Self-contained, Internet-less Wikipedia machine. | Hacker News Reader

Self-contained, Internet-less Wikipedia machine.

thewikireader.com

98 pointsasciilifeform69 comments

I know this sounds trite and stupid, but I have a recurring fantasy of travelling back in time and using my modern knowledge to change history. My greatest [imagined] downfall has always been a lack of access to Wikipedia, and how to get around it. This solves my hypothetical problem perfectly.

It's much more interesting to try to actually learn enough about how things work that you could, theoretically, reconstruct some of them. Take gasoline, for example: could you work out how to refine oil by partial distillation? If you needed to measure the temperature, could you make a basic thermometer and calibrate it? Do you know how to make metals from ore? If you go back to before the bronze age, being able to mix copper and tin could be a big deal.

This is really fun stuff to learn about. Plus, this way you won't have to worry about battery life.

It is, but we run into the barrier of expected utility. What is a better thing for me to learn: how to scale rails apps, or how to construct a solar panel from iron age technology?

Don't get me wrong, both are awesome. Absent a working time machine or the ability to leap into a novel or Z-Day, however, only the former will have any benefit for me.

Time travel could be impossible, but the end of the world could be right around the corner. I’m not a doomsdayer, but I do think it’d be fun to contribute my brains to a survival group. Better than contributing them to a zombie’s maw, wouldn’t you say?

How far down the rabbit hole do we go, and how much of my irredeemable nerdiness am I willing to admit to?

If we're talking Z-Day, then the most economical solution seems to be either to buy and convert an old mine, or to grab an old nuclear silo[1]. Either way, one could reasonably assume that a full, automated backup of Wikipedia and other helpful sites would be stored on local servers.

If you're not a member of the "I can buy Minuteman silos, whatever bro" club, then your currency is best spent on ammunition and the pre-determination of likely gathering points.

If, in the unlikely event that you're just a nerd putzing around on the Internet when you should be doing Real Work, you have access to none of these, then I suggest you move quickly towards a zombie-proof survival plan.

[1] http://www.missilebases.com/properties

Learning to make charcoal in bulk or blow glass are valuable purely for entertainment: it's not likely to be useful to you, but you'll have fun doing it. On the other hand, this was also my motive for learning about things that later turned out to have really high expected utility, like the concept of "expected utility".

You might enjoy reading about the difficulties Thomas Thwaites had in building something as simple as a toaster from scratch:

http://www.thetoasterproject.org/

I've seen the TED presentation about this and there is one thing I particularly disliked: all the energy and effort spent molding the plastic enclosure. I think it wasn't faithful to the concept of "recreating civilization from basic premises" that appeals more to me.

I have the same fantasy. I guess Ryan North does too, he produced this t-shirt http://www.topatoco.com/merchant.mvc?Screen=PROD&Store_C...

Which is semi-helpful. By wearing it I've discovered this fantasy is not uncommon.

sigh

Ordered.

Then I recommend Lester del Rey's novel Lest Darkness Should Fall. It's about someone from the 20th century who suddenly and inexplicably gets transported to ancient Rome, after a while he accepts he can't get back to his time and decides to do something worthwhile where he is: keep the middle ages from occurring.

Also, 1632 by Eric Flint (and its numerous sequels) about a Virginian town circa 1999 that suddenly and inexplicably gets transported to 1632.

me too! I've thought a lot about when would be good times to do great things and which paths would be easy and which would be hard. Unfortunately it usually ends up with me realizing that going back to 2004 and starting youtube would be the greatest reward per time and difficulty. :(

I'd prefer going back to '97 with $10MM and buying Apple stocks, then sitting on my ass. My reverie could possibly be interrupted by getting in on Google's IPO. I'd also offer to buy each viral video from the last decade before it went big for $1k, then pop ads on it and let the money roll in.

but you could go to 2004 and make > $1B in 18 months without needing any initial capital. Im not sure theres ever been a 'if only I had know this then' get rich quick plan quite like that in history.

I am disappointed by the lack of comments referring to this as the Hitchhiker's Guide to the Galaxy finally realized.

Don't panic. I'm sure they are coming.

They are already here.

http://kmkeen.com/tmp/wikireader1.jpg

I think this weekend I'll release the image view for the wikireader I've been working on. It can smoothly pan/zoom images that are thousands of pixels wide.

Obligatory XKCD: http://xkcd.com/548/

The H2G2 sent queries over the sub-etha-net.

Asimov's Encyclopedia Galactica is a better literary reference. The tl;dr for Wikipedia is not "Don't panic."

That would be a Wikitravel reader.

You can get it for $27 on Amazon at the moment (normally $99). http://www.amazon.com/Pandigital-Handheld-Electronic-Encyclo...

I'm almost inclined to get one to replace mine, whose screen went permanently dim at some point.

The code is open source, so you could theoretically use it with custom content that you adapt to their build formats. They also provide downloads for a bunch of other-language Wikipedias, as well as Wiktionary, Wikitravel, and Project Gutenberg. You can swap out the built-in microSD card (8GB?) for a much larger one and just put everything on it.

https://github.com/wikireader/wikireader

I've got one of these too, and really like it. We take it when traveling, so if we find ourselves at, say, a Civil War battlefield, we can look up the story behind it.

One might think that you'd just use the browser on your phone. But very frequently, places of interest are also places with no cellular coverage. Having this data offline is great.

They also provide downloads for a bunch of other-language Wikipedias, as well as Wiktionary, Wikitravel, and Project Gutenberg

I tried to do this. Bought a card just for the purpose, downloaded the GBs of torrents, and set up the card as they suggest (as best I could decipher). I was never able to get it to boot up with my substitute data card.

I would dearly love to see a ruggedized, solar-powered version of this device created for our descendants.

In case industrial civilization, well.. fails to go the distance. A Wikipedia-equivalent volume of paper is rather bulky and fragile.

Again and again, this topic keeps coming up and the surprising truth is: paper is the answer. Don't discount it.

There was entire writeup about this (I think it was posted to HN) that I can no longer find and would be grateful if someone could point me at (my Google-Fu is failing me on this one) that of all the data storage mechanisms we have, paper is the one that we best know how to maintain and preserve. It's nowhere near as "bulky" as people make it out to be, and when done right and taken care of properly, can last centuries.

Here's a link that touches on the topic of preserving info for future generations, a great read: http://www2.library.ucla.edu/about/2483.cfm (What do you preserve? How do you do it? Why? For whom?)

The other article I was referring to earlier pointed out the pitfalls with all the modern "wonders" of storage and how woefully inadequate they would be for attempting to preserve knowledge/data in the big picture. Namely interfaces, specs, requirements, readers, electricity, and all the other goodies that silently play their role in powering your USB stick or bluray disc.

Paper is a good answer. Unless civilization has collapsed, and libraries are burning. (They are burning - or rather, being closed down permanently, one after another, right now in the USA.) Paper is the answer, unless you are a refugee living in your car (should you still have gas) or running with what you can carry.

Paper is not the answer for a pauper with no room to call his own, nor for a factory hand living in a dormitory. There are many people alive right now for whom paper is already not the answer. But some of them could benefit from a Wikipedia "gameboy."

Yours or my paper library will carry knowledge into the ages quite well - until some vagrant burns it to warm himself in the wintertime. A piece of electronics, on the other hand - assuming we could make one that really lasts - could be passed down as an heirloom. And, should it begin to flicker out, the words could always be read off and set to paper.

> Again and again, this topic keeps coming up and the surprising truth is: paper is the answer. Don't discount it.

Paper is good, but clay tablets FTW.

If we wrote "Don't Panic" on it, it'd be perfect.

Damn you; I already consider the absence of these words to be a design flaw.

I was considering their claim that it's easy to turn on but difficult to turn off to be a design flaw.

The real technical challenge would be figuring out how to store 16GB of data reliably for that long, I would wager. You'd probably need to use some sort of optical storage that would be chemically stable over the time you expect it to keep.

I'm no expert in this domain, so I'm not sure what this would look like, or if there's already a way to do this in a small enough form factor to be feasible and relatively inexpensive.

How about mask ROM? 100+ years easily (at least in theory.)

The first thing that comes to mind is that any inaccuracies or vandalism are going to be stuck on their devices until they make an update.

Here is an idea for vandalism. Find the last non-bot, non-ip edit. Verify that this user has been on Wikipedia for a while, has a talk page with no vandalism warnings (again you'd have to check the history for the last edit made by someone other than the user), and/or has made more than 100 contributions. Won't get everything but should be effective.

Exactly. Whenever I read about one of these projects that are distributing Wikipedia content offline, my first thought is - How do they check all the content for vandalism? They most probably neither have the time nor the expertise. On the WP site, such things get detected and rolled back quickly and that is one of the reasons the whole thing works. Once you take that out, it is shaky territory.

They don't curate it at all, they just load snapshots.

In practice, it's no shakier than before. Whenever you load a Wikipedia page, there is a certain probability it's vandalized. I'd be surprised if the probability were going down, since as it is more widely used it attracts more trolls. If the probability is staying relatively constant, it doesn't matter if you are looking at the current data or a snapshot from six months ago, the probability of a random page being in a vandalized state is about the same. And of course, if the vandalized page rate is going up, a snapshot would be better than the live site.

You could argue that, if you discover a vandalized page, you can't do anything about it on the Wikireader. You could argue that once you find a vandalized page, it's going to stay vandalized until you load another snapshot. But those are consequences of the tradeoff for the size and network independence of the device.

You are right when considering an individual. When you consider a group of people, the probability that all of them saw the vandalized version on Wikipedia site goes down as the number of people increase. With the offline version, it stays the same.

If you consider the case of disadvantaged children whom some of these projects aim to reach, the whole peer group now has the same vandalized information and there may not be a secondary source of information.

That's a good point. Ultimately it's going to come down to risk versus benefit.

Gameboy pockets seem to run around $20. They seem almost perfect for this sort of thing. They can run on two tripple-A batteries continuously for 10 hours. And since they boot instantly and actually turn off when you tell them to, under normal usage that will last you weeks. The screen resolution could be better but for reading short articles it's good enough.

And when you take the wikipedia cartridge out, you have a gameboy.

I wonder how much it would cost to make a Wikipedia gameboy cartridge?

The Game Boy can address a maximum of 4 megabits of ROM: http://fms.komkon.org/GameBoy/Tech/Software.html

Quite a far cry from the gigabytes of Wikipedia data you get with this, I'm afraid.

The cartridge interface is general-purpose, so you can stick whatever electronics you like in the cartridge. If I was producing these at scale I'd burn into an asic the compressed data, the decompression logic, and some paging/addressing logic. So you say "give me article 1234" to the cart and it streams you the decompressed article. You could do that with an fpga if you were making them in low volume.

For an example of the crazy things you can do with gameboy cartridges look at Gameboy Camera.

The official Nintendo cartridges never used more than 4 megabit of ROM. But many unofficial "backup solutions" integrated a microSD reader and had quite a bit more space available.

$15 shipped - http://slickdeals.net/newsearch.php?forumchoice%5B%5D=9&...

Based on this (see: 1saleaday), this item got excessed/liquidated by a distributor last... fall probably. Judging by the low ratings/views on Slickdeals most people just aren't interested in such a device, even at $20/pc. Several players (maker/distributor/liquidator) probably took losses unless they were on consignment.

Judging from the initial response on HN, this item totally could've been remarketed better. :)

The $15 deals are over. Cheapest price currently is $25.

Just trying to share some experience here since I used to do retail pricing analytics. The $15 prices will come back, if not lower (though perhaps through alternate offline discount-store channels).

If you look at how the price continues to fall after October through Christmas shopping season by one inventory holder (1saleaday, probably has 10k+ pieces to liquidate) and how poorly the deal has been moving even with guerilla marketing (see: views on each thread, as well as thumbs-up)... all I'm saying is that this deal's been a dog in retail.

I believe this. I bought one from 1saleaday for 15 shipped. Since then I've seen it on 1saleaday twice more. Another 15 and also one day for 9.99 with free shipping.

This was made by the OpenMoko guys, after that project failed. It's a pretty neat device, that's surprisingly powerful, but unfortunately there is no commercial backing to drive development.

Wow, user stories/bios are pretty darn powerful. Convinced me of the usefulness of a device I'd initially discarded.

If you're looking to have a Wiki dump on your iPhone/iPod Touch, I can recommend this software: http://www.brilliantish.com/allofwiki/

Disclaimer: I got a free copy because the dev is a nice guy. Well worth the $9.00 though.

I use wikidroyd on my android phone. There may be something better by now, as I installed it a couple of years ago and haven't investigated since.

It would be cooler if they at least gave a token amount of the purchase as a donation to Wikipedia.

I don't see the use a new device for a single function like this. Why not include this kind of function into any e-reader instead? Who wants to carry tons of devices around?

Where is the $100 e-reader that runs for a year on one set of AAA batteries?

While Kindle touch won't run a year on AAA batteries, it's not that far away of it. Based on my calculations I'd guesstimate that Kindle should run at least a month with 2xAAA batteries (assuming same "multiplier" that WikiReader uses, 90 hours = 1 year => 10 hours = 1 month). I'd argue that it's good enough.

I think they were posting from the opposite side of things, more "Given a $100 device that runs for a year on one set of AAA batteries; why not put Wikipedia and Project Gutenberg books and OpenCourseware and everything else possible on it."

Also http://en.wikipedia.org/wiki/WikiReader

Open this page on your WikiReader to get fully meta :).

Making one of these available to every 5-7th (or so) student would be massive improvement to education than just about anything of comparable cost/effort.

According to the wiki article on the WikiReader: The device can also run programs written in Forth, so you can write WikiReader 'apps' also.

I bought one of these recently during a sale. Picked it up for about $30.

It's really nice. microSD is easy access, and AA is awesome as a power source for this little thing. I love the idea of giving these to kids in school.

The software is open source, and you can add custom wikis to the existing software.

specs http://en.wikipedia.org/wiki/WikiReader

cpu S1C33E07 http://www.epson.jp/device/semicon_e/product/mcu/32bit/ (bottom of table)

this looks awesome and what not, but i feel the screen itself is my biggest gripe. i'm only judging this based on the pictures on amazon, so it could look better in person. the screen makes it look like a device you would have found in the 90's :\

What it really looks like is one of the shitty "handheld videogames" your dick uncle gave you for Christmas instead of a GameBoy game. You know the ones, with predefined sprites in predefined positions that would sort of move around with the eventual refresh? The kind your Casio calculator outperformed? The kind you could find in the dollar stor for $6?

I love the idea, but I'm with you on the screen. I'd be willing to shell out more for something Kindle-like.

This is kinda what I wanted to do with the peek devices that were being given away last month. Unfortunately very little progress has been made with the Peek and Linux.

And so the encyclopedia has come full-circle!

I might buy one of these someday for the sheer enjoyment of developing something silly for it.

I think 'internet-less' is a doomed technology

Isn't there an app for that?

I'm still eagerly waiting for an e-ink version to show up.

Reply on news.ycombinator.com