In my professional life, I advise clients on document processing. This may involve document management systems (DMS), content management systems (CMS), web forms systems, and document capture and storage solutions. And in this niche, I see a place for an iPad - or a similar device.

One of the DMS products I work with, Autonomy’s Interwoven WorkSite, has for many years had a web based interface, enabling users to search for, retrieve and store documents in a repository with a web browser. This is a feature found in many modern DMS implementations. And while this web interface would be a natural for tablet devices like the iPad, as a practical measure it is not currently possible, since browser support is limited to Internet Explorer. But Autonomy is a smart company, and I would not be surprised to see them add support for some other browsers, such as Safari.

But that’s not the end of the story.  Last May Autonomy showed it’s iPhone client (iTunes link) for it’s WorkSite DMS. They previously had support for Blackberry devices through their Mobility Server add on, but with the iPhone client, they extended document access via Mobility Server to the iPhone.  Since the iPad runs iPhone apps natively, and Apple is touting the “easy” transition of iPhone apps to iPad’s bigger screen, I expect an iPad app from Autonomy relatively soon.

So what does this mean?  Imagine a corporate meeting where the participants, instead of using a three ring binder filled with paper documents, use iPads connected to the corporate DMS. After the meeting ends there are no stray paper copies to pick up, to shred, or otherwise worry about.  Or imagine a courtroom, where the litigators refer to documents on an iPad, instead of boxes filled with paper. (Note that there is a big use of laptops in this setting already, but it’s rather awkward to hand a laptop to a witness)

Still, both of these scenarios have gaping voids. Document annotation on an iPad is an unknown. Without annotation, the corporate meeting scenario would be limited to review only. Changes would be difficult. And since the iPad, like the iPhone, does not support multitasking, running a secondary app - even something as benign as a calendar, would be impossible. And the iPad lacks a stylus. This makes collecting a cursive signature impossible, and limits the ability of the iPad to process some web based forms.

In the end, though, the possibilities for future advances in document management that the iPad represents is inspiring. I’m optimistic that a few years from now, those of us working in the DMS space will look back on the introduction of the iPad as a seminal moment in the history of our business.

Do you agree? Let me know your thoughts.

Not earth shattering in itself. But the call was from some folks at Administaff, responding to a posting on Twitter I made a few days ago.  The post - in its entirety is:

Administaff says new web site is better cause its “cookie enabled” & users wont have to remember a password. How is less security better?

Now, that’s not a really earth shattering observation, if you read the e-mail excerpt that led to my post:

One of the hallmarks of the new MarketPlace is convenience, and you’ll notice it the first time you log in – because if you bookmark the site, it will be the last time you’ll have to enter a username and password to start saving. The new site is cookie-enabled, so it will recognize your computer every time you access the site.

I received this e-mail as a participant in Administaff services - my company uses Administaff for HR functions.

Although I could belabor the point here, I am not a security expert. But it seems to me that the storage of authentication information in a cookie, as is suggested in this snippet, is not a best practice. Anyone who gains access to the computer on which the cookie is stored could potentially open the web site and be automatically authenticated without any user action.

The folks who called me sought to calm my fears, insisting that the site in question, MarketPlace, does not use or have access to Administaff data, and that they operate independently - although the site uses my Administaff username and password, which seems to partially contradict this statement.  I was also told that the site is not operated by Administaff directly, but by a contractor.

The security folks can weigh in on this and make their observations. For me, the interesting thing to note about this is that the posting was made on November 23, and I received the phone call on November 25. That’s a two day turn around for Administaff to see the Tweet, make an assessment, and track me down. And note that I got contacted directly - by phone. Not e-mail.

So the point here is that regardless of what you may think about social media, companies are listening - and responding - to Tweets posted that affect them.

And a pat on the back to Administaff for being one of those companies.  I don’t have a huge following on Twitter, but posts can go viral quickly. So Administaff’s response to my post was a terrific example of a company paying attention.

But - I’m not going to use MarketPlace without disabling cookie caching on my browser. And I’ll be checking out MarketPlace shortly to see how that works.

Accounting departments have always been document intensive - invoices, purchase orders, statements, expense reimbursement, payments, etc. And much of this is concentrated in accounts payable. In this environment, a document management system shines as a corporate record keeping tool. The architecture doesn’t vary substantially from what was illustrated in part 1 of this post, but application integration with office productivity applications such as Microsoft Word and Microsoft Excel is probably not used. In fact, in many cases, the types of documents stored are limited to document images - either in the form of TIFF or PDF files.

