dariusk / NaNoGenMo-2014

National Novel Generation Month, 2014 edition.
257 stars 17 forks source link

A Web Browser's Diary #78

Open WhiteFangs opened 9 years ago

WhiteFangs commented 9 years ago

The title says it all! But I won't be able to start before mid-november and won't have a lot of free time to work on it... Anyway, I'll try to do something that works, and then expand it later if needed!

MichaelPaulukonis commented 9 years ago

[history deleted]

WhiteFangs commented 9 years ago

So I couldn't wait to participate and did some basic stuff today to begin, you can find the repo here: https://github.com/WhiteFangs/NaNoGenMo2014

You can try a demo of the current version on my website: http://louphole.com/apps/NaNoGenMo2014/

I thought at the beginning about doing something using the browser's history but then I changed my mind...

I express some of my ideas and hesitations in the README, so I'll just copy/paste here to spare you a click:

Current progress

The novel needs a title that is taken as an input by the user. The first entry of the diary shows general informations about the browser and the client (name, OS, IP, localisation, cookies parameters). Then, it searches on Google for the title given by the user and displays 8 results of the Google search at a random page between 10 and 50.

The content is generated asynchronously so that one can read the beginning while the rest of the novel is being written. The header of each entry is a UNIX timestamp and the date related to it. The beginning date is chosen randomly between 2011 and 2012 and then every new entry is in the week following the previous one.

What's next?

The rest of the novel is not defined yet, I have two ideas:

It would eventually react to some informations about the page (title, language, JS files, CSS files, etc...) and display the main headers and random sentences from the paragraphs of the page. Using meta tags (description keywords), it can link one page to another one already visited (in its "memory"). The external links provided on the web page would be "paths" to somewhere else for the browser and one of them could be chosen randomly to go to the next page. If no external link or a 404 is found, it could react and go back to the previous page and choose another one.

Let's imagine

We could include "emotions" with the use of adjectives and also implement a bit of evolution. For example, two arrays of happy and sad adjectives, and a probability of picking the adjective in the first one decreases as for the second one increases with time. Or we could imagine that the more it sees Google analytics, Facebook or Twitter integrations, the more it could express paranoïa of being tracked.

WhiteFangs commented 9 years ago

I implemented the crawling and display, now I just need to make it more interesting. You can try the current version here : http://louphole.com/apps/NaNoGenMo2014/ It already achieves to reach 50k+ words in most cases... I need some more error handling still.

I thought about making something inspired by Gogol's Diary of a Madman, which would then be Diary of a Mad Browser, but it's not sure yet. I opened some issues that I will focus on in the coming days to make this more fancy and readable by anyone : https://github.com/WhiteFangs/NaNoGenMo2014/issues/

If you have any ideas, comments or suggestions, I would love to read them ! Stay tuned for progress in the coming days...

WhiteFangs commented 9 years ago

DONE !

I'm glad to present you: A Browser's Diary - Writing a novel (unofficial name is "Sisters of Cleveland", you'll understand if you read until the end)

You can read it here: http://louphole.com/apps/NaNoGenMo2014/novel/A%20Browser's%20Diary.html 50195 words!

Generate your own Browser's diary here: http://louphole.com/apps/NaNoGenMo2014/

And see the code here: https://github.com/WhiteFangs/NaNoGenMo2014

As I said before, I was a bit inspired by Gogol's Diary of a Madman so the Browser's Diary goes worse and worse, there's at least this dynamic. Something interesting though is that the first pages are taken from Google and thus are usually about news, like Ferguson, because of Google's result system. I had to do some ugly fixes because my server was returning an error at some point after calling lots of time the same request but anyway, it works now! Enjoy!

cpressey commented 9 years ago

:heart: the madness! :scream: