Hacker News Reader

Comment by acqq | Hacker News Reader

I believe the logic behind HemingwayApp is misguided:

Hemingway the writer actually wrote long sentences and they were actually important in his writing.

Passive is also important in good writing.

You can't use machine metrics to force "good writing" you can only enforce mediocrity and the following some random rules "because the rules have to be followed."

Likewise, I as a writer of the software would absolutely hate to run some program to tell me "this function has more than 10 lines" or whatever. If I wrote 500 lines function it doesn't mean it shouldn't be that long: there are examples where exactly such functions are still necessary and good. Such automatic evaluations are for managers who probably don't understand what they enforce. Pointy-haired bosses, if you will.

So I see HemingwayApp as the pointy-haired-editor app.

(Edit: Improving the text based on the human input, thanks Agathos!)

View on HN

You might not be able to force people to write well, but you can help them not write badly. That's the point of apps like this one - mediocre is better than bad. Once someone has improved their writing to the point where this app is no longer useful they can easily switch to something else.

Writing is a skill that few people have really learnt to do well. For them, following some "random rules" will improve what they write.

Regarding writing 500 line functions in code: very occasionally you need that, but there's nothing wrong with an IDE that points it out and asks if it's really a good idea, and for new developers who would do that sort of thing all the damn time it'd be very useful in improving what they come up with.

Learn the rules first, then learn when they don't apply.

Forcing somebody to write short sentences wouldn't improve his writing.

Unless I missed something, this app doesn't force edits, it merely highlights things that writers should take a second look at.

Showing someone overly long and complex sentences, and pointing out examples of passive voice could only improve an editing process. Once you manually point out problems, the burden to change/leave as is falls onto the writer.

I find on average that it does, especially with less skilled writers.

Yeah it does. Long sentences are harder to read. Bad writers write too many long sentences, and it makes their writing difficult to parse. So just forcing them to write shorter sentences will improve their writing a great deal.

That said, a good writer will know when using a long sentence is the right choice for emphasis. And there's no way this program could ever develop that level of good judgement. This program isn't good for writers. It's good for non writers who want to quickly improve their shitty writing without effort.

You can't force someone to write well by any means.

Good writing is sensitive to multiple contexts and multiple rules. Essentially, one has to be able to shift-gears and notice the results of your writing, how your audience will receive it.

The only way a person becomes a good writer is through self-direct study. A bad writer who writes long sentences without noticing the results will remain a bad writer if someone forces them to divide their sentences, because they are still not being aware of the results of their writing.

All that said, a lot of people learning to write create long sentences because they don't yet have the skill to communicate with shorter sentences. Demanding that a person break up their sentences into smaller chunks can help them become more skillful - if they put in the effort to learn this.

> Bad writers write too many long sentences, and it makes their writing difficult to parse. So just forcing them to write shorter sentences will improve their writing a great deal.

See, I don't think the above is any better as two sentences, and I think it works better as one. Perhaps I just don't like starting sentences with "So".

You couldn't be more wrong.

I know people already pointed this out, but I'll add my 2 cents: Every single teacher I've had told me to shorten my sentences.

Take Twitter for example. It didn't take me long to realise just how fluffy and useless my writing was...

I think your point is that there are times when breaking the rules is a good thing. No one would argue with that. The problem is that most people have NO IDEA what the rules are for good writing. On top of that, it's easy to slip into overuse without realizing it. I'm a good writer and yet, one thing I'm terrible at is far too many adverbs. They aren't useful when overused. This app helps me see when I have relied too heavily on them.

Hemingway occasionally wrote long sentences and they proved more power because of their scarcity. The passive voice is occasionally important in good writing but bad writers use it too much. Most of the time they don't even realize it.

An app like this isn't about enforcing a standard rigidly, it's about showing an author the places in text that break general rules. It's making an author aware of the issue, not saying "Don't ever do this". In fact, the "rules" are defined as "Try to do X or fewer". The rules don't have to be followed but if you can't even recognize the rules, you have no idea when you should break them for effect.

From your example, somewhere along the way, someone taught you that long functions are GENERALLY bad (just like adverbs). That doesn't mean they are always bad (just like adverbs). This app doesn't provide an automatic condemnation of breaking the rules. It allows you to know what the rules are and how well you are following them.

Who says that the rules from the app are even right? Anybody acually eveluated them on real books by known good writers? I want statistics: how often good writers are outside promoted rules?

The rules in the app are Hemingway's rules from his book "On Writing". Lots of other authors have written similar books and generally they all agree on what works.

Henry Miller and Stephen King have both written books on the subject, and both called their books "On Writing". E.M. Forster wrote one called "Aspects of the Novel". King actually goes further, to the point of saying things like "Don't use adjectives." Extreme for most written work but if you want to get your audience to use their imagination less direction is often better.

