I had a density thing but it took too much time to load once the book count got higher. Now I just do avg frequency per word instead which takes like 2 milliseconds to calculate.
Best thing I could think of for an actual “difficulty measurement” would be average characters per sentence, with kanji counting as 1.5-2 “characters.” That would give you a sense of how complex you can expect a book’s sentences to be. Someone could do it in javascript with wanakana + some basic logic.
I’m just super torn because I really do wanna read Haruhi. I’m going to try to muscle my way through it a bit longer before I decide.
prepares to commit seppuku
sees StarLi’s comment
puts sword down
Y-yes! I shall help! 責任を取る！
The book is definitely too hard for me to read right now but I do want to read it. Because I loved the anime when I first watched it years ago (in the airing order and not chronological order) and am curious about the source material. And because struggling with something hard makes you come out on the other side levelled up… But I absolutely wouldn’t be able to do it just by using this club to ask about stuff I don’t understand (because that would be practically every line). I’m getting the english version as well and using both. I’m conflicted about this method of study/reading… But it might be the only way for me to read material a little less juvenile than Zenitendou…
Just to warn people on the fence: the next week’s reading starts of with a whopper of a sentence, but then it gets easier again (I promise!) as we get some more dialogue.
But seriously though, what a sentence.
On the up side, these huge freaking sentences are good grammar review.
While I admire your optimism, I know for me, personally, I won’t improve that much in so short a time. Only a few days left in the week and I’m sitting at one and a half pages, only.
I’m curious to hear what everybody’s ideas for “other” was.
@Belerith - the idea was to decrease the workload…
I completely agree with this, and also wouldn’t read samples of books I’m not interested in. So, unfortunately, Naph, I don’t think that’s likely. It might be best if 1 or two people (with a lot of time) who are more advanced could actually speak to the difficulty/do the comparison, so at least it’s one person’s scale… Or conversely, Instead of a poll, have people chime in with what WK level and approximate grammar level they think they are, + “looks way too hard”, “looks hard”, “looks tough but manageable”, “about right” “fairly easy” “should be a picnic!” kind of answers. *shrugs*
It’s more subjective, but at the same time, gives people of a similar WK/grammar level a better idea of how it might fare for them. Also the idea of the FloFlo list could work too - but would be best if it could be done based on 1 specific WK level (say… 20? 35? Somewhere in that region?) and only count vocab that’s 5 or more times in the novel, and how many are new compared to the WK level). The stat could be helpful.
@ekadish - the airing (non-chronological) order is best order. ^^
Haha, didn’t mean to suggest we should increase the workload. Right now I’d vote to keep the 15 pages pace if only so it doesn’t drag out even longer than the 4 to 5 months scheduled already. (
)
But then I’ll be starting コンビニ next week, so I might change my opinion. And, you know, vote for decreased page count here. Maybe.
Sorry I wasn’t clear. 
Oh, it’s okay!
I thought you thought I wanted us to speed up and read more pages a week. ^^; Glad that while I misunderstood you, you hadn’t misunderstood me.
Oh, the joys of the internet…
The original airing was non-chronological order because if it were chronological, the “main storyline”, such as it were, would be done by episode six, leaving the remaining eight episodes as just filler.
My vote for “other” was the same reason as Ditto20: I had no specific preference if the group decides to slow or speed things up.
That’s probably not going to work. It’s really difficult to imagine what someone else will find difficult. For instance, I have no idea what JLPT level grammar points are. All I could say would be “this is fine for me”, which isn’t helping in any way.
That’s more like it, but that’s pretty much already what we are doing. It would be a logistical nightmare to have polls based on people’s level or to cross reference their vote and their level. That’s why the club itself is supposed to be indicative of the average level. We do also warn people to check the samples for themselves rather than blindly trust the difficulty rating.
I do like @seanblue and @Raionus solution, though. It’s fairly easy to do some stats on the sample pages. I haven’t played much with text parsers, but we may even be able to extract some grammatical constructions (those that are basically vocabulary) and cross reference them against a JLPT list. I don’t believe that it would be too informative (since I think a book aimed at natives would use grammar from all levels), but who knows.
Based on the current discussion, median sentence length +/- standard deviation could be informative as well.
Well, again, slowing down for a couple of weeks would not change the time spent at all. Slowing down for the next two weeks, for instance, makes us lose only 10 pages. It’s going to be annoying to re-establish all the break points, but anyway, on week 9, we had only ~11 pages planned, so we could bump that back up to 15. We also have only the epilogue (7 pages) on the last week, which we can bump instead to 12 (in the ideal case, it would take some really weird cuts to get to that). Et voila, no delay.
That sounds very reasonable. We do have a lot of time until having to make that decision though. I think we should consider it when we are about 6 weeks from finishing kombini (or whatever is standard for having the next poll, I forget). If it’s just me and @QuackingShoe still reading along, I guess no one will mind anyway. (I hope we’ll manage better than that though)
I’m doing it to the end too, even if I don’t like it (I already bought it, right?). So at least three people 
Still hoping there’s more people, though
I’ll be reading to the end as well. I may be too busy to comment much at times though.
I’ll be trying to read to the end as well because why not.
Also, because ending on the baseball episode would be a weird letdown… The airing order ramped up the action, and increased tension.
I intend to struggle as long as I have time to… So at least the first half. Might like my head in every now and then after/if I drop out due to classes.
Okay, so I made a small app to see if it addresses the problem. It measures average characters per sentence.
I tested a couple books. “Difficulty” scores are as follows (* marks children’s books):
22.01 Katakoi*
23.89 Kino no Tabi
24.38 Majo no Takkyubin*
26.35 Hyouka
28.43 Ryuugajou Nanana no Maizoukin
29.65 I Want to Eat Your Pancreas
29.70 Re: Zero
29.91 Haruhi
30.45 Shin Sekai Yori
32.03 Shin Sekai Yori 2
34.03 Ojousama Tantei Arisu 3*
The only worrying one is the last book but it’s also written with a ton of 敬語 (mc is a butler) so maybe that could account for the sentence length.
The scores are average sentence length where any character = 1 except for kanji which count as 2.
Ex: 私はアリス = 6 but わたしはありす = 7. There’s room for improvement but it seems like a decent test.
I googled around and found the term “readability score”. Specifically, I found this paper: Paper that describes a readability score developed specifically for learners of Japanese as a foreign language.
It has an online implementation here:
JReadability.net
Thanks for implementing that! It looks reasonable, based on my experience: I have read Kino, I want to eat your pancreas, the first ~30 pages of Re: zero, the first 15 pages of Haruhi and Shin Sekai Yori, and I would sort them in that order. However, I don’t think there’s such a gap between kino and I want to eat your pancreas, but that may be just me.
I plugged in some stuff from wikipedia in JReadability and it grossly underestimated the difficulty, since it only computes it based on word it knows… and apparently technical Japanese isn’t its forte. 
That being said, it should be more than adequate for novels, since the vocabulary is much more likely to be in their dictionary.
All in all, I guess either (both?) tools could be used for nominations and may be more informative that the difficulty poll.