What-If 0059: "Updating a Printed Wikipedia"

[higgs-boson]

Updating a Printed Wikipedia

If you had a printed version of the whole of (say, the English) Wikipedia,
how many printers would you need in order to keep up with the changes made to the live version?

Marein Könings

Short Answer:

(for text only I guess one powerful matrix printer would be sufficient)


For having fun:

Wait for the next WP blackout, order the needed bunch of text markers...
... open all caps, and inhale deeply.

[najodleglejszy]

A few years ago, student Rob Matthews printed every Wikipedia featured article, creating a book several feet thick.

while in the source given it says 1 ft 7 in. tut tut.

[EvilHom3r]

I'm not sure what would be worse, having to sharpie every page, or rerunning them through the printers to print them entirely black and paying for the ink.

[peewee_RotA]

There was nothing in the original question that forced him to use inkjet printers. It was fun to see the figures on it, but leaving off on a note that he'll have to switch to laser without exploring it further was just lame. Especially the number of times that a question is altered to randomly fit in a crazy piece of technology.

We can't use laser pointers, we have to equip all of india with high powered lasers, but goodness me, we can't talk any further about the cost of toner! That would be extreme!

[Klear]

I'm not sure what would be worse, having to sharpie every page, or rerunning them through the printers to print them entirely black and paying for the ink.

Or maybe just boy a lot of black paper?

There was nothing in the original question that forced him to use inkjet printers. It was fun to see the figures on it, but leaving off on a note that he'll have to switch to laser without exploring it further was just lame. Especially the number of times that a question is altered to randomly fit in a crazy piece of technology.

We can't use laser pointers, we have to equip all of india with high powered lasers, but goodness me, we can't talk any further about the cost of toner! That would be extreme!

Yeah, I thought so too.

[wolfticket]

Surely it would make sense to use some sort of large scale continuous ink supply system.

I actually think collation would be the biggest issue

[Flumble]

Is sharpie the one and only manufacturer of markers in the USA? I'm used to calling it a 'marker', not the company that makes them (actually I say 'stift' or 'markeerstift' but you get the idea).

Or maybe just boy[sic] a lot of black paper?

That would be more fair towards 'printing an edit', yes, as Randall wants to inconsistently edit the paper. I'd just apply a lot of black paint or setting it on fire in a low-oxygen environment.
Forget the painting, charring all the paper is much more fun later on.

[TimeLurker]

Is sharpie the one and only manufacturer of markers in the USA? I'm used to calling it a 'marker', not the company that makes them (actually I say 'stift' or 'markeerstift' but you get the idea).
<snip>

In America we tend to use a brand's name in place of a product name when one brand has a large market share. Such as 'Legos' vs. 'small plastic blocks'

[Klear]

Screw inking/bruning. The idea is that the wikipedia is unavailable, right? So just lock the whole thing somewhere dark. You won't even have to print it out again the next day.

[cellocgw]

Mmmmmm, sniffing Sharpies. Not as much fun as sniffing Magic Markers, which isn't as good as the long -lamented Ditto Masters.

[Sandor]

There is a "footnote" link to an image of IBM punchcards in the Federal records center in Alexandria, Virginia, November 1959:



According to the link what you can see represents 4.3 billion characters of punchcard data, so less than one DVD's worth

[Whizbang]

What If BHG.png (12.07 KiB) Viewed 16170 times

What if Randall wrote an article about editing Wikipedia; how many printers would you need to keep up with all the troll edits?

[Flumble]

The idea is that the wikipedia is unavailable, right?

The implementation is that there appears an overlay as soon as you request a document.
But then you miss out on the great charcoal fire.

According to the link what you can see represents 4.3 billion characters of punchcard data, so less than one DVD's worth

That character is most likely an EBCDIC character, for a a total of a little over 32GiB. Fits nicely on a modern thumb drive.

In America we tend to use a brand's name in place of a product name when one brand has a large market share. Such as 'Legos' vs. 'small plastic blocks'

Surely we play with lego(s) in Europe and probably anywhere in the world, as it's the sole manufacturer of fitting lego bricks. But I didn't expect a(n apparent) monopoly in markers. Thinking about it, something trivial as chocolate milk is often called chocomel in Dutch even though they have just a large market share.

[cellocgw]

Spoiler:

The attachment What If BHG.png is no longer available

What If BHG.png (11.13 KiB) Viewed 16084 times

FTFY

[higgs-boson]

In America we tend to use a brand's name in place of a product name when one brand has a large market share. Such as 'Legos' vs. 'small plastic blocks'

Surely we play with lego(s) in Europe and probably anywhere in the world, as it's the sole manufacturer of fitting lego bricks. But I didn't expect a(n apparent) monopoly in markers. Thinking about it, something trivial as chocolate milk is often called chocomel in Dutch even though they have just a large market share.

On a side note: There are other manufacturers, and of course their products are not compatible to LEGO, or to LEGO Duplo (since Duplo seems to be accompanied more often than the standard sized LEGO). And it is beside the point: Of course the described phenomenon exists in Europe, too. For the German speaking part, this page provides good examples (well, most are better than LEGO, anyway). And one or two awefulsome good examples^Plexiglas.



^PlexiglasNo kidding.

[Klear]

In America we tend to use a brand's name in place of a product name when one brand has a large market share. Such as 'Legos' vs. 'small plastic blocks'

Surely we play with lego(s) in Europe and probably anywhere in the world, as it's the sole manufacturer of fitting lego bricks. But I didn't expect a(n apparent) monopoly in markers. Thinking about it, something trivial as chocolate milk is often called chocomel in Dutch even though they have just a large market share.

It doesn't have to be monopoly. I think it's more important for the company to introduce this kind of product.

Incidentally, just yesterday I was reading about Orphaned Etymology, and I feel it's unfair that such a nifty and useful linguistic term is actually only used on TVtropes...

@cellocgw
Hey, keep that stuff in your weird little thread, will you?

[Himself]

In my experience, it seems to me that reverting an edit occurs far more often than an edit that affects multiple pages. There is always an edit war going on somewhere.

[enumerated powers]

Now here's an oddity. The piece included a link to a site which tracks edits...

http://toolserver.org/~emijrp/wmcharts/wmchart0001.php

And it looks like this...



What do you suppose happened on August 10, 11, 12 --- particularly the 11th ???

(Even 'bots took the day off.)

[rhomboidal]

So I'm guessing a live-updating audiobook version is out of the question, then. I'll delete it from my Amazon wish list.

[mcdigman]

There is a "footnote" link to an image of IBM punchcards in the Federal records center in Alexandria, Virginia, November 1959:



According to the link what you can see represents 4.3 billion characters of punchcard data, so less than one DVD's worth

Also, at 4.3 gigabytes that warehouse is half the size it would need to be to hold just the article text after compression with bzip2, and 1/442 as big as it would have to be to include all the multimedia files, assuming you printed the binary instead of the image, which you definitely shouldn't.

[peewee_RotA]

Is sharpie the one and only manufacturer of markers in the USA? I'm used to calling it a 'marker', not the company that makes them (actually I say 'stift' or 'markeerstift' but you get the idea).
<snip>

In America we tend to use a brand's name in place of a product name when one brand has a large market share. Such as 'Legos' vs. 'small plastic blocks'

Sharpie markers are permanent fine tipped markers. Fat tipped felt markers, dry erase markers, fine tipped pen markers, washable markers, etc. aren't called by their brand name normally. Sharpee manufactures a lot of types, but only the fine tipped permanent marker gets that common name. If you're looking for an equivalent example, we use binder clips when you guys use bulldog clips. (Bulldog being the brand name across the pond)

TimeLurker is not quite accurate as adapting brand names for things is pretty common. It only sounds unusual if the brand has survived the test of time. For example, Frisbees don't exist (Apart from Wham-o holding onto the trademark and being a trade-mark troll with the name). We coat our pans with teflon. We give fevered children tylenol. We sit in our lazy-boy. We wear khakis. We use our bifocals to read our thesaurus while eating our cheerios.

The point is that brand names, trademarks, and catch phrases are a historical way to coin terms. The winner just seems to be the one who gets their version of the word out to the most people for the longest time. Sounds like a good description of someone who holds the market share.

[keithl]

What if Randall wrote an article about editing Wikipedia; how many printers would you need to keep up with all the troll edits?

How much would BHG spend on plane tickets if he tracked down and "edited" the trolls?

Detective: Can you describe the assailant?
Troll (in a death rattle voice): Black hat, round face, extremely skinny ...

[fizzfaldt]

By comparison, if you were trying to keep up with the edits, you'd print out 300 cubic meters every month.

I think this is wrong by an order of magnitude.

Assume 30 days in a month (43200 minutes)
Use a standard ream (500 pages is 2.25 inches thick)
Assume pages are 8.5 inches by 11 inches

(100 edits / minute) * (1 page per edit) * (43200 minutes in a month) * (8.5 inch * 11 inch) * (2.25 inch / 500 pages) = 29.785783 m^3

[keithl]

On the other hand, a pair of decent desktop scanners could scan the 32640 pages of the final revised print edition (2012) of the Encyclopedia Brittanica in less than a day. Until the lawyers took down the website, the posted scans could tell the world such timely information as who won World War II. With court costs, more expensive than the laser toner needed to print Wikipedia.

But for printing Wikipedia, who needs toner? AFAIK, the fastest laser printer ever made was a one-off monstrosity built by Lawrence Berkeley Labs in the 1970s. That used a multi-kilowatt laser to burn images directly onto 14 inch pin-feed line printer paper, passing between big rolls at a few hundred pages per second. It took a while to start and stop the multiton paper rolls, and quite a while to manually "burst" the print jobs. But hey, what are expendable graduate students for?

[marsilies]

According to the link what you can see represents 4.3 billion characters of punchcard data, so less than one DVD's worth

That character is most likely an EBCDIC character, for a a total of a little over 32GiB. Fits nicely on a modern thumb drive.

Actually, EDCDIC wasn't invented until 1963/1964, while that photo is reportedly from 1959. So it was likely EDCDIC's precursor BCDIC that they were coded in, which was 6-bit per character, or 480bits per card, for 25.8GiB total:
http://en.wikipedia.org/wiki/BCD_(6-bit)

Interestingly, there was also just a straight-up binary coding format for punch cards that would hold about 864 bits per card. The IBM 704, made in 1954, could use this format:
http://en.wikipedia.org/wiki/Punched_ca ... cter_codes
http://en.wikipedia.org/wiki/IBM_704

This page shows how many of various punch-card types and formats it would take to fill a 16GB micro-SD card:
http://ferretronix.com/march/computer_cards/sdcard/

[gmalivuk]

Is sharpie the one and only manufacturer of markers in the USA? I'm used to calling it a 'marker', not the company that makes them (actually I say 'stift' or 'markeerstift' but you get the idea).
<snip>

In America we tend to use a brand's name in place of a product name when one brand has a large market share. Such as 'Legos' vs. 'small plastic blocks'

Sharpie markers are permanent fine tipped markers. Fat tipped felt markers, dry erase markers, fine tipped pen markers, washable markers, etc. aren't called by their brand name normally. Sharpee manufactures a lot of types, but only the fine tipped permanent marker gets that common name. If you're looking for an equivalent example, we use binder clips when you guys use bulldog clips. (Bulldog being the brand name across the pond)

TimeLurker is not quite accurate as adapting brand names for things is pretty common. It only sounds unusual if the brand has survived the test of time. For example, Frisbees don't exist (Apart from Wham-o holding onto the trademark and being a trade-mark troll with the name). We coat our pans with teflon. We give fevered children tylenol. We sit in our lazy-boy. We wear khakis. We use our bifocals to read our thesaurus while eating our cheerios.

The point is that brand names, trademarks, and catch phrases are a historical way to coin terms. The winner just seems to be the one who gets their version of the word out to the most people for the longest time. Sounds like a good description of someone who holds the market share.

Yeah, everyone genericizes brand names, which companies tend to hate when it means people buy anything that performs the same function, but secretly love when it means theirs is the only brand with any recognition. And it isn't more common on either side of any pond, either. In many cases it' s just a matter of different brand names being used.

In any case, though, I feel like Sharpies are more akin to LEGO, in that people (in my experience) almost always really are using Sharpie brand fine-tipped permanent markers when they use that word. It's like Polaroid or Kodachrome for those cameras, rather than Kleenex for any type of facial tissue.

[Flumble]

Actually, EDCDIC wasn't invented until 1963/1964, while that photo is reportedly from 1959. So it was likely EDCDIC's precursor BCDIC that they were coded in, which was 6-bit per character, or 480bits per card, for 25.8GiB total:

I stand corrected. I didn't even know there were other and older encodings in those days.

[PM 2Ring]

Interestingly, there was also just a straight-up binary coding format for punch cards that would hold about 864 bits per card. The IBM 704, made in 1954, could use this format:
http://en.wikipedia.org/wiki/Punched_ca ... cter_codes

True. Those cards had 12 rows of 80 columns, giving a total of 960 bits in binary coding format, but cards with too many holes punched out tended to be rather fragile. But they did have their uses, eg that format was used for the IBM 360 bootloader, so it could be used for small general purpose programs, too. It was fun optimizing a machine code program down to 120 bytes so that it could fit on one card.

[mcdigman]

Is sharpie the one and only manufacturer of markers in the USA? I'm used to calling it a 'marker', not the company that makes them (actually I say 'stift' or 'markeerstift' but you get the idea).
<snip>

In America we tend to use a brand's name in place of a product name when one brand has a large market share. Such as 'Legos' vs. 'small plastic blocks'

Sharpie markers are permanent fine tipped markers. Fat tipped felt markers, dry erase markers, fine tipped pen markers, washable markers, etc. aren't called by their brand name normally. Sharpee manufactures a lot of types, but only the fine tipped permanent marker gets that common name. If you're looking for an equivalent example, we use binder clips when you guys use bulldog clips. (Bulldog being the brand name across the pond)

TimeLurker is not quite accurate as adapting brand names for things is pretty common. It only sounds unusual if the brand has survived the test of time. For example, Frisbees don't exist (Apart from Wham-o holding onto the trademark and being a trade-mark troll with the name). We coat our pans with teflon. We give fevered children tylenol. We sit in our lazy-boy. We wear khakis. We use our bifocals to read our thesaurus while eating our cheerios.

The point is that brand names, trademarks, and catch phrases are a historical way to coin terms. The winner just seems to be the one who gets their version of the word out to the most people for the longest time. Sounds like a good description of someone who holds the market share.

Yeah, everyone genericizes brand names, which companies tend to hate when it means people buy anything that performs the same function, but secretly love when it means theirs is the only brand with any recognition. And it isn't more common on either side of any pond, either. In many cases it' s just a matter of different brand names being used.

In any case, though, I feel like Sharpies are more akin to LEGO, in that people (in my experience) almost always really are using Sharpie brand fine-tipped permanent markers when they use that word. It's like Polaroid or Kodachrome for those cameras, rather than Kleenex for any type of facial tissue.

The worst was when I caught myself saying 'I googled it with egrep'

[Klear]

The worst was when I caught myself saying 'I googled it with egrep'

Actually, I do that consciously. For me, to google has completely replaced "search on the internet", no matter which search engine you use.

[PM 2Ring]

The worst was when I caught myself saying 'I googled it with egrep'

Actually, I do that consciously. For me, to google has completely replaced "search on the internet", no matter which search engine you use.

But egrep is for searching for patterns in local files. I guess you could retrieve stuff from the internet using wget or curl and pipe the output through egrep, but it would be a bit slow.

[scribbler]

The idea is that the wikipedia is unavailable, right?

The implementation is that there appears an overlay as soon as you request a document.
But then you miss out on the great charcoal fire.

Yeah, that's the part I don't get.

If the blackout is an overlay, wouldn't the analog equivalent be a guy slamming a shutter on a reader's hand when they try to take a volume?

Saves buying those sharpies.

[laptopleon]

Even the printers themselves wouldn't be too expensive, despite the terrifying replacement cycle.

As for the main question: Of course this all depends completely on what kind of printer is used. I'm sure there are much faster printers available that are also much more expensive. But for the sake of the article, let's look at the rest:

The paper would be about one cent per sheet, which means you'll be spending about a thousand dollars a day.

This seems a bit steep. Even at consumer quantities, one can find paper at € 0,004 in the Netherlands, so in the US it should be less, despite the somewhat cheaper dollar. Also, at 100 pages per minute, one would need 8640 reams (500 pages) per month. I'm sure you'll get a much better price when you order so much paper. All in all, paper should cost no more than about $500 per day.

You'd want to hire people to manage the printers 24/7, but that would actually cost less than the paper.

Now here's were it gets really interesting. You'd need at least 1 full-time employee to mind the printers 24/7. Refilling the paper and ink, taking away the printed paper, helping out when paper gets stuck etc. In the EU, the average cost per hour is €23,50 x 24 hours = €564 = ± $750 (today) to manage the printers, more than the cost of the paper.

Of course, if you'd really want to maintain a paper version of wikipedia, you'd need a lot more people to keep up with the 100 pages per minute print-out, replacing the pages on the fly. Even if they work without breaks, in 8 hour shifts, they would be very fast if they could replace 1 page per minute in the huge set of volumes. At 100 pages per minute, you'd need 100 people x 24 hours x €23,50 = € 56400 = $75000 per day. So as usual, in a real-life situation, labour would be the largest cost, by far, not the laser printer toner.

[mikrit]

Michael Mandiberg prints out all of Wikipedia in book form.
http://www.dailymail.co.uk/news/article ... pedia.html

[stopmadnessnow]

...in a real-life situation, labour would be the largest cost, by far, not the laser printer toner.

Thousands of people volunteer to write for Wikipedia every day. Why not employ those people to print the pages out? (And Wikipedians could also donate to the cause as well...)