We did the same sort of thing in the pkzip source code 20 years ago. There were a number of pointless routines in the program that would be clear evidence of code theft.
These days I suppose an optimizing compiler would be smart enough to strip away any truly dead ends.
Google Maps directed me down a trap street once on the way to a new tutoring gig. I got lost, was late, and had to sheepishly call the family to ask for directions. Just glad I had a printout of the directions to show them.
Never used it for anything essential after that. Thanks a lot, Google Maps. You'd think they'd have the decency to add a "don't route down this street" flag to their trap streets.
I wonder if it was truly an intentional 'trap street' or a genuine mistake? It would be hard to tell the difference. In Google's case they may be worried about their back end data being copied wholesale, in which case flagged streets would be too easy to remove, or perhaps it was a trap street in a third party data source, and they couldn't tell the difference to remove it. Or maybe it was just an honest error. Interesting that trap streets and errors make good cover for each other...
That doesn't really make any sense: "trap streets" make sense and are used in, say, the base map that you see when viewing a Google map (that is, in the raster data). But why would Google bother putting a trap street into their own (proprietary) street data used for routing? Unless you are saying that Google Maps "directed" you in the sense that you were looking at a printout of a map that included a trap street, not actually a turn-by-turn.
I agree, it doesn't make any sense. But it happened. This was back in '07 or so, so before Google did live turn-by-turn. I had a printout of Google's directions from online.
It led me on a street that's never existed, and that would have had to go either over or through a very substantial hill to actually exist.
I make mountain bike trail maps[1] using OSM data, but when generating my first drafts of the final PDF files and distributing them for review I'll include similar trap landmark.
Along similar lines, the shareware assembler A86 watermarked the binaries it produced, so that unauthorised use of unregistered copies could be proven. It did this by - or so it was claimed - choosing one specific instruction over another, when it had a choice:
This seems like it gets posted every week. I have to say so what? Is it not totally understandable? I mean, cartography is an invaluable tool and because there information is public knowledge there has to be some way to stop those would cut corners and copy a map. Does it lead people astray sometimes? Sure. But is it causing so much harm that the practice needs to be called into question? No.
I'm not really sure why this gets posted here without any context though. Is this supposed to be a statement about copyright? If so, I have to say I'm really sick of these clever attempts to show how copyright is bad. Copyright is not good or bad in and of itself. It can be both and any discussion of it in absolute terms is ridiculous.
I don't think the original posting was taking aim at copyright infringement. I think it's just a neat idea that appeals to the hacker instinct - which is why it keeps popping up again and again.
Although trap streets specifically refer to maps, I can't help but think of the trap that Google set up to bust Microsoft on copying their search results:
22 comments
[ 3.5 ms ] story [ 30.6 ms ] threadI was googling for examples in Google Maps, and came across the corresponding article on open street maps[1], which is a nice supplement.
[1]: http://wiki.openstreetmap.org/wiki/Copyright_Easter_Eggs
These days I suppose an optimizing compiler would be smart enough to strip away any truly dead ends.
Never used it for anything essential after that. Thanks a lot, Google Maps. You'd think they'd have the decency to add a "don't route down this street" flag to their trap streets.
http://www2.census.gov/geo/tiger/TIGER2012/ROADS/
It led me on a street that's never existed, and that would have had to go either over or through a very substantial hill to actually exist.
[1] Examples: http://mmba.org/library/maps/riverbends/latest.pdf http://mmba.org/library/maps/clintonriverparktrail/latest.pd... http://mmba.org/library/maps/stonycreek/latest.pdf http://mmba.org/library/maps/addisonoaks/latest.pdf
http://www.preterhuman.net/texts/computing/general/DOCUMENTA... (see section 6)
Revisited here, 40 days ago, http://news.ycombinator.com/item?id=4818584
I'm not really sure why this gets posted here without any context though. Is this supposed to be a statement about copyright? If so, I have to say I'm really sick of these clever attempts to show how copyright is bad. Copyright is not good or bad in and of itself. It can be both and any discussion of it in absolute terms is ridiculous.
http://searchengineland.com/google-bing-is-cheating-copying-...