> You don’t want to go through life being one of those smug Internet people whose sole contribution ...
smug people on the internet? frankly, i don't appreciate his overgeneralizing mischaracteristizations of people who use the internet. it's 2012! everyone and their mom uses the internet! phrases like "Internet people" make it seem like the only people on the internet are defensive, awkward neckbeards who can't relate to regular people.
honestly, things like this call into question his entire argument. if this is what he thinks about people on the internet, he can't possibly know what he's talking about (which, is presumably, colors or something)
he can't possibly know what he's talking about (which, is presumably, colors or something)
The “Light” theory of color says that Black is not a color. The “Pigment” theory of color says that White is not a color. Therefore, the author really doesn’t know what he’s talking about!
overgeneralizing mischaracteristizations
of people who use the internet
You could read it like that, but I garnered a different understanding. I read it as there are some people on the Internet whom are categorized as smug, not that everyone on the Internet is smug. We've all encountered them before surely. They tend to jump to conclusions and rush to dismiss others based on the tiniest perceived slights.
Reminds me of [Poe's Law](http://rationalwiki.org/wiki/Poes_Law) which states "it is impossible to create a parody of Fundamentalism that SOMEONE won't mistake for the real thing".
Though I think it should generalise to "there exists no parody which can unambiguously be identified as such". Doesn't that make parody's great?
Just as important as trying not to be one of those people who can't see the forest for the trees, so to speak, is trying not to care too much about them. Because there's always going to be someone around to remark on spelling mistakes or taking hyperbole literally.
It's just inevitable, and to me the best response has always seemed to be to quietly mock those people behind their backs. ;)
As a programmer, I see this phenomenon as related to bikeshedding [1]. Instead of taking in and processing a complicated and interconnected argument, the gut reaction that people have is to find the smallest and simplest mistake they can and attack that. In a similar way, I try not to get too worked up about those reactions either. It's just human nature.
Before you disagree with someone on something, you need to understand it. Part of this process is discovering and understanding what is implied, not just what is explicitly stated.
An the primary problem an author has is the need to explain to the reader. Often a good way to do this is to make statements, such as those given as examples in the article. Fully qualifying every statement can get in the way of the explanation.
Of secondary importance is the need to actually prove it to the reader. Writing a proof is hard, long, and often unnecessary.
How do we resolve this situation as rational readers? When you notice a "problem" in something, you must state why that problem is really relevant to the argument at hand. It may not be relevant to the argument being made.
One of the problems here is that in order to understand the statements relevance to the argument, you need to actually read the argument. Stating "I stopped reading at x", it is likely that you do not understand the argument at hand, since you do not know what it says.
Of course, everyone has limited time, and certain cues in the body of some writing can show if it is a waste of our time. However, to say "I stopped reading at x", and then to spend 30 minutes composing a comment on HN is engaging in intellectual dandyism, not improving the world through understanding.
I'm grateful this exists, so that I can link to it.
When you try to point out a trend and give five or six examples, and people nitpick one but ignore the general point... I don't know, it seems almost perverse.
The magic of Hacker News (or Reddit, a very long time ago) was that it didn't have much of that. It's changing. Still the best public discussion on the internet, but pedanticism is on the rise.
HN used to feel like having a coffee and snacks in the living room of a very gracious host who created an eclectic mix of people -- just to get a good discussion going. I liked contributing, and though I'd at times make poorly thought arguments (I hadn't studied enough history to see the politics I advocated weren't quite right), I always felt like I had a shot to engage someone in discussion.
Now, the place feels like you've got to be aware of rhetoric in front of a potentially hostile crowd, fighting for the payoff of... the crowd's fleeting ephemeral approval. Which is one reason why I don't comment much here any more.
I stopped reading when he discussed Joel's software pricing model, and then proceeded to make a crack about how Joel's model was more correct than "some mythical supply and demand model."
If you don't know enough about economics to recognize that Joel's pricing model is a supply and demand model, you probably shouldn't try writing about the subject.
(Disclaimer: Yes, of course this post is tongue in cheek. If you had any doubts, you may need to get your sarcasm detector calibrated.)
Most products showed as examples in the previous story have a very consistent profile: entrenched professional packaged software, basic for the operations of technical departments in medium to big companies: MS Tools and Office, Autocad, Photoshop... more than exceptions, it seems as a category in itself.
> You don’t want to go through life being one of those smug Internet people whose sole contribution to every discussion thread is “I stopped reading after…”
This is my biggest pet peeve on HN, and usually leads to an immediate downvote. If you stopped reading, you're not qualified to comment on the post, so don't.
I come from an industry (AEC) where software tools all (in the reganwald sense) fall into that gap and somebody (again in the reganwald sense) uses them.
Why?
Because the industry leader, Autodesk, bought into $75,000 a pop software concepts from day one in the early days of PC's. They had a mini-computer model rather than a scaled up Apple II approach. Timing helped as well, early versions at $2000 a copy were not driving the total cost when the hardware was three times as much.
Now their products and those of their serious competitors all fit in the gap between roughly $2500 and $6000 a license. Those price points are pretty common among scheduling and project management software for construction contractors as well.
Let’s have a look at Mr. Braithwaite’s résumé: Hmm, he won a Jolt Programmer Productivity Award for leading the JProbe team. Fine, let’s see how much JProbe costs: $3,000 for the multi-OS concurrent version.
Haha, he ought to know better before quoting Mr. Spolsky!
The question is how much does the price of the software drive that purchasing process?
Autodesk was selling $10k solutions in an industry where the purchasing process was largely built to buy a dozen concrete blocks, a hollow core door, a tee square, a lamp and a stool (if the blocks and door could not be acquired gratis from a friendly contractor) and where $600 for a drafting table and $100 for a Mayline was seen as outrageously extravagant.
Their pricing model created a whole new tier of purchasing within the industry - in 1980 no small firm had a recurring five figure equipment expense. Autodesk's pricing changed the capital requirements of the industry. As they have shifted to a subscription model over the past twelve years, they have further changed the financial considerations within it.
I perhaps used "tool" too loosely seeing the comparison it brought. Perhaps, I ought to know better.
I think there is a point here. If your competitor software costs $X, and you offer similar features for $Y, where $Y < $X and $Y is within this range, it could sell anyway.
To throw a point out there, fogbugz/Jira quickly starts to "grow" into the 1000+ range once it has over 20+ users.
The key though is that they usually start as something used by a small group of folks and then grows as more folks realize it has some niceities over the open-source solutions. It also means however that Fog Creek/Atlassian do have a sales force that will call you over the phone, personally answer your questions, and sometimes negotiate on price.
Having gone through a phase where I tried to write like that, I can tell you two things: First, "no". Second, it's a complete waste of time. Add in the caveat that you don't literally mean that there are 0 software packages sold in that price range, and you will still get flamed for the claim you just disclaimed. A lot of people are sloppy readers and putting in that level of detail is like putting fine detail in your painting for a farsighted person without their glasses.
I've come to view this sort of misunderstanding as an unavoidable tax for trying to communicate with humans. I say "humans" because it's not "English's" fault. Natural language is fuzzy mostly because human ideation is fuzzy. Noticing when English creates fuzziness is a noteworthy event precisely because it is an anomaly in the vast sea of fundamentally fuzzy ideas being expressed in English.