I don't know anything about what Hemingway's rules are or what writers suggest in general. I can say, though, that during my AP English class, these are principles that we were taught in school for good writing. No doubt, you always take everything with a grain of salt so the writing doesn't have to be so rigid. But I find this app to be a great one, especially for people who want to improve and analyze their writing.

Have you ever read Tense Present by David Foster Wallace? If you haven't, find a copy online and enjoy! Based on your other comments, you'll appreciate DFW.

Short answer is that you should pardon my language, but English is fucked. So fucked that those of us native speakers should apologize for the state of our language. Not only is English constantly evolving (in different ways, in different English speaking countries), but the limited 'rules' are contradictory.

So, to answer your question about stats, we could take 10,000 English experts and lock them in a room until they write the English spec. In actual practice, they would die before they would ever agree on a set of rules. And then, even if they did come out with a spec, most famous writers would violate the spec at certain times.

I think you under-appreciate how simple English is. Sure, creating a set of rules for perfect writing is probably impossible, and in German you could probably fit them on a single page (I'm exaggerating, I don't speak German well enough to precisely judge that). But getting to a decent level of English is probably the easiest of all widely spoken natural languages. That is if we forget your atrocious non-existing spelling and pronunciation rules.

How often are bad writers following the rules? http://imgur.com/nXwdDC6

If you choose to slave yourself to the machine, you will produce output that reflects that. If you use the machine to improve your output, then you produce improved output with little effort. Either way, it is not the machine at fault. A frequent user would probably learn that some streaks here and there are OK, but if your entire text is colored blue that's likely to be a problem.

Personally what I need is an extension or scriptlet that embeds this into an arbitrary text box. I've struggled with adverbs in my writing. I am less worried about Hemingway about longer sentences or the use of a larger vocabulary since I'm generally writing for an audience other than the general public, though. Perhaps some tuning parameters would be nice.

Hmmm... thanks for commenting, acqq, but four out of nine sentences in your comment are hard to read. You also used four adverbs, try and aim for two or less.

Yo my man! Four legs good, two legs bad!

(I really, really hope you don't get down-voted from those who miss the point).

> try and aim for two or less.

I hope this was meant as provocation. I'm guessing you actually meant, "Try to aim for two or fewer," yes?

It was a reference to the Hemingway app. If you paste acqq's comment into Hemingway, it suggests "Four adverbs used. Try to aim for two or less."

As humour goes, it's a few levels of indirection away from Seinfeld. But you're on a forum full of people who spend all day thinking of abstractions for their abstractions, so what do you expect?

> If you paste acqq's comment into Hemingway, it suggests "Four adverbs used. Try to aim for two or less."

Okay, that made my day. :)

The passive voice has its place. But sentences in the form, "x is something that is y," are difficult to defend. Just write "Passive is important in good writing."

(Even worse: https://twitter.com/Graham_LRR/status/387778330054774784 )

The point is not whether passive voice has a place in good writing, it's to make young writers aware of the difference between active and passive voice, and when each is appropriate (or "strongest").

The advocation of a limit discourages that interpretation.

My point is that writing is like walking -- at first, you only have one way to walk, the one that keeps you from falling onto the carpet. Only later do you learn how to dance. It's the same with writing -- avoid obvious mistakes until you're skilled enough to use them for an intended effect.

That seems to be a different point or a muddled metaphor - people learning to walk are accidentally dancing (badly)?

I have not seen it established that passive constructions - where they are appropriate - are harder to use correctly or well. I have seen it asserted countless times, usually by people who don't know their passive from a hole in the ground and always without accompanying data.

You probably can use machine metrics to have a good overview of some properties of the text; one of the more useful machine metrics that comes to mind is vocabulary estimation.

I guess personally I'm a bit put off by the UI. I'm reminded of a scene in the recent movie Her, where the AI offered some suggestions but only after being explicitly asked to do so. In HemingwayApp, the immediate red/yellow highlight will nudge the user into fixing the paragraph, whereas a more subtle colour and/or style might allow him more space to think "the computer is wrong this time, I like it more the way it is".

But most people are lazy and will slip into writing poorly on a regular basis. Having an automated tool highlight potential problem spots can help you to write more consistently the way YOU want to write. If what is desired is passive voice then ignore the highlighted messaging.

note: most programming syntax style applications have the ability to selectively highlight special cases to flag sections of code to ignore particular errors.

Denotationally, there are never examples where such functions are necessary in most languages (with operational constraints and a sufficiently dumb compiler, you can make anything necessary). This doesn't mean that there are never examples where such functions are appropriate.

Your comment about machine metrics reminded me of corporate use of process.

Reply on news.ycombinator.com