As you know, I scan books. Why is com­pli­cated, and best saved for another day. But over a period of more than five years I’ve scanned more than 3000 books and peri­od­i­cal issues, by hand, on my own time, as a vol­un­teer. And I bought the books myself.

And I bought the scan­ners myself.

In the house we have two Plus­tek Optic­Book 3600 Plus scan­ners. They’re admirable, stolid, slow, pur­po­sive hard­ware. Work­horses, if not very spry or smart. I’ve known for some time, from the ads and the cat­a­log pages over at Plus­tek, that the 3600 was sup­planted by some later mod­els whose overblown ad copy men­tion lit­tle things the 3600 lacks: speed and… well, speed, mainly.

That said, the Optic­Book 3600 Pro has been a good old thing. I can sit and lis­ten to the TV while I scan a 200-​​page book in a cou­ple of hours. But of course I have to run Win­dows to use it, because hard­ware man­u­fac­tur­ers who build consumer-​​grade opti­cal tools for the graphic arts can’t be both­ered to develop dri­vers or util­i­ties to run the damned things.

That’s not strictly true. There is sup­pos­edly a MacOS dri­ver for the Optic­Book 3600. [I won’t even hint at a link to it, for rea­sons you’ll see in a moment]

I installed it when it was announced. Well, tried to. It failed to install, since it attempted to install the KEXT as a supe­ruser, but didn’t man­age to do so even when you ran it using sudo. It crashed while try­ing to do the sim­plest pos­si­ble thing an installer script can do, on a Mac.

In other words, it was released with stu­pid mis­takes that even the most iso­lated basement-​​dwelling novice pro­gram­mer wouldn’t have been allowed to make, if it had been released as free­ware.

Flag that; we’ll come back to it.

So for five years now, I’ve been pleased with the OpticBook’s stolid (and slow) per­for­mance. They still do a thou­sand pages a week with no prob­lems, though there’s a bit of book-​​dust inside the platens by now. But I have to get the Mac­book Pro out, fire up Win­dows 7, run the awful non-​​resizable Book Pilot inter­face, fill out a stu­pid form every time I change vol­umes (you know, with obvi­ous stuff like paths to where I want to save it [no default], what for­mat I want it to save in [no default, and no LZW com­pressed TIFFs], no saved pre­sets… the usual Win­dows crap).

To sum up: Plustek’s Optic­Book 3600 was (and still is) a kludgy and noisy but very work­able solu­tion for the non-​​destructive dig­i­ti­za­tion of books on a small scale. Both small scale in the sense of a few at a time, and small scale in the sense that it only has an A4-​​sized platen.

So yes­ter­day, for my birth­day, my wife bought me a Plus­tek Optic­Book A300.

It’s a lovely piece of hard­ware. Huge, first of all: an A3 platen, so I can start to address the grow­ing pile of quarto and folio vol­umes, and old news­pa­pers and mag­a­zines that I can’t phys­i­cally fit on the other. And fast as hell: nice buffers, obvi­ously faster dri­vers and some onboard stor­age and buffer­ing so I don’t have to take a breath between pages like I do on the 3600. And clean, of course; it doesn’t have five years of fin­ger­prints and infin­i­tes­i­mal book pow­der­ing inside it.

But the software…

Oh god, the fuck­ing software.

Win­dows hard­ware peo­ple, I am talk­ing to you: you do not merely suck at User Expe­ri­ence design. You suck at imag­in­ing what any human being who has not sol­dered the hard­ware together him­self might want. If you wrote this for the gov­ern­ment even they would think it was oner­ous and risky to use.

First of all, the instal­la­tion process was a joke. A 60-​​page man­ual included with the DVD—wait, what, there’s 5 Gb of soft­ware? no, there’s a few hun­dred megs of second-​​rate third-​​party intro sam­ple software—says next to noth­ing in seven lan­guages. The Auto-​​run doesn’t auto-​​run on my VMware machine. There is no direc­tory struc­ture, just fold­ers labeled “Eng­lish”, “Brazil­ian”, “Chi­nese” and so forth. Inside them are raw piles of 8-​​plus-​​3-​​named Win­dows crap, “blah.dll” and “vm7a.exe”. No “read me”, no “READ.PDF”, no man­u­als, no URL files, noth­ing. A few things just popped up a Chinese-​​language dia­log box that said “TODO: [Chi­nese]”, with two Chi­nese buttons.

Finally, I found by trial and error the dri­ver installer. “Plug the hard­ware in, then push ‘Next’…” fol­lowed imme­di­ately by “Well, that didn’t work! Sorry; I got nothin’.”

Just to be per­fectly clear: this is a $1500 piece of hard­ware, and worse if you don’t shop around.

I finally had to just cross my fin­gers and run what­ever seemed runnable. Along the way it tried to install obso­lete off-​​brand OCR soft­ware from years ago, stu­pid also-​​ran Win­dows image “edit­ing” pack­ages that made me puke in my mouth a lit­tle, and… ah yes! A new inter­face app I finally dis­cov­ered, called “Book Pavil­ion” (as opposed to the Optic­Book 3600’s “Book Pilot”).

Yes, each piece of hard­ware has its own user client. Because, you know, you have to write a new client inter­face for each piece of hard­ware. Even though the but­tons all say the same things….

So even­tu­ally it runs. There are some slight improve­ments, in five years, but I still have to fill out a lit­tle tax form for every project I want to scan: What’s it called? How do you want to save it? What res­o­lu­tion do you want? No, you can’t have LZW-​​compressed TIFFs, we’re only allowed to give you JPEG-​​compressed ones. No, you can’t resize the win­dow so you can see a closer view of the fuck­ing huge pages you can scan now; who would want to do that?

In the old, slow Book Pilot soft­ware, every time you scanned a page there was a lit­tle pause while the image was post-​​processed and saved to disk. After all, even a lit­tle A4 page at 700 dpi color is going to end up being a 30 Mb TIF (since you can’t com­press it!).

Not so in Book Pavil­ion! Instead, it just eats up pages, and there’s a lit­tle but­ton in the inter­face you need to push called “Trans­fer” which—check this out—makes the entire inter­face dis­ap­pear com­pletely, with no dia­log box or any­thing, while the images it’s buffered are post-​​processed and saved, and which then pops back with no expla­na­tion of where it’s been or what’s been happening.

So what does one do, first time through? One pokes around—because there is no writ­ten man­ual, no elec­tronic man­ual beyond some Win­dows “help” files, no website—and one sees what one thinks is hap­pen­ing, and so one scans a few pages and “Trans­fers” them, and they look pretty good.

Then one scans a whole book. Maybe 75 huge 12×14 pages, at 600 dpi color of course, and then one pushes “Transfer”.

And the fuck­ing thing dis­ap­pears for­ever. Because 75 huge 12×14 pages at 600 dpi color are more than you’re sup­posed to want to do, as far as this user inter­face is concerned.

Even though the inter­face has been let­ting you poke the but­ton on a fuck­ing large-​​scale 600dpi opti­cal res­o­lu­tion 24-​​bit color scanner.

So you lose that work, because you have noth­ing so fancy-​​dancy and mod­ern as a dia­log box with a progress bar or any­thing to explain what’s up, and then you re-​​open the util­ity again man­u­ally before it mag­i­cally reap­pears to tell you it might be done. Well oops. There are no saved temp files, no buffered images (even though while you’re work­ing it tells you they’re there). There’s no popup ask­ing you if you want to add pages to the same project, or start a new one. No con­ver­sa­tion about what you want.

Just a form to fill out, which will mainly be ignored.

Harsh? No, because of course you crashed the piti­ful lit­tle user inter­face, so it never both­ered to save pref­er­ences or cache its set­tings. So the next book you scan (after re-​​scanning the orig­i­nal one), you acci­den­tally save at the wrong res­o­lu­tion, in the wrong place, in the wrong format.

Then you start to get mad. Because man­u­ally scan­ning $100 books, even on a $1500 piece of hard­ware that’s made to be min­i­mally dam­ag­ing, dam­ages the books if you have to do them twice. So on the whole this is going to be a very expen­sive and annoy­ing set of “con­ver­sa­tions” you have with the stu­pid user inter­face before you agree to disagree.

So here’s the rub: Plus­tek seems to be mak­ing excellent—maybe even amazing—hardware.

I can imag­ine how good they would be to use on a real com­puter, with a real user inter­face. If I am very very care­ful I can man­age to scan a book all the way through. I have to remem­ber to check the set­tings every time. I have to check them again, because some­times the stu­pid lit­tle tiny text fields can’t show you the salient part of the infor­ma­tion you’re putting in them. I have to remem­ber to push the “Trans­fer” but­ton often enough that the invis­i­ble buffers don’t over­load. And of course I still have to LZW-​​compress my TIFFs after­wards, and run real OCR and image edit­ing soft­ware on the Mac.

The thing that makes me sad, frankly, is that these peo­ple are clearly smart. This is a nice box. It makes nice sounds. It’s quick, and light, and it worked imme­di­ately as soon as I jumped through the right hoops.

But their soft­ware is a design sin. It’s almost like they’re sell­ing an Aston Mar­tin that only shifts gears if you blast Vanilla Ice on the radio. No, worse: because at least in that case the auto­mo­tive engi­neers would have clearly done it out of spite and evil.

This is just thought­less banal­ity.

Plus­tek: Spend the fuck­ing $50000 to hire a pair of real soft­ware devel­op­ers, take three months, and write a real adult piece of soft­ware so we can all actu­ally use your lovely hard­ware, just like the Big Boys do.

Plus­tek: Spend another $50000 for a real User Expe­ri­ence team. Sit them down, talk to peo­ple like me who buy this stuff for our­selves (instead of hav­ing some cor­po­rate or insti­tu­tional boss whose back we can com­plain behind), and make it work nice.

Spend that $100000, and you will fuck­ing nail the mar­ket for per­sonal scanners.

These machines are lovely. But they suck to use. Epson’s soft­ware sucks less. Canon’s soft­ware sucks a lit­tle less. Even Nikon’s soft­ware sucks a tiny bit less, but at least it sucks on the Mac.

As things stand, you should be embar­rassed to the point of apol­ogy. Your mar­ket­ing peo­ple should be sent to re-​​education camps. Your “soft­ware devel­op­ers” should be… well, you don’t have those, clearly. The nephew who can run Visual Stu­dio should be sent back to college.

I’ll keep the damned thing. It’s an invest­ment of five years of my life so far in the pre­vi­ous clunky effort, and it cost my wife a whole day’s wages to buy this one. So I’ll play your Book Pavilion’s lit­tle rit­ual games, and I’ll chuckle wryly when it plays hide-​​the-​​scans or woops-​​I-​​crashed with me.

But you will lis­ten to me the whole damned time.

Because I feel cheated. I expected more, after five years.

You just don’t give enough of a damn to even try.

I want the best for you. Those librar­i­ans are bitch­ing just as much as I am, behind your back. The cor­po­rate records-​​scanners are bitch­ing just as much. The cen­sus scan­ners, just as much. But they are bitch­ing against a back­ground of bitch­ing about their work, their jobs, the fact that they have to use Win­dows, their gray lives.

Some of us choose to buy these lovely things from you. Because we want to make a bet­ter world, and we’re will­ing to put our own time and money on the line.

You’re cost­ing me that time and money.

Stop it. You could win, you could soar at mak­ing scan­ners for archivists and librar­i­ans and edu­ca­tors and stu­dents and schol­ars and his­to­ri­ans and artists and geneal­o­gists and accoun­tants and pirates and kids all over the world.

If only you could stop being so blithely idi­otic about the software.

So close. So much poten­tial wasted.