Half a croissant, on a plate, with a sign in front of it saying '50c'
h a l f b a k e r y
The best idea since raw toast.

idea: add, search, annotate, link, view, overview, recent, by name, random

meta: news, help, about, links, report a problem

account: browse anonymously, or get an account and write.

user:
pass:
register,


                                                   

Junk Mail Recycler

Become content with junk mail delivery.
  (+3, -2)
(+3, -2)
  [vote for,
against]

OK, so sometimes I work away from home for extended periods of time - the one thing I hate is coming back to a pile of advertisements, free offers, pre-approved credit-card offers (yeah because my credit is THAT good!) and advertising papers that have been issued through my letterbox.

Now, if I had an assistant, I'd be able to task her with organising my post prior to my arrival, but I don't - So, what if I had a machine that lurked beneath my letter- box ready to catch any unwelcome (and welcome) postage. Using OCR and a fuzzy logic memory it would learn what was definitely junk, what was real post, and what it wasn't sure about.

The junk mail, it would shred, mix with water and some glue or something and make papier maché plant-pots, or fuel- briquettes, or something, anything that means I don't have to deal with the stuff.

The good mail, it would sort neatly into a shelf, toast-rack style, and the unknown mail would be presented to me on return to identify whether it really was junk or not - helping it to make a better judgment next time round.

This machine would act like a spam-filter for your house, only it would put to good use the paper effluent that's dumped through your letterbox on a daily basis.

zen_tom, Aug 14 2004

Heat your home with it http://fastq.com/resemail_policy.php
"...we laughed when we read about a person in Alaska who tried to get on every junk mail list possible, so he could burn it and heat his house." [Klaatu, Oct 05 2004, last modified Oct 17 2004]

<link from Around TUIT> http://www.wired.co...,1848,61140,00.html
Post Office Gets Pressured to Pry [Around TUIT, Oct 05 2004, last modified Oct 17 2004]

<link from Around TUIT> http://www.opex.com/mps7.htm
Scans Up to 7,200 items per hour! [Around TUIT, Oct 05 2004, last modified Oct 17 2004]

Please log in.
If you're not logged in, you can see what this page looks like, but you will not be able to add anything.
Short name, e.g., Bob's Coffee
Destination URL. E.g., https://www.coffee.com/
Description (displayed with the short name and URL.)






       Sorry mate! - bandying phrases like 'fuzzy logic' around isn't going to save you!. We all know that junk mail is far too sophisticated to be so easily sorted.
* m-f-d magic (box technology) charitably removed.
gnomethang, Aug 14 2004
  

       Now hold on a minute! Just because you may not be able to imagine a system being able to tell the difference between junk mail and the more normal stuff doesn't mean it's magic.   

       Fuzzy logic is a real and practical solution to providing analysis of real- world data. That does not mean it's magic. Let me elucidate...   

       Suppose a system with OCR capabilities were able to test the size, colouring, and shape of a letter, as well as being able to read the text. Each letter received would have certain properties that might allow it (or another letter sent from the same source) to be recognised again.   

       A scoring system (which is basically all fuzzy logic really is) is used, so that mail sent to 'The Occupier' is assigned a low 'Good Mail' score and a high 'Junk Mail' score, whereas mail addressed to you might receive a more positive rating. Brightly coloured mail, or mail containing words like 'Important Documents Enclosed' or 'You have won £1000', or 'Pre-Approved Application Inside' or other such rubbish could well receive high 'Junk Mail' scores.   

       Some algorythm that compared various scores with recognised letter sources could be trusted to make reasonable choices - perhaps after a period of training, during which time, the recycling module is left in the 'Off' position.   

       We're using technology that is already in existance in email spam-filters across the world today.
zen_tom, Aug 14 2004
  

       tech-driven post cosumer waste meets nature and everybody wins?!   

       Croissants for you!
Around TUIT, Aug 14 2004
  

       Hmmm! - I still think that the junk mail that comes through my door today is less and less recognisable as such. Your training period would be a good deal longer than you think, I suspect.
I am aware of fuzzy logic in automated control solutions for heating and cooling applications but I still think that your whole box has a lot of problems including the OCR bit. You are assuming that you can identify and manipulate the letter to be the right way up and I am pretty sure that the variety of handwritten letters and even printed text that resembles handwriting will cause big problems.
I am also troubled by the fact that the device will be so large that you will not be able to open your front door!.
gnomethang, Aug 15 2004
  

       Handwritten postage is often correctly addressed - direct marketing would soon collapse if the marketeers had to employ people to hand-write the addresses on.   

       This had led many of the direct marketers to try printing faux-handwritten addressed envelopes. It may be difficult to tell the difference between real handwriting and fake handwriting, but it may be possible. However, being able to read is not the primary feature of the machine. What is key, is the ability to recognise the source of a letter. Once the direct marketer has sent you one faux-handwritten letter (where the text is so-big, and of a certain colour, the stamp is franked, and the envelope of a certain size, shape etc) the machine will recognise future letters from the same source and treat them appropriately.   

       Getting a letter the right-way up shouldn't be an issue - at least the part recognising which side is 'up'   

       And finally, the size of the device...hmmm it may be rather lager than your average letter-catcher, but that doesn't mean you wont be able to open your door, you just may not be able to open it as *much* as you perhaps are accustomed to. Those of a fuller figure who may be otherwise unable to gain access to their homes, may prefer to use the mini-sort, which simply sorts the mail into junk and non-junk variants, allowing you to deliver the rubbish to it's sister-unit, the mini-mash, which will turn it into something useful.   

       Or you could get your assistant to do it.
zen_tom, Aug 15 2004
  

       YEah! but you forgot the bit of your machine that will do the papier-mache bit. Too Large Sir!, Too Large!
gnomethang, Aug 15 2004
  

       [gnomethang] yes, the recycling unit may well have to be a separate piece - however, it is an integral part of the 'system'.   

       [Contracts] You may have a point, but I've noticed that most envelopes that feel the need to express their importance one way or another are often doing so with the express intention of goading you into opening them.   

       If you were going to send important documents, would you really advertise the fact that they were of some value as they passed through the postage system?   

       However, since [gnomethang] has a point re the size issue, we may have to settle for a simple sorting device, and a separate recycler that you can manually put the rubbish into after checking that deceased relatives are going to remain in situ.
zen_tom, Aug 15 2004
  

       (corrected)Sorting isn't really a problem, just tell the software to look for stuff you DO want (drops it in a box), then take the obvious junk mail (ex.anything not in an envelope, flyers, coupons, whatever you tell it) and drops it into a blender) and put the unknowns in a pile.   

       The obvious stuff can be shredded right away. Hand inspect the rest of the unknowns.   

       Worried about missing anything? Have it scan and save anything it sees. That way the stuff you do want gets to you faster!
Around TUIT, Aug 15 2004
  

       //YEah! but you forgot the bit of your machine that will do the papier-mache bit. Too Large Sir!, Too Large!// [gnomethang]   

       Just have the buckets the mail drops into be blenders. Each would have a small water tank in the same place. Put drains and flowing points all in the same places. The could be set to auto (programable) or manual.   

       Mail in bucket. Blender shreds as water and glue are added. Paper sticks to walls. Blender off. When it gets to the right size (determined by an adjustable lever in the bucket), empty.   

       Stack them 3 high. Put outside and cover in rock or wood. Now it's an outdoor mail box! Put two in your office! The paper plate model hangs on your door!
Around TUIT, Aug 15 2004
  

       By the stuff you do want, I mean regular bills, hand-written stuff, etc. It is easier to tell the software what you do want to keep, than everything you don't.   

       Anything else?
Around TUIT, Aug 15 2004
  

       [Around TUIT]. I am actually coming round to the idea in general as stated by [zen_tom] since he is coming back with reasonable arguments.
Your last two missives have been serious nonsense!
gnomethang, Aug 15 2004
  

       //Your last two missives have been serious nonsense!// [gnomethang]   

       How would my suggestions not get closer to the stated goal than what has been proposed in the statements?
Around TUIT, Aug 15 2004
  

       1) //Just have the buckets the mail drops into be blenders. Each would have a small water tank in the same place. Put drains and flowing points all in the same places. The could be set to auto (programable) or manual.//
q.v. Too large Sir! Too Large!
2) //Stack them 3 high. Put outside and cover in rock or wood. Now it's an outdoor mail box! Put two in your office! The paper plate model hangs on your door!
Explain yourself!
3) //It is easier to tell the software what you do want to keep, than everything you don't. //
This is not at all obvious and there is no explanation as to how or why.
BTW, I am actually aware of the concept of Fuzzy Logic (It is not just the same as 'woolly thinking' but more so) and I a not sure that it is suited to this application.
I also thought that the bakesperson MFD'ded this about 10 minutes ago. Either she removed it or it got deleted by the proposer. I happen to agree with her.
gnomethang, Aug 15 2004
  

       //q.v. Too large Sir! Too Large!//[gnomethang]   

       Okay, make long flat box, put a fan of blades in it and make paper plates or whatever.   

       //2) //Stack them 3 high. Put outside and cover in rock or wood. Now it's an outdoor mail box! Put two in your office! The paper plate model hangs on your door! Explain yourself!//[gnomethang]   

       I don't get my mail at my door. I get it in a box by the road.This is a variant model for sorting,which is the main stated goal (ie. //Now, if I had an assistant, I'd be able to task her with organising my post prior to my arrival// [zen_tom]   

       //3) //It is easier to tell the software what you do want to keep, than everything you don't. // This is not at all obvious and there is no explanation as to how or why.// The list things you don't want is going to be much larger than the list of things you do want. Each thing on each list needs an action (what to do) associated with it. Bigger the list, the more stuff to do.What I'm saying is, the list of stuff you do want is easier to work with because the list small.
Around TUIT, Aug 15 2004
  

       Tried to change my orginal comment to make it more clear.Any better?
Around TUIT, Aug 15 2004
  

       [gnomethang] I missed the MFD, so no, I didn't delete anything - as for the fuzzy logic thing, perhaps it is an overused term - any system designed to make decisions is eventually bound to choose one option or another.   

       The way I would propose it would work is by either employing a whitelist, or blacklist approach(potentially both). The more difficult part is providing the system with enough senses to be able to recognise the differences between senders. This could be done as stated earlier by checking the type of stamp used (vs franked/pre-paid post), any return address information, the addressee (some companies might have me down as 'Zen Tom', or 'Z Tom' or even 'Zen Thomas' depending on where they originally got my name from) then there's the dimensions of the envelope, colouring etc. Each of these properties could be given a different 'vector' in an evelope phase space, and saved to a database, along with the information provided by the user as to which envelopes are good, bad, or unknown. Future mails could have their properties tested against the database in order for them to be recognised as having been previously classified.   

       As I said before, the technology has been proven in spam-filters employed in the email world, this idea shifts the application to paper-based communications.
zen_tom, Aug 15 2004
  

       Couldn't it sort out anything that is not a standard size envelope? Could it look for something that is only one size?
Around TUIT, Aug 15 2004
  

       Content based filtering in email spam filters is much harder to get 100% right 100% of the time because the people sending the emails are able to quickly change their methods to get around new filtering methods.   

       So yes, there is always going to be the odd exception that gets through. Doing this with real mail however would have a larger impact on the direct marketers due to the higher costs involved in their having to change printing methods, envelope standards etc to keep up.   

       For a spammer, 1 million messages can be generated using some new exploit for very little cost - it is for this reason that the complexity of spam-filters and the methods used to circumvent them is escalating as rapidly as it is. It is also the reason why employing something like this in the real world would be more successful, for a longer period of time.
zen_tom, Aug 15 2004
  

       For so much talk, only 2 votes?   

       I vote (+) on this just because I'm an optimist, and hate junk mail too.
cloudhedd, Aug 15 2004
  

       Thanks [zen_tom]. I did see the MFD and she (the bakesperson) was being less charitable than me!.
Never mind [Around TUIT]'s arguments, My main problem is that your invention will be huge and he in no way alleviates this difficulty.
gnomethang, Aug 16 2004
  

       Links for you!   

       One is about the pressure on post offices to make mail more identifiable.   

         

       By the way, they do have these machines already in businesses that receive large volumes of mail. See link. One of the ways it does this is by the size of the envelope. At the govenrment facility where I worked, there were machines were running 6 hours a day at 1200 documents a hour. One of the ways they would make sure they were only scanning documents they needed was to take out any envelope that wasn't a standard letter size. Junk mail often is in a non-standard size because it is easier to sort in at the post office (in the country I live in the post office takes contracts to put in junk mail). The post office also adds inserts, one page flyers, and mini-newspapers. This system catches this stuff.   

       Like I said, sorting isn't a limitation. First figure out how to get the mail you do want. Then figure out how the bad mail gets sorted in at the other end (post office) so you can sort it back out. Combine the two to get the mail you do want in one pile, mail you don't want in one pile, and the "unsures" in another pile for inspection.   

       //My main problem is that your invention will be huge //[gnomethang]   

       At what point in the idea does it state that the size of device is a requirement?   

       I wouldn't buy this product. It takes out a simple, cheap solution and replaces it with an expensive, compilcated solution. But that is what makes it halfbaked.
Around TUIT, Aug 16 2004
  

       [+] just because it conjures up images of a Wallace & Gromit type invention.
Emi-chan, Aug 16 2004
  


 

back: main index

business  computer  culture  fashion  food  halfbakery  home  other  product  public  science  sport  vehicle