On the surface, the accounts payable process, shown in figure 1, looks straightforward. All the incoming paper documents (invoices and their supporting documents) are converted to TIFF files in a scanning process. As part of that scanning process, the A/P clerk views the document, enters the index values for the document and stores the image into the document management system. In an A/P environment, these index values include such data as the invoice number, total amount, invoice date, vendor, etc. These indexes typically match, or relate to, the data used in the ERP system to track transactions.

Figure 1 - A DMS in an accounts payable environment

Once the invoice image is scanned, indexed and stored, the invoice can be processed for payment. And we’ve only scratched the surface of what is possible. Once the base system is in place, it can be extended with additional tools enabling higher levels of automation. But leaving automation aside for now, there are still huge benefits to the indexes used the store the documents.

With the index data, the accounting department can search for any A/P invoice. Taxonomies can be developed using the available index fields, by matching a document to vendor, vendor invoice number, date and so on. This can have a substantial benefit to audits, and to legal actions requiring document production. Instead of searching through paper records, or requesting boxes from offsite document storage facilities, these records can be retrieved from the DMS.

This ability to search for documents and produce them on demand is an incredible enabler for an enterprise in other areas too.

In human resources, a document management system provides the ability to store, index and retrieve personnel data including certificates, licenses, employee intake documentation such as W-4, and I-9, employee reviews, and so forth. Rather than a paper file, all this information can be collected in a DMS. To make this even more interesting, the DMS can be tied to a number of workflow engines, asuring that activities for employee intake, promotion, termination, etc. comply with both organizational policies and applicable employment law. Once again, legal actions show the benefit of DMS usage. All employee records can be produced quickly in that event, for delivery to corporate counsel.

As a last example, consider an environmental engineering department. I’ve done some consulting for an organization which operates gas stations, and is affected by environmental law. Every gas station must have a license or permit from the state in which it operates - in fact several permits. In Texas for example, there are air quality permits, storage tank permits, waste water permits, etc. If any systems are repaired or installed, all work orders, service invoices, etc. must be kept as part of the environmental records. The DMS enables the organization to store, index and retrieve these records, demonstrating compliance with regulations.

So there you have it. In two short posts, basic information on document management systems, how they work, and what they can do. And I’ve only scratched the surface. In future posts, I’ll be expanding on one some additional aspects of document management that make implementing a DMS even more compelling.

A DMS is probably in your organization’s future, if not already there.

In some cases, this is also called “content management” but I’m going to avoid that term, because content management can also be applied to systems that manage web based content, and the type of document management system I’m going to talk about is not limited to web content.  In fact, web content is not often stored in enterprise document management systems - although it can be.  But I’m getting ahead of myself.  Let’s step back and start from the beginning.

To start, you need to grasp that there are two basic components to a document management system.  The first is a document repository.   Think of this as a file room, filled with filing cabinets, filled with folders, and files.  In a modern DMS, the repository is usually just disk storage, with a specific hierarchy of folders.  The hierarchy can vary, and to the end user, isn’t visible.  And because we are sticking to basics, we won’t go there.

The second component is a database.  This could be one of several relational database offerings, including Microsoft SQL, Oracle, Sybase, DB2 - and others.  The database tracks the location of the document in the repository, along with document metadata.  Metadata may include the author of the document, the last edit date, the document type (Microsoft Word, Microsoft Excel, Adobe PDF, TIFF, TXT, etc) and any number of other classification types.  These classification types are also known as index values.  Think of this database as the cardfile in the library.  Find the card for the document you want, and it tells you the “room”, “filing cabinet”, “drawer”, and “folder” the file is in.

An example of this basic structure is shown here

Figure 1 - A two tier DMS

In reality, the architecture shown in figure 1 is rarely seen for several reasons.  One is the need for more functionality.  The other is the direct connection between the user and the DMS resources is not the most efficient, or effective way to do things.  A better way is to insert another tier of service into the model as shown here

Figure 2 - A three tier DMS

In the three tier model, the DMS services broker all the database transactions, handle all the file operations between the repository and the user, and interconnect the other services, such as the web interface, the full text indexer, and any other services that may be present.

Now, those are the basic architectural pieces.  But, how do users access a document in the system?

For DMS access, we can give the user a few different types of interfaces.  One is a web browser.  The user can log in to the document management system, search for documents using the metadata, select a document from the search results, and act on the document.  The actions possible include viewing the document, downloading a copy or “checking out” the document, very much like checking out a library book.  In addition to a web browser interface, the DMS may have a standalone client.  Or it may have a client that is a “plug-in” for Microsoft Outlook, which is a very popular option.  And last, but not least, some DMS variants offer Blackberry, Windows Mobile, and iPhone clients.

Another part of the user interface may include application integration.  With application integration, a user opening a file is presented with an alternate dialog box rather than the standard one.  The alternate dialog box provides the application access to files stored in the DMS, and typically also provides search functions the user can employ to find the document they need. So, for example, a user running Microsoft Word and opening a file might open that file directly from the document management system.  Here is an example of the file open dialog a Microsoft Word user might see (this is a dialog from Autonomy’s Worksite DMS);

Figure 3 - Autonomy Worksite integrated file open dialog.

This particular view shows the user’s worklist - the documents the user has worked on recently.  This is roughly equivalent to the recent file list that most applications provide as part of their user interface.  That might not seem very impressive, but remember that a DMS is a multi user library.  The real power of a DMS starts to be more apparent when you start searching for a document.  To give you a little hint about that search power, see Figure 4.  This is a sample search dialog (again from Autonomy’s Worksite Server).  Notice the multitude of indexes that can be used as search parameters (the index field captions have not been customized in this example).  Being able to build a taxonomy with this level of detail lets a user narrow a search result set even in a repository of millions of documents - assuming of course the documents were correctly classified when they were stored.  And this ignores the power of searching a full text index, which can be combined with the index search.

Figure 4 - Autonomy Worksite search dialog

That’s one example of document management, and in this case, it’s the type typically found in professional services firms, such as law firms and accounting firms. In fact, the dialog boxes I have shown are from a product oriented at those environments.  But there is more to document management than tracking a legal pleading or a general ledger spreadsheet - as you’ll see in part 2.

Ah, the naivete of youth.

I suppose that a day may come when we achieve a paperless office, but it won’t be soon. Why? I can think of several reasons.

First, we are too enmeshed in business processes that rely on the paper document. Want proof? Walk down to your accounting department, and watch all the A/P clerks at work. They are all processing invoices - paper invoices. Amazing when you think about it. Accounting systems have come a long way, but in the end, payables and receivables rely heavily on paper. The only way we get away from this in reengineering the business process globally. Make invoicing use a standard data transmission. Make that data a file if you like, and attach it to an e-mail. But make the data universal in format. We could do that right now using XML if all the accounting vendors could agree on some standard XML field names for data exchange. But you are going to have to get all the accounting vendors to agree, and craft an intake process for those data objects. Good luck on that.

But even if we convinced the ISVs to add these facilites to their accounting packages, there will always be the small business that uses a paper invoice. So on this count, paperless? No. Less paper? Yes.

Another prime example of an old model business process is anything having to do with a fax machine. Think about it - we still rely on a document transmission method that can trace it roots back over 160 years. That’s amazing. But it’s still getting used, in mailrooms all over the world, right now. I’ll admit there are advantages to fax - reliable, low bandwidth, uses readily available analog phone lines - but the receiving fax machine still frequently ties us to a piece of paper. Granted, faxes are now frequently processed as images but many start life as paper, and I’d bet a large majority of the received image faxes are printed and filed. And those images frequently have to be processed by OCR software, adding complexity and expense to this old school technology.

But like the accounting example, getting rid of fax is a tough prospect. E-mail has certainly reduced the need for fax, but I’m waiting for the day when someone asks me to set up a new network infrastructure and tells me they DON’T want fax. So no paperless here either. But faxes as images do mean less paper.

And for the trifecta, look at a business process that uses paper forms - any process. What an incredible model of inefficency in a world where everything is connected, and a good e-forms system can model a business process with a higher level of compliance and consistency than ever possible. Paper forms are a business dinosaur, and they should be extinct.

But in our world, the entire business environment is a Jurrasic Park. And again, there are those small businesses buying sales slip pads, “While you were out” pads, and countless others at the local office supply store. So again, not paperless, but maybe less paper.

The second reason paper stays is convenience. Right now, I can print a document, grab a pen and go anywhere and read, annotate, and edit that document. It’s cheap, it’s portable - and it’s a data security nightmare. I’d bet more organizations have lost sensitive data through dumpster diving for paper than have ever lost it through stolen laptops, or network security violations. Think about that next time your sitting outdoors at Starbucks and a quick breeze blows away a few pages of confidential financial data. But that darned convenience factor overrides all.

So what about the paperless office? To quote my Brooklyn friends, “Fugetaboutit”. Not going to happen. No matter what a vendor of a “paperless office” product tells you. But - and you knew the “but” was coming - we can move to a “less paper” office.

And it’s not really that hard. Three steps I’ve already outlined - eliminating paper invoices and fax machines, and implementing e-forms - not only reduce paper, but can have a real impact on operating expenses. For example, think of the man hours spent keying data from paper invoices, or validating that data from scanned images. Think of the risk (and printing cost) eliminated by modeling business processes with e-forms.

And there are more steps that can, and must, be done in parallel with these. One absolute requirement is a well implemented enterprise content management system. All the documents, document images, etc, must be managed and contained in repositories which are connected to the users who require them through the line of business application they use, whether it is Microsoft Excel or SAP. And think of the operating expense reduction we could all see by eliminating all those filing cabinets full of paper, and all those boxes stored offsite at your favorite document storage location.

But we’ll still want to print that file and read it on the plane, or while laying on the sofa, or sitting on the patio. And we’ll still jot notes on pads of paper. It’s just the way we are.

Disclaimer - the author is employed by a developer of middleware which connects document capture and enterprise content management systems to line of business applications. The opinions expressed are those of the author.

In the vernacular: OMG.

For those of you out of touch, Twitter is a popular social networking site, and it’s exploded. Growth in the past year has exceeded 1300% by some calculations. And it’s had growing pains. Outages, called “Fail Whales” for the icon twitter uses when a failure occurs, are frequent. And recently, the service has been attached by a worm which infected a number of Twitter accounts, an inevitable result of being highly visible on the internet.

So what will happen when Oprah tweets? Think of it - lots of new users on the service, lots of tweets from those users. All that adds up to more traffic for Twitter - an already somewhat fragile service. If you are a user of Twitter, and don’t care about Oprah, it means your user experience, and your view of Twitter as a company, could be negatively affected. Not a good thing for a nascent company that is trying to become profitable.

I really can’t blame Oprah. Using Twitter with a talk show isn’t even a new idea. Kevin Pollack started a web based talk show a few weeks back, and his primary viewer interaction channel is Twitter. But - and no offense to Mr. Pollack - his audience is small compared to what Oprah’s is, and her show will stress Twitter much more than Mr. Pollack’s.

Still, the service is free so it’s a good move from a business perspective: no cash outlay, lots of viewer interaction, and lots of buzz from using the hot social media of the moment. And it’s a way for Oprah to connect with her audience, estimated at 7.4 million viewers. To put that in perspective, Twitter had 9.3 million visitors to it’s site in March.

If we assume that’s an accurate count of actual Twitter users, and if only half of Oprah’s viewers participate on Twitter, it means Twitter will experience 3.7 million new users in the next few days - or roughly 40%. Ouch!

The final results will be known in the next few days, as Oprah starts tweeting. But for now, I bet the business folks at Twitter are ecstatic. In fact, I’m sure Evan Williams, the CEO of Twitter, is grinning from ear to ear about the prospect of Oprah using, and thus promoting, his company. But if I was his CTO, I’d be afraid of what 7.4 million potential new users could do to the system. Very afraid.

Update - 4/18/09
Well, it’s over and nothing bad happened to Twitter. I’d say “I was wrong” but I think the jury really is still out. It’ll take those 7.4 million viewers time to get up to speed on Twitter. Let’s see what the longer term effect is of all those potential new users…

There’s nothing wrong with these events. In fact, I like them, because they are a good way to get answers directly from the vendor’s reps. And a lot of the time, I get a pretty good meal out of it. But I left today’s event pondering the state of desktop virtualization. More correctly, I left it pondering the need for desktop virtualization - or at least application virtualization.

In application virtualization, an “invisible” thin virtual machine runs the app, totally isolated from the system it runs on. No direct registry access, no usage of locally installed DLLs, etc. This means the app can be copied to the PC it runs on easily, run without regard to what might already be on the PC, and updated fairly easily by updating a master image of the application.

But come with me now, as we step into the Wayback Machine, and return to a simpler time.

In days of yore, Windows applications stood alone. Configuration information was carried in things we knew as INI files. Program libraries were kept in the directory with the program’s executable. An application could be installed by creating a directory, and copying the application files to the directory.

Ah, how things have changed since those days. We now have a centralized application configuration database - in the form of the registry. We also now have shared libraries - DLLs - populating our hard drives. These things have been with us on the desktop since Windows 95, and things have been complex ever since. The complexity has grown as desktop operating systems became capable of application customization based on the logged on user - sections of the registry that were user specific. And now in the timeline, it’s application virutalization. An amazing effort to abstract the registry and the DLLs and all their inherent management problems.

Think how easy it would be to “virtualize” an application if it didn’t use the registry, always looked in it’s own folder for it’s own DLLs, didn’t rely on COM, etc. etc. Pushing a “virtualized” app to the desktop would be as simple as running xcopy. Look for an example to the “portable” apps being