MSPA Forums
Page 7 of 7 FirstFirst ... 4567
Results 151 to 171 of 171

Thread: Measuring Homestuck (Data! Charts! Graphs!)

  1. #151
    Has all the et ceteras. Etc. psychoticBark's Avatar
    Join Date
    Mar 2011
    Location
    Edinburgh, UK
    Posts
    230

    Re: Measuring Homestuck (Data! Charts! Graphs!)

    Quote Originally Posted by Fish View Post
    No passive voice (even if it is the most accurate, it just adds hollow words)
    Heh, my writing style sucks. Sounds like a fine rule. (Unless it puts people off - any writing is better than none!)
    Recap a location if we last saw it more than a certain number of pages ago
    Good point - yes, I already do something like this for some of the word count stuff: ignore recently repeated words, but count them again if we return to a scene and the words help to re-establish the narrative context. So sure, recap just like you would if it was a purely textual narrative - that's what we're trying to mimic here.

    I'll bash through some more images myself this week. Fish, Flutterknight, anyone else - descriptions of the same form you already did will suit my purpose fine - so more of them would be most welcome.
    Last edited by psychoticBark; 02-11-2013 at 07:52 PM.
    Have an offline friend to recruit? A complete archive of Homestuck (and the rest of MSPA.) | There is a (MATH) % chance I've calculated MSPA stats. | oh my god so many words - transcripts of everything. | All of the words, etc. - extended search.

  2. #152
    rude duodeo Fish's Avatar
    Join Date
    Sep 2011
    Location
    a really small island
    Pronouns
    [other] (see signature)
    Posts
    4,603

    Re: Measuring Homestuck (Data! Charts! Graphs!)

    I had a couple hours to kill w/o internet today sooooo
    yes it is actually ridiculously easy to lapse into passive voice for descriptions; I dunno, as I keep going, maybe it makes more sense than talking about what "the screen shows ___?" Do you think what, specifically, the screen is zooming in on is important?

    also you are def right about narrative style (I need to stifle this alien/robot voice :0)

    e: I'll get spelling errors in the morning
    Last edited by Fish; 02-12-2013 at 01:03 AM.

  3. #153
    Seer of Doom Firecrow's Avatar
    Join Date
    Nov 2012
    Location
    Land of Snow and Apologies
    Posts
    72

    Re: Measuring Homestuck (Data! Charts! Graphs!)

    I don't think this has been done yet, so here's my contribution to the thread:

    Flash chance graph; what percentage of pauses X days long or longer have ended in a flash* or major HTML** update.



    *any SWF
    **specifically Openbound unless other similar updates happen
    Last edited by Firecrow; 02-15-2013 at 04:06 AM.
    Disappointed in poster's lack of clowns.

    Also, Spectralysis, a daring forum adventure of aliens, and surely some non-SBURB related world destroying as well? It is pretty awful actually ok, and stay out of my signature, dude!

  4. #154
    rude duodeo Fish's Avatar
    Join Date
    Sep 2011
    Location
    a really small island
    Pronouns
    [other] (see signature)
    Posts
    4,603

    Re: Measuring Homestuck (Data! Charts! Graphs!)

    oh man that's

    HELLA SWEET

  5. #155
    Has all the et ceteras. Etc. psychoticBark's Avatar
    Join Date
    Mar 2011
    Location
    Edinburgh, UK
    Posts
    230

    Re: Measuring Homestuck (Data! Charts! Graphs!)

    Quote Originally Posted by Firecrow View Post
    what percentage of pauses X days long or longer have ended in a flash.
    Great chart! I assume you've used some editorial judgement here since technically there was a 14 day hiatus (i.e. run of whole Hussie days without updates) between "AH: Close tome." (HSA5A2 6/005997, August 22 2011, 20:41 EDT) and "==>" (HSA5A2 6/005998, September 06 2011, 02:40 EDT) - but he was working on Cascade during that time, he just put out a few pages before starting another huge 49 day hiatus preceding it being finished.

