Friday, September 05, 2008

For sale: one filing cabinet

Over brunch at University Cafe.

RB: "Did you get my card?"

Me: "Oh, I haven't gone through my mail in a month. It takes so long to file the papers into my filing cabinet."

RB: "Why don't you just dump them into a shoe box?"

Me: "I file them into individual folders for bills, insurance, mortgage, etc."

RB: "How often do you go back to look something up?"

Me: "Almost never. Maybe once a year."

RB: "So you do frequent writes, and infrequent lookups. Why would you choose to pay the cost on write? Just dump everything into a box and then sort through it if you ever need to look for a particular item! You would never design a software system this way."

Me: [stunned into epiphany]

41 comments:

Wes said...

I actually have one garbage bag with all my documents in it sitting in my closet. The system works surprisingly well. :)

Anonymous said...

Lol.

I do dump them all into a shoebox, (one for bank bills, one for investments & tax, one for fuel bills, one for medical bills) and when I do have to lookup, I feel it too much load on sort and figure out just one document, I was looking for.

I'v always thought, I should the more organised and file them properly etc...blah.

Yishan said...

That's actually exactly how I think about my filing system design.

Anonymous said...

I used to use the shoebox method because I'm lazy. But the latency to look stuff up is agonizing. You underestimated the lookup usage and the discouragement of high latency to do anything with these papers.

Fortunately, I acquired a "secretary" who uses a file cabinet to store the papers in sorted folders, which is really not too much hassle if you do it whenever you have incoming papers. Since then, every tax season, we got refunds long before people started their map-reduce jobs.

Unknown said...

Well, unless you are a computer and have no feelings and emotions, I think a lot of us, me included, like to have a certain level of comfort that we can access our records easily effortlessly at any time.

So unless you yearn to become Data (and he yearned to have human emotions and feelings), take the classified ad off for your filing cabinet.

The human in you will appreciate it. Don't let the Google Borg assimilate you. Resistance is not futile.

- Spock

konberg said...

Suddenly I feel better about myself.

Bob said...

Niniane I do the same exact thing. I, however, will keep my filing cabinet.

As for RB's logic, I don't really agree. A lot of effort is put into search creating efficient search algorithms, just so people can quickly access that one piece of information that was stored away at some point in time. Your company for example spends a lot of money and time trying to retrieve data faster than your competition.

So, I say sell that file cabinet and use the money to get a much bigger one. :)

Anonymous said...

Scan, shred and search it with Google.

If you can't already do this...then this is your next 20% project.

Noah said...

hmm ... there's electronic billing over at your side ? If you sign up for e-bill, you can file them virtually and maybe google through when you need something ?

Anonymous said...

Use the best attributes from each system. Take a standard four-drawer filing cabinet and label each drawer with the name of a general category (bills, statements, personal, junk, etc). If present, lock the internal drawer partition to the size of one 8.5" x 11" sheet of paper. Place a stack of 48 hot pink sheets of construction paper or cardstock in a standard wooden office tray on top of the cabinet. When your mail arrives, throw it all into a single basket and carry it to your filing cabinet. Open the cabinet drawers in a cascade from top to bottom (with the bottom drawer fully extended, to maintain stability) so you'll have access to all drawers at the same time. If this is the first time you've deposited mail this month, throw a pink sheet into each drawer as a separator page. Now, start grabbing pieces of unsorted mail from your basket and throw them into the appropriate drawer until your basket is empty. You get timewise sorting for free, and on the rare occasions when you do need to find a particular item, you can drill down past the pink sheets that mark approximate monthly boundaries (and if you wanted to improve on this, print the month in large letters on each sheet and laminate it, for reuse each year). Hopefully, your friends won't mistake you for Ms. Monk...

Anonymous said...

"You underestimated the lookup usage and the discouragement of high latency to do anything with these papers."

Agreed.

But on the other hand, RB reminds me of that classic Dilbert cartoon about the records retention clerk.

Also, if you hire an accountant, you can fob off the delayed filing onto them for a fee. I think.

Finally, have you considered going through your non-financial mail in a separate pile?

Anonymous said...

I really like siphonophore's solution. But test it out before you buy, cause some filing cabinets have the safety feature of only being able to open one drawer at a time, for stability.

Anonymous said...

Download and set up prsnl and write a Google Desktop plugin for it. :)

Rusty Rider said...

I feel like there's a missing requirement here. When the filer/searcher needs an item once a year, how urgently and quickly do they need it? I would think that data point would be needed to decide what level of pre-sorting you do.

p.s. You can take the dork (me) out of computer science, but you can't ever take the computer science out of the dork...

Anonymous said...

I have a hybrid approach. Throw everything into a shoebox, but whenever I need something from it, I pick up each document one at a time and separate them into relevant piles until I find what I'm looking for. Then I put them all onto one stack again (maybe with some dividers if there are any within arms reach at the time), with the unsorted documents at the bottom, and put it back in the box. Next time I open the box, it's partially easier to find things. Each repetition gradually cleans things up but costs me no extra time.

Mark said...

here's what i do:

1) this scanner is awesome. it'll scan a stack of 20+ irregular sized documents, both sides, run OCR & place a searchable PDF w/ the date in the filename into a predefined directory. this step takes about 10 seconds of my time.

2) i then put the original document into a box.

3) every once in a while i take everything in the box and put it into a big manilla envelope and label it with a data range. (e.g. "April 2008 - September 2008."

if i need a document, the searchable PDF's are nearly always sufficient. if not, I can always track it down by looking in the appropriate envelope.

this method saves me from having to search the stack should i find that i need something, and the up front time cost is minimal.

the method does require spending $400 on the scanner, but in my opinion, it's well worth it.

IvyMike said...

If you place them in the shoebox in chronological order, they *are* presorted in a way that is often useful so even the lookups aren't that expensive.

For example, it's not difficult to find the last ten months of electric bills just counting back a month at a time. After you find the first electric bill, you know the approximate day of the month they arrive, and since the stack is chronological, it's easy to find bills on that date from previous months.

Anonymous said...

I stuff all mail in a pile which I sort through into folders about once a year. And I almost never go through it … you've given me reason to dump the sorting part. Thank you :)

Anonymous said...

Take a look at the Noguchi filing system: http://communicationnation.blogspot.com/2005/12/noguchi-filing-system.html

It's basically a LRU system.

Anonymous said...

For any bills that have to come via snail mail, have them sent to Earth Class Mail, download the PDF, and delete them from ECM. Why keep old bills?

Oh, and you are beautiful! ;)

Anonymous said...

do earthclassmail (i DO NOT work for them) --- they scan snail mail for u - ... then search with goooooogle and .... ur done.

Anonymous said...

I have an accumulator-based system -- a "to be filed" pile that I tackle every couple of weeks, two months tops (else the backlog is too much work). I then go through this stack, and distribute each of the items into the appropriate folder in the filing cabinet.

Each folder is more-or-less unsorted; when I need to find something, I then sort the contents of just the one folder as I look for what I need.

Once a year, I take the thickest couple of folders, sort 'em, and pull out everything older than January 1 of the previous year, and that gets archived (if there's a reason to keep it) or shredded.

Nothing is religiously sorted until it needs to be, but neither is everything so disorganized that the task of finding a document becomes overwhelming. And because most of the time you need a document quickly (because something's gone wrong or you would like to double-check something suspicious), the additional stress from getting everything organized on top of whatever stress that is resulting from your need for a document can easily result in a meltdown.

Meltdowns are critical failures. It's important to take this into account when computing the amortized cost of a filing "system".

Put-it-all-in-a-box-and-sort-when-you-need-it seems akin to invoking the garbage collector only in an interrupt handler. Or flushing the disk buffers only when the system detects a loss of power. Or only performing system maintenance when the auditors show up.

Chinmay Kulkarni said...

So many opinions and advice on how to do it right :)

However, I think the assumption is wrong to start with. This is not a case of infrequent reads and frequent writes. It is one that requires hard time guarantees on reads; besides zero false negatives. (That is, if you don't find an item in the hard time deadline, you can't simply say the item doesn't exist).

Of course, whether you do a bucket sort; or a chronological one is a choice.

Btw, I have similar problems putting my wardrobe in order. Frequent reads, writes, and deletes, and hard-deadlines. Any thoughts?

Anonymous said...

over the years i've developed this system:

i save my receipts, then put a checkmark on them when i enter them in my checkbook ledger (a google spreadsheet). then i put them in an envelope.

at the end of the month, i label that envelope with the month and year and put it in a shoebox. finding receipts is trivial because i can cross-reference the spreadsheet to find the month/year.

it's fairly likely i'll be looking for something once a year, and sorting through an entire year's worth of stuff is a lot of work. the amount of work for my system is extremely minimal.

Fabien said...

I use the "Is this needed?" method.
When I receive a document, I think seriously about whether I need to keep it or not. If so, I put in on the stack; if not, I do whatever I need with it and then discard it.
That way, I don't have many papers, so the stack stays a good sorting method (i.e. easy retrieval of documents) for several months.
Twice a year or so, I spend half an hour actually sorting the documents into folders, once again making sure I discard whatever isn't useful any more.

Anonymous said...

Coolness, I like filing cabinets!

Jiff
www.anonymize.kr.tc

Anonymous said...

I put large papers in one folder for each year, 2006, 2007, 2008, etc, and receipts/small papers in one folder for each year. If I need something, I spend the time to go through the stack. It's worked well for many years.

Japherwocky said...

I get emails sent to relevant addresses at relevant domains.

Sort once. When you sign up.

wtf sends mail? :)

Anonymous said...

wish I can "/" or goog my shoebox :)

Anonymous said...

Whoa, girls hot. And she went to my college 10 years ago.

Aww, and you probably won't approve this message.

Anonymous said...

I think Iwantofbeingme's solution works better. Just like a hashtable!! Fast lookup and organized too.

Anonymous said...

I'm terrible about keeping up with filing, and almost never have to retrieve papers that I've previously filed. I finally found something simple that works quite well:

1. I put papers that don't require permanent retention into the appropriate partition of a 12-month expandable folder. For example, this month's routine bills go into the September partition.

2. At the beginning of each month, I empty that month's partition -- which contains papers from 12 months back -- and shred its contents after quickly scanning them to see if there's anything I should keep permanently. For example, earlier this week I shredded the documents from September 2007.

That, plus annual folders for tax papers, and a folder for proofs of purchase (for warranty purposes), and I'm pretty much set.

Anonymous said...

Wow, lots of options. Good! :)

Chinmay:

Use clothes hangers with an embedded RFID circuit, a bit of flash memory and an attached microcontroller with a red LED strobe (all powered inductively by the closet, or with lithium batteries or rechargeables). During, um, "writes" the article of clothing is RFID scanned, digitally photographed, categorized and optionally labeled before being hung in the closet (thus, each hanger has some idea of what was hung on it, allowing a computer to build up a profile of the contents of the closet in realtime; of course, there would be optional caching for the sake of speed so long as you dealt with the requisite sync issues). For retrieval, select an article of clothing using a laptop (either visually, or by category and/or label) and the hanger for the item/s will begin flashing. For meeting hard deadlines, position the requisite number of friends along equidistant points across the length and breadth of your closet to guarantee minimum interrupt latencies, taking into account their respective fields of view. :)

Anonymous said...

I can tell who has been audited by the IRS!

I have found it of great comfort to be able to go into my files (of 15 years ago) to find the record of my having paid income tax.

Before you scream that the limit is 7 years, consider that there is no time limit on how long an IRS fine can accrue. According to the IRS, I owed them many ten's of thousands of dollars in interest and fines. Since I had proof that I filed the return, they set aside all the charges and closed the case.

In another case, my business was audited by the IRS. At the hearing, my accountant (a former IRS agent) took the contents of a big box of receipts out and put them on the IRS agent's desk. They were all nearly organized and caragorized in separately bound stacks and file folders.

When the accountant was done laying the receipts out, she crossed her arms, looked at the IRS agent and said "We will be expecting a positive adjustment." And, we got a bigger refund, and the case was closed.

Also, there is what seems to be my regular need to go back into my files and pull out a specific receipt. It is easy to do that since everything is so well organized.

Finally, there is the yearly grind of filing taxes. Filing the taxes is time consuming, but having everything organized in files makes it much easier. I usually get a healthy refund because I have the receipts to document my itemized deductions and my accountant can make quick work of filing my returns (business and personal.)

If I had to sort through a garbage sack full of receipts, I'd likely never find what I was looking for.

For those who are ADD challenged, consider hiring professional organizer to help you keep things organized by helping you organize a filing system and helping you to keep your receipts filed. Final an organizer who is patient and kind. You do need one who is anal about being organized, but they need to be relaxed enough not to impose themselves on you in an overpowering way.

Jonathan said...

This reminds me of how I do my laundry. My ex-girlfriend would always want to turn each sock right-side-in and fold each matching pair of socks together, but that takes more up-front effort and time. I just take matching pairs of socks and roll them into a ball, regardless of whether they are reversed. Later, when I want to put them on, I just unroll 'em and make sure they are right-side-in.

Anonymous said...

Wow.. I never realized that my lack of organization was in fact optimal.

Made my day :)

N said...

re: Jonathan.

I'm not sure if the analogy holds. Presumably you wear the socks that you wash, so the same cost is paid on folding or paid on wearing.

Anonymous said...

What would be the optimal configuration for a closet? A circle with lit crenellations for non-hanging items, all articles within one hop of the designated center? Or maybe a spiral, so you can see more of your clothes from the side, instead of edge-on? Magnetic hangers and coated panels would let you shingle a wall with your outfits, for viewing the entire collection at a glance. Jesus, I think I need to get back on decaf...

Anonymous said...

I use the sort and trash it method...throw away/shred bills after they are paid (I always have the CC or check blank for a receipt), anything extra I toss it after I am done, usually within an hour of receiving it. Anything that may be needed for taxes or kept for awhile... I place it in a small folder in a desk cabinet thing that I go through every few months and toss out what is not needed. I really only keep 4-5 pieces per month so theres not much to sort through.

Unknown said...

I store incoming correspondence in a red/black tree. Admittedly, this tree is horribly unbalanced, since most documents are printed with black ink, but that is made-up-for by the fact that documents with red ink are usually high priority (Third Notice, Final Notice, etc.)

Documents printed with other ink colors (blue, green, etc.) don't fit into the tree and are immediately shredded (sorry Grandma).

This system works equally well for your wardrobe, as long as you can get away with only wearing red and black.

Anonymous said...

40 comments on this? really? YOU FUCKING NERDS!!!!!