Decent way to scan hundreds of kilometers of receipts and register tapes?

[gomipile]

I know someone who would greatly benefit from digitizing their records. However, among the records that need to be kept indefinitely are 1000+ km of register tapes and receipts.

Is there a decent way to scan that much? Is there something like an automatic document feeder that can handle multiple 90 meter rolls of register tape without human intervention?

Also, the smaller business-to-business receipts are of varying sizes on curled flimsy paper, many with staple holes in them. The automatic document feeders we've tried choke on them. There are tens of thousands of these.

Any suggestions?

[Eldan]

I've worked in some companies that had Archives to digitize, and I think the answer is "no". Or at least, "not really". The most cost effective way was Always to hire some housewives and students to do it by hand.

[halfeye]

I've worked in some companies that had Archives to digitize, and I think the answer is "no". Or at least, "not really". The most cost effective way was Always to hire some housewives and students to do it by hand.

That will be fault prone, best to key everything twice and check for differences between the two keyings. It will be difficult and very expensive.

[Eldan]

Oh yeah, absolutely. Quality Control of the scans is an entire second layer, but I thought that wasn't relevant to this question.

[tomandtish]

As others have said, there's no easy way to scan these in. Time and effort is pretty much the only solution when it comes to scanning.

Now, what MIGHT be a bit easier is taking pictures of them and storing them that way. You could probably get several receipts in a picture. You'd just have to make sure your naming system let's you find them if/when needed. And then do scanning going forward.

[Tvtyrant]

My assumption is the easiest thing is to print out a bunch of barcodes, put the item on the coded paper and take a picture. Then just name the barcode sequences in your system, easy peasy.

[gomipile]

As others have said, there's no easy way to scan these in. Time and effort is pretty much the only solution when it comes to scanning.

Now, what MIGHT be a bit easier is taking pictures of them and storing them that way. You could probably get several receipts in a picture. You'd just have to make sure your naming system let's you find them if/when needed. And then do scanning going forward.

My assumption is the easiest thing is to print out a bunch of barcodes, put the item on the coded paper and take a picture. Then just name the barcode sequences in your system, easy peasy.

That might work with the B2B receipts, since they're all less than 50cm long. The images would still have to be cut apart manually to OCR for each separate document. And the focus on the camera would have to be close to perfect. I suppose with enough light and a small aperture setting, and some sort of rig that lights the receipts with no glare... I'd essentially be reinventing the flatbed scanner to get a flatbed scanner with a square area close to one square meter.

It seems like for the 90m tapes I still don't see a solution better than paying people to operate receipt scanners. Which would mean it wouldn't get done, despite the benefits. Most of the benefits are hedging against future risk and saving time in the future. There's no direct monetary profit to be gained that would offset the expense of scanning in any sort of labor intensive way.

[Tyndmyr]

I know someone who would greatly benefit from digitizing their records. However, among the records that need to be kept indefinitely are 1000+ km of register tapes and receipts.

Register tape ages fairly quickly. A couple of years in storage, and it may well be illegible, depending on exactly how it was printed.

Depending on how long it has been stored, it may already be a non issue.

As for the rest, I'm sorry to say that it is harder to automate than one might think. OCR isn't flawless. People aren't flawless. If you automatically push records electronically when making the sales/etc, the situation is not so bad, but digitizing vast quantities of paper documents is generally pretty tedious indeed. It can absolutely be done if you want to hire a bunch of temps to key it all in, but it's going to cost.

I would suggest reconsidering how bad the financial implications of loss of data actually are, and if they are low enough, just accepting that the information is effectively already lost.

[gomipile]

Register tape ages fairly quickly. A couple of years in storage, and it may well be illegible, depending on exactly how it was printed.

Depending on how long it has been stored, it may already be a non issue.

In this case, these are legally required to be kept as records for a time that may be longer than the deterioration time. So, deterioration of the tapes makes the preservation more urgent, not less.

If it is found that some older ones are illegible, reconstructing the records will be required by law in some cases. That's a completely different conversation than this thread, though.

[Eldan]

From what I've seen in our lab archives, thermal paper might still be readable after 20 years if carefully stored (dry, cool and dark), but it's dicey. Very little chance after 30 years. Which is why it was such a huge pain in a pharmaceutical lab, where it was legally required to keep everything for lifetime of product + 30 years.

Edit: from googling, it seems that different suppliers guarantee 7-10 years.

[halfeye]

In this case, these are legally required to be kept as records for a time that may be longer than the deterioration time. So, deterioration of the tapes makes the preservation more urgent, not less.

If it is found that some older ones are illegible, reconstructing the records will be required by law in some cases. That's a completely different conversation than this thread, though.

This is probably going to cost millions or maybe billions (this is an approximate estimate, it should be possible to work it out to about an order of magnitue), I would suggest (I am not a lawyer or accountant) that somone checks with the owner whether declaring bankruptcy is a better option.

[Manga Shoggoth]

I don't know how far it will go, but the old Logitech Scanman series could scan a long, narrow (up to 5in wide) sheet. You would be able to scan more than just A4 lengths, but I don't know how far. 90M might be impossible in one go, but you might be able to break it up a bit.

I used to have the model linked, but that one won't be much help (apart from being on ebay, it used an custom ISA card, and I don't think the drivers went much beyond DOS/Window 95), but the newer models may be of some help.

I have also seen a couple of sheet-fed scanners that are designed to scan recipts, but it isn't clear how much length you would get out of them. The Fujitsu SP-1120/SP-1125/SP-1130 series apparantly handles up to 3M/120"

Not seen anything that will do a full 90M roll, though. A dedicate scanning service may be able to help (at a price, of course).

[J-H]

I would reach out to several local companies in the imaging and printing industry. Most copier dealers are big on MPS (Managed Print Services) and scanning now.

[Algeh]

First, set up a system where new records are digitized as they're created or collected (or, in the worst case, as part of some kind of weekly/monthly/etc. archiving process if there's no way to build it into the creation process). This means the backlog won't be getting larger while you're working on fixing it.

Then, research if "indefinitely" has some kind of end. Pretty much every field does have an end date for how long things need to be stored. (For example, personal income tax records are 7 years in the USA, and the lawyer I used to work for's client agreement was that she'd keep records for 10 years. At school, we have to keep gradebooks for five years.)

If such an end date can be established, figure out when the various existing records have or will age out of it. Then look at how degraded the oldest records you still need to keep are, and sample the level of degradation from various newer records. If records about to fall off the "retain date" are still generally legible enough to use, then start your digitization project with the newest records and gradually work your way back in time.

Set aside the amount of person-hours you can afford to this project and work on it gradually over time. If possible, make this what someone does on a specific day or the week or month. You will never get around to digitizing the oldest records you currently need to retain, but some of them will fall off of the "retain date" cliff each year and so if you needed to keep, say, 10 years of records and are able to digitize your new records in an ongoing way, plus scan 2 months of backlog each year, in a year you will have gone from 10 years of un-digitized records to 8 years and 10 months of un-digitized records (1 year from scanning new things as you go, plus the two months from scanning the most recent backlog). In another year, you'll be at 7 years and 8 months, and so on. If you can manage to scan a backlogged year each year, it would of course go much faster.

If the oldest ones you're supposed to still have are too degraded to read (meaning your current document retention scheme is not leading to compliant records being kept in addition to not being digitized), then the choices get tougher. You have to weigh how much harder it is to preserve the most degraded ones you're supposed to have versus how much worse it'll be for you if you don't try, with consideration of both how bad it'll be if you need them and how likely it is that you'll need them.

Assuming that the pattern is mostly going to be that newer stuff is more likely to be needed, I'd probably pick the oldest year of stuff that mostly hadn't degraded enough to need special techniques to preserve and scan that year first, then work my way newer until I caught up before looping back and trying to resurrect the oldest things. Sure, I'm in trouble if the older stuff is needed since I'm letting it get worse this whole time (if there are simple steps you can take to keep it from getting worse in storage, do them!), but in the event that I need those records I'll have to pay specialists to give it a try and hope for the best and at least I saved what I could so the problem is getting smaller and older with time. In general, people are going to be more understanding about not being able to locate a document from 50 years ago than 5 years ago even if you're supposed to have saved them both, and particularly so if you do have the 50 year old one but it's just really faded and illegible.

Finding out how other similar businesses that need to keep things this long handle it is also a good idea. Generally "I did it the way that is standard in my field" is a good shield against a lot of compliance issues even if it seems like a stupid or inefficient way to do things on its own. Talking to a records storage company used to the standards for your industry could also be a good move. I assume a lot of the places where you can send records for off-site long-term storage also offer digitization services these days.

This is all assuming that you're keeping the records for some kind of legal compliance reasons rather than historical or sentimental interest. If you need to archive things for historical interest, you should consult archivists working in that field to see what's important to them so you're preserving the right information about the documents when you digitize them. If sentimental interest, then you would want to focus on which things carried the most emotional weight first rather than chronological order.

[halfeye]

Please let us know in broad terms how and whether this was resolved.

[gomipile]

Please let us know in broad terms how and whether this was resolved.

No decisions have been made about this yet.

[Bohandas]

edit: oops! wrong thread.