The dream of the paperless office is just that - a dream. Even though we can receive faxes electronically, transmit documents over networks, and broadcast memos via e-mail, paper continues to dominate much of what we do.

Paper is portable, easy to read, and doesn't run out of battery power somewhere over Indiana. The reality is that the first thing most of us do when we receive an electronic missive is print it. I met a fellow on a plane a few weeks ago carrying an enormous stack of paper. It was all his e-mail from the weeks he was away from the office. He printed it so he could study it on the trip!

Managing the huge preponderance of paper flooding offices is becoming increasingly critical. One solution, believe it or not, is to convert it back to electronic form so it can be filed, searched, faxed, edited, and otherwise electronically manipulated. Although traditional flatbed scanners can do the job, for offices on a budget, a new class of small, inexpensive scanners may be a decent solution.

One candidate to consider is Umax's PageOffice, a handy 12 x 6 x 4-inch desktop black-and-white scanner that can, via its optical character recognition (OCR) software, turn stacks of memos, reports, and other printed business minutiae into PC-readable ASCII text. It can also capture line art and 256-grayscale images for in-house publishing projects. And the PageOffice is darn fast. The downsides: It's limited to scanning about 10 single sheets at a crack (books and oversized documents are out), and worse, the scanning, OCR, and image-manipulation software have some major usability problems.

Installation

PageOffice comes with a "switchless" SCSI card, which means you won't have to fiddle with jumpers to set DMA or interrupts. The card works great - as long as you don't have another SCSI card in your system. On two different test systems with existing SCSI cards, the Umax scanner could not be recognized by its software, and eventually, locked up the system. The manual fails to mention that if you already have a SCSI card in your system, you can simply add the Umax scanner to the chain, and it works fine. You discover this interesting tidbit via a dialog box when you install the software. This is just the first of many problems with the manuals and software.

The Software

Home base is Umax's PageManager software. You get a toolbar across the top of the screen, while icons for various other tools stretch across the bottom. These utilities let you take a scanned document and send it via e-mail, fax it, or print it on the spot (both Microsoft Mail and cc:Mail are supported); edit the scan with a mini-Photoshop-like program; run Umax's OCR program to convert the scan to text; add text to a nonconverted image; or add a scanned page to a "stack" (Umax's metaphor for organizing pages). Most operations are drag-and-drop. For example, to e-mail a scan on the spot, drop the thumbnail image of the page on the e-mail icon and you'll be prompted for a destination.

Scanned pages are kept in stacks. For example, you might keep all the scanned pages of a contract in one stack, and family photographs in another.

Clicking a stack displays thumbnails of all the scanned pages. You can click and drag documents between stacks, and start a new stack by dragging a document thumbnail to an empty portion of the stack area (which is not particularly intuitive). Double-clicking a thumbnail gives you a full-screen view of the document. You can then crop and otherwise manipulate the image.

In the first of many usability glitches, you can use the generic pointing tool to select an area of the scanned document, then select Edit*Crop. But if youselect the actual cropping tool, the Edit*Crop option isn't available - you must right-click to do the job. (A little trick poorly documented in the manual.) Another hassle: If you minimize PageManager, it closes all the stacks, so you must reopen the stacks you were working on when you restore PageManager. You don't lose any of your changes, but it's a pain.

Do You Scan It?

Scanning pages is simple. Just slip up to 10 sheets into the paper input tray, and press a button on the scanner or click a software option. The sheets feed through the scanner, and a thumbnail of each newly scanned page appears in the current stack. PageOffice automatically recognizes the width of the page, and sizes the results to match. Unfortunately, it had problems recognizing a standard 8.5 x 11-inch page, and the resulting scan captured only a narrow strip of text. Turn off the "auto-recognize width" option and it works fine. According to Umax, this hardware glitch has been corrected in later units.

As desktop scanners go, the Umax is pretty speedy. It can scan a document page or line art in 12 seconds, a photo at 100 dots per inch (dpi) in 6 seconds, and a photo at 300 dpi in 18 seconds. Accuracy in converting a scanned document into readable text isn't exactly stunning (about 85 percent), and that's when scanning pages using so-called normal fonts like Times Roman, Helvetica, or Courier. Fancy fonts like Balloon, Caslon, and Gallery totally defeated the product, but that's hardly unexpected.

Once a thumbnail of a scanned page appears on the desktop, it's yours to manipulate. As noted earlier, just drag it to the desired tool and you're prompted along the way. Drag a thumbnail to the fax tool and a dialog asks you to enter the name and fax number of the recipient; the name and address can be stored in a phone book for later use. You can't import this information from standard databases, or even from popular fax programs, such as WinFax or FaxWorks. Dumb.

Converting a scanned page to text is fairly automatic. You can also convert just a selected portion of the page. Unfortunately, once you do this, you can't go to another section and convert that; you must first clear the software of the previously scanned text. Exporting scanned text is just as problematic. With only the Umax software, Dashboard, and Word for Windows 2.0c running, Windows refused to copy the scanned text to the clipboard due to low memory errors. This one is a puzzler - according to Dashboard, I had plenty of system resources left, and my 16MB machine wasn't maxed out.

Filing It Away

Umax's file-management software organizes scanned pages into a hierarchical structure of cabinets, drawers, folders, and documents. (Only cabinets and documents are required.) One window displays a File Manager-like tree, where you move pages around by clicking and dragging. A second window displays the thumbnails of the currently selected document, which you can double-click to view at various magnifications. You can also attach annotations to a scanned document, and categorize it by attaching a marker, such as Urgent or Personal.

When you drag a page to the file-manager icon, the page ends up in an in box; you can then attach it to an existing document or create a new document attached to an existing folder. Unfortunately, adding pages to an existing document is rather confusing, thanks to a poorly designed dialog box. To print a document or save it elsewhere, you must drag it to an out box. It's a workable system, but it's hardly elegant.

There are lots of little problems with Umax's file manager. For one thing, it doesn't install its DLL files in the working directory, so it won't run until you manually specify the working directory in Program Manager. Another problem: Identical File menu options work differently depending on which window (file-management or thumbnail) is selected. You can drag multiple stacks into the Umax file manager from PageManager, but the program ignores everything except the pages in the first stack. The only way for the other stacks to be seen is to add them - one page at a time - to the in box.

Modifying a Document

You can modify a scanned document two ways. With the PageType tool, you can add text to a document, and change the font, effects, kerning, and so on. One neat trick: If you don't convert the scanned page to ASCII text (if you treat it like a picture), you can add to the page fields that contain editable text. This makes it easy to quickly modify documents like contracts or to fill out forms without converting them to ASCII text. And it eliminates the possibility of Umax's OCR program introducing errors - a disaster with legal documents.

Unfortunately, PageType has a nasty bug that can trash your work. When you save your work, the program places it in a temporary directory. However, when you exit PageType, it empties this directory, wiping out everything you saved. Earth to Umax: This is not a good feature.

The other tool, PageImage, lets you do mini-Photoshop-like modifications to a scanned image or any PCX, BMP, or JPEG file. You can draw objects, adjust colors in specific areas, and apply special effects. Unfortunately, when I opened JPEG files, it usually resulted in a General Protection Fault.

One More Time

PageOffice is a great idea at a great price - but it's not ready for prime time. There are enough bugs and rough spots that I can't recommend it to even fairly experienced users, although a PC veteran could sort things out. The four manuals are helpful but poorly organized; the early sections assume knowledge that isn't explained until later. An entire page is devoted to things you should do before calling technical support - but the tech support phone number isn't listed! Once Umax cleans up the bugs, straightens out the manuals, and streamlines its software, it could be a great way to beat the paper chase. Until then, steer clear.

© 1995 David Plotkin. All rights reserved.
David Plotkin is a business analyst at Integral Systems in Walnut Creek, California, and the author of Using Lotus Approach for Windows (Que). He can be reached as DPlotkin on America Online, dplotkin@aol.com via Internet, or care of Computer Currents.