    [Edit: oh, unless you're counting every SWF as a "Flash" - seem to remember that's what notEgbert did. I guess 6/005999 and 6/006000 are SWFs. Most people mean an "[S]" when they talk about "Flashes", though.]

    Fish, thanks for the descriptions: I'll start churning numbers this week end. I wouldn't worry about spelling mistakes, they won't affect the word counts, which are my driving metric. Also if you run into any images with lots of text, remember that it is already transcribed - so a more abstract "X talks to Y" description not mentioning exactly what was said will do fine there.
    Last edited by psychoticBark; 02-14-2013 at 02:39 PM.
    Have an offline friend to recruit? A complete archive of Homestuck (and the rest of MSPA.) | There is a (MATH) % chance I've calculated MSPA stats. | oh my god so many words - transcripts of everything. | All of the words, etc. - extended search.

  6. #156
    Prince of Doom Flutterknight's Avatar
    Join Date
    Aug 2012
    Posts
    244

    Re: Measuring Homestuck (Data! Charts! Graphs!)

    After I get out of school and stuff today I should have time to crank out more pages for your evaluation too.


  7. #157
    rude duodeo Fish's Avatar
    Join Date
    Sep 2011
    Location
    a really small island
    Pronouns
    [other] (see signature)
    Posts
    4,603

    Re: Measuring Homestuck (Data! Charts! Graphs!)

    Quote Originally Posted by psychoticBark View Post
    Also if you run into any images with lots of text, remember that it is already transcribed - so a more abstract "X talks to Y" description not mentioning exactly what was said will do fine there.
    Ahhh yeah I wasn't sure if you had everything from the images or just flash stuff. Noted!

  8. #158
    Has all the et ceteras. Etc. psychoticBark's Avatar
    Join Date
    Mar 2011
    Location
    Edinburgh, UK
    Posts
    230

    Re: Measuring Homestuck (Data! Charts! Graphs!)

    Quote Originally Posted by Flutterknight View Post
    I should have time to crank out more pages for your evaluation too.
    Excellent, thanks.

    In combination with having word counts for example image descriptions, it seems it will be important to know the relative proportion of images that show a whole new scene, or new props/complex actions that beef up the narrative and so need a longer description. So I built a little extension to my local archive that would let me quickly type notes against the page I was viewing and put them into browser localStorage for later analysis. Then I zipped through Act1 making such notes re the approx 250 images it contains: I was just typing 'S' for a new scene (about 30 images), and 'P' for new/busy props (about 60 images); only took fifteen minutes or so. I plan to at least sample other parts of the story too.
    Have an offline friend to recruit? A complete archive of Homestuck (and the rest of MSPA.) | There is a (MATH) % chance I've calculated MSPA stats. | oh my god so many words - transcripts of everything. | All of the words, etc. - extended search.

  9. #159
    Prince of Doom Flutterknight's Avatar
    Join Date
    Aug 2012
    Posts
    244

    Re: Measuring Homestuck (Data! Charts! Graphs!)



    I wasn't entirely sure what to do with a two-picture page (though it's probably already come up), so I included descriptions of both just in case. Same for the one Scratch section page I had which had an important banner.


  10. #160
    Seer of Doom Firecrow's Avatar
    Join Date
    Nov 2012
    Location
    Land of Snow and Apologies
    Posts
    72

    Re: Measuring Homestuck (Data! Charts! Graphs!)

    Quote Originally Posted by psychoticBark View Post
    Great chart! I assume you've used some editorial judgement here since technically there was a 14 day hiatus (i.e. run of whole Hussie days without updates) between "AH: Close tome." (HSA5A2 6/005997, August 22 2011, 20:41 EDT) and "==>" (HSA5A2 6/005998, September 06 2011, 02:40 EDT) - but he was working on Cascade during that time, he just put out a few pages before starting another huge 49 day hiatus preceding it being finished.

    [Edit: oh, unless you're counting every SWF as a "Flash" - seem to remember that's what notEgbert did. I guess 6/005999 and 6/006000 are SWFs. Most people mean an "[S]" when they talk about "Flashes", though.]
    Yeah, should have been more specific. By "Flash", I'm referring to every SWF, plus Openbound. I figured including the minor flashes might help balance out when it's been updated before midnight, with a flash after. (Not sure how to figure out time of posting; I believe it was mentioned earlier in the thread, but even if I used that, it would just add a bunch more difficulties determining how close two updates have to be to count as one. Just going by day of posting gives me a standard to go by so I can calculate this without wasting several hours or having to write a program.)
    Disappointed in poster's lack of clowns.

    Also, Spectralysis, a daring forum adventure of aliens, and surely some non-SBURB related world destroying as well? It is pretty awful actually ok, and stay out of my signature, dude!

  11. #161
    Has all the et ceteras. Etc. psychoticBark's Avatar
    Join Date
    Mar 2011
    Location
    Edinburgh, UK
    Posts
    230

    Re: Measuring Homestuck (Data! Charts! Graphs!)

    (So I'm glad the forum didn't disappear into the ether forever and take those picture descriptions with it. Was getting slightly worried.)

    By the way, here's a pretty cool site filled with visualizations based on stats derived about Lord of the Rings and related works: http://lotrproject.com/
    Have an offline friend to recruit? A complete archive of Homestuck (and the rest of MSPA.) | There is a (MATH) % chance I've calculated MSPA stats. | oh my god so many words - transcripts of everything. | All of the words, etc. - extended search.

  12. #162
    Knight of Pulse craZboy557's Avatar
    Join Date
    Mar 2013
    Location
    Land of Slavers and Rampancy (playing through Marathon)
    Pronouns
    he/him/his
    Posts
    780

    Re: Measuring Homestuck (Data! Charts! Graphs!)

    Will you ever be updating the original graphs? I want to see if homestuck is bigger than the bible yet.

  13. #163
    Knight of Light amiableTemplar's Avatar
    Join Date
    Aug 2012
    Location
    Land of Stone and Word
    Pronouns
    he/him/his
    Posts
    882

    Re: Measuring Homestuck (Data! Charts! Graphs!)

    Quote Originally Posted by craZboy557 View Post
    Will you ever be updating the original graphs? I want to see if homestuck is bigger than the bible yet.
    I'm not sure it's possible to compare Homestuck and the Bible anymore. Homestuck is so internally self-referencing and non-linear. Furthermore, it's stuffed with images, has lots of music, and (worse yet) multiple interactive components. All of these massively pad out the only reliable means by which we can compare the Bible and Homestuck: data size.

    We can't use raw word count because (a) Homestuck is coded in many places, which presents a thorny problem, (b) so much of Homestuck is non-word components (music/interaction/etc) that direct-word comparisons will leave out a huge chunk of the comic's meat, and (c) there are a number of recaps, outright repetitions, and self-reinserting loops (such as when a conversation gets seen from both sides, separated by stretches of real-world time between updates).

    Basically, Homestuck is a monolithic multimedia work, something that is literally only possible now that we have the Internet; it's one of the first great works of hypertext, rather than just text.

  14. #164
    Knight of Pulse craZboy557's Avatar
    Join Date
    Mar 2013
    Location
    Land of Slavers and Rampancy (playing through Marathon)
    Pronouns
    he/him/his
    Posts
    780

    Re: Measuring Homestuck (Data! Charts! Graphs!)

    Didn't you see the graphs? They may not be 100% accurate but the dude thought of some equations to estimate how many words images and flashes are worth.

  15. #165
    Has all the et ceteras. Etc. psychoticBark's Avatar
    Join Date
    Mar 2011
    Location
    Edinburgh, UK
    Posts
    230

    Re: Measuring Homestuck (Data! Charts! Graphs!)

    notEgbert who compiled the original report and came up with the "adjusted word count" figure he used in those comparisons hasn't been around this forum for a while.

    But I'm publishing MSPA stats on an ongoing basis (tracking as Hussie updates), and I do calculate his "adjusted word count" figure. http://readmspa.org/stats/#size

    If you trust the equations for adjustment (and I agree that they are far from perfect) we're way bigger than the 773472 word King James Bible by now (even the non-adjusted word count of Homestuck is 641594 today) and nipping at the heels of the entire Harry Potter series.


    Edit: digging into a few particular reasonable concerns:

    Quote Originally Posted by amiableTemplar View Post
    (a) Homestuck is coded in many places, which presents a thorny problem
    You're talking about non-plain text words, yes? I think this is solved. Not intelligently, but by brute force: I transcribed all the words from all the non-text media.

    (b) so much of Homestuck is non-word components (music/interaction/etc) that direct-word comparisons will leave out a huge chunk of the comic's meat
    This does indeed make any comparison ultimately very arbitrary. notEgbert's formula is a fair try, but yes, we're comparing the sizes of chalk and cheese here. Just before the forums took a holiday I did begin a little project to ground the adjustments in better data: a few of us have been writing text descriptions of the fresh narrative non-word content of images in Homestuck to get a handle of how many words a picture is really worth. See some of the messages above. Want to help?

    (c) there are a number of recaps, outright repetitions, and self-reinserting loops (such as when a conversation gets seen from both sides, separated by stretches of real-world time between updates).
    This one turns out to be not quite so much of a problem. When I went over the story, I marked the text content of many images and a few sections of prose as simple "repeats", and excluded them from my counts.

    Now, sure, this was yet another subjective editorial decision. For example, I kept the recaps (regular prose works can have them too!) and I did double-count the troll romance exposition (because it's part of the art.)

    But, turns out those micro-decisions don't sway the count very much. My pure (non-adjusted) word count taking all that into account is 641594 today; whereas a dumb word count of my extended search page (which contains every repeat, walkthrough instructions, and overcounts various other stuff too) is 678452 words. So the uncertainty is pretty bounded. Like I say, the "adjusted" (b) stuff matters much more.
    Last edited by psychoticBark; 03-17-2013 at 01:50 PM. Reason: more thoughts on amiableTemplar's issues
    Have an offline friend to recruit? A complete archive of Homestuck (and the rest of MSPA.) | There is a (MATH) % chance I've calculated MSPA stats. | oh my god so many words - transcripts of everything. | All of the words, etc. - extended search.

  16. #166
    Knight of Pulse craZboy557's Avatar
    Join Date
    Mar 2013
    Location
    Land of Slavers and Rampancy (playing through Marathon)
    Pronouns
    he/him/his
    Posts
    780

    Re: Measuring Homestuck (Data! Charts! Graphs!)

    We're coming for you Roling!

  17. #167
    Niklink's Avatar
    Join Date
    May 2011
    Location
    Venezuela.
    Posts
    1,124

    Re: Measuring Homestuck (Data! Charts! Graphs!)

    Great news all!

    As of MSPA page 8012, Act 6 of Homestuck has broken the 2,000 page barrier! Since there is almost basically no way that Act 6 Act 6 isn't going to be at least as long as Act 4, it means that Hussie's insistence of Act 6 being shorter than Act 5 is LIES LIES LIES!

    Still waiting on a new version of the original report! This thread is awesome, don't let it die out!
    "What? Funning? I never fun! I do not fun, and have no history of funning, and even if I were inclined to funning, I would not dream of funning with you."


  18. #168
    Time for nostalgia Blaperile's Avatar
    Join Date
    Dec 2010
    Location
    Antwerp, Belgium
    Pronouns
    he/him/his
    Posts
    8,509

    Re: Measuring Homestuck (Data! Charts! Graphs!)

    To bump this thread for great justice, I just found some interesting statics about MSPA that could be interesting.

    Wolframalpha knows how many daily page views there are, and how many daily visitors there approximately are.

    This is its current information:

    daily page views | ~~ 3.3 million
    daily visitors | ~~ 130000
    views per visitor | 24.6
    site rank | ~~ 10137th
    traffic fraction | 1 in 1.1 millions of all web traffic

    Check it out over here for yourself.
    A Member of the Midnight Crew

  19. #169

    Re: Measuring Homestuck (Data! Charts! Graphs!)

    Finally got caught up on HS after letting it drop off for a while, have updated the flash list on page 4. Decided not to follow the DOTA rule for the flash where Dirk engages trickster mode, because it involves so many hits of the right-arrow key that my own button-hit rate varied as it went along; I timed 3:10 for me to finish it, but your mileage may vary. Probably need to go back and re-time the original "worldbuilding" flash, but I wasn't sure when it started a new loop the first time through.
    Hobbestuck: You'll never look at Calvin and Hobbes the same way again.
    Website: http://www.morganwick.com. Tweeter: http://twitter.com/morganwick.

  20. #170
    Mage of Heart NorthernDruid's Avatar
    Join Date
    Mar 2013
    Location
    Land of Reflections and Sorrow
    Posts
    59

    Re: Measuring Homestuck (Data! Charts! Graphs!)

    Quote Originally Posted by NotEgbert View Post
    Masters Thesis? No, I already have one of those, and it's enough work as is.

    I agree that cascade is overestimated by the current metric, and would probably aim for something closer to 1500-2500 words, but as I mentioned in the report, I think the underestimation of many simpler flash pages (like Strife) generally makes up for the overestimation of one big one.

    There was also a suggestion of using a log or logistic function for equivalence rather than a linear one, which is probably a better option all around.

    "Time Taken To Watch" is a measure I'm trying really hard to automate, but it's proving difficult as hell. Currently my plan is to use the links on this page to fetch the duration of the music attached to each flash file, and then use that as the duration of the file (maybe -15% or something to correct for trimming). By this metric, Cascade would be 13 minutes, 14 seconds. However, this doesn't provide a great option for the adventure-type ones (like Myststuck) or the flash animations without music links.
    This is old, and i do suppose i shouldn't reply to this particular post, but i feel i can make a point about the adjusted word count.

    So let me open with stating the obvious.

    The Cinematic Flashes, in particular the non-linear ones, often cover multiple events, sometimes with the events running alongside each other in a manner that no one would describe anything in plain text. So for the more complex flashes a "simple" description of events is not a filling substitute, because it achieves something text alone would struggle with doing. That's before you count in the music.

    The MSPA wiki's detailed description of [S] Cascade comes up to more than 1400 words.

    While for the static or moving images (at least the ones of a single action), an approximation to a description of what happens in the image is fine, but for the more complex flashes, the medium actually grants something which could never be implemented (well) with words. So for that reason...

    But i'm not really making a point, i just felt the need to convey how small part of the flahes you can convey with words alone.


    On a completely different note. I suppose this is obvious, but i assume when you get a new method of generating the AWC you'll cross check with the old data to see how it compares.


    I might try to do an in-depth description of [S] WV: Ascend for the purpose of word counting. If i don't forget, or give up, or never have the time.
    Bah, i hate coming up with good signatures

  21. #171
    Human of Alfandra simon.clarkstone's Avatar
    Join Date
    Feb 2010
    Location
    Oxford, UK
    Pronouns
    he/him/his
    Posts
    4,549

    Re: Measuring Homestuck (Data! Charts! Graphs!)

    Quote Originally Posted by Morgan Wick View Post
    Decided not to follow the DOTA rule for the flash where Dirk engages trickster mode, because it involves so many hits of the right-arrow key that my own button-hit rate varied as it went along; I timed 3:10 for me to finish it, but your mileage may vary.
    Well it's just 1111 divided by you keyboard repeat rate. You know you can just hold down the right key for that one?
    [U4G] OREG (A space adventure)
    Starring Opabiniataurs, because humanoids are boring.

Page 7 of 7 FirstFirst ... 4567

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •