“Help! We went into production and documents are being rejected and throughput is slowing to a crawl. What’s wrong with the product you sold me?”
Providing document management technology to the world’s Fortune 2000 is a challenging effort. Though it is true that managing one company’s documents can be similar to managing another’s, say, for example, a banking company and a shipping company, there are areas where things aren’t similar. This often creates room for surprises.
When we sell our products to a bank, they are most commonly used for what you’d expect: financial documents (mortgage statements, deposit slips, credit card bills and so on). However, as the project gets deployed and succeeds, the bank often looks to further standardize its other departments using the same system. After all, documents are documents. Though these large institutions have learned to be very careful about deploying software without a lot of pre-testing, they can sometimes overlook the need for testing within their own organization when expanding.
In situations like this where unexpected, unplanned and untested scenarios arise, major problems can spring forth. What is important to remember is that not all documents are similar and the load they present onto your processing systems can vary greatly. Small documents impact your systems differently than 10,000 page color documents and international double-byte documents present new challenges that have to be anticipated.
Here are a handful of other irregularities to keep an eye out for…
International documents – Different languages can have a great variety of issues, including accents, alphabets and even direction of reading. You can’t assume that a system that reads Anglo-American text will readily handle Chinese, Japanese, Hebrew or even Norwegian.
Large Documents – Because human beings handle color and black and white interchangeably, it may not be obvious that computer systems today are impacted by color to a much greater degree than black and white. Where a single page, black and white document may take 20 KB, a full color page–whether in PDF or JPEG–can be 1 MB. So a 20 page document can be 400 KB in B&W, but 20 MB in color. Today’s computers can easily handle either one if that’s all you have. But when you’re processing these documents on a server shared by multiple users and/or multiple departments, there are scenarios where you might have 200 users opening up 20 page documents at the same time. Now you’re talking 80 MB for B&W, but 4 GB if they’re all color.
Many corporate servers in service today are not designed for handling images and documents and are only provisioned with 1-2 GB of RAM. This is sufficient for many day-to-day processing needs, but it can be a big problem when working with documents and photos–especially the large color ones discussed earlier. You should also be aware that even if the document appears visually B&W, it may be stored as color and require much more memory. Without sufficient pre-production testing, you may find out too late once you get into production and the server slows down to a crawl.
Very Large Documents – Some companies deal with 10,000 page documents or larger. In a large company, as you incorporate more divisions into your processing center, you may not be aware in advance of these special documents. The problems they can cause are much greater than the large documents described above. The memory requirements are greater and they eat up more processing and bandwidth resources than you’re accustomed to. If your system downloads the full document to the client, you can imagine the time lost compared to a system that only downloads the pages you need.
Unusual Documents – Not all PDFs are created equal. Somehow it is easy to assume that all document formats work the same. That isn’t true. Many PDF documents don’t conform to the full specification – whether they’re PDFs or TIFFs or AFPs or even JPEGs. If one of your departments has been using a system for many years, those kinks may have been worked out previously. However a new system may not handle these files in the same way. You don’t want to find out on the production floor. It is really wise to gather a full range of representative documents from each new business entity that is being included into the corporate document management center.
Yes, it is hard to test, particularly exhaustively. But it’s necessary. To avoid risking production meltdowns, make sure you’ve gone to your user community to find out how they work with their documents. Truly examine as many documents as you can. Look at the workflow process and try to find out how many people will be working simultaneously. Get use cases written down and review them within your department. And remember that not all documents are created equally even though they have the same file extension. There is no substitute for exhaustive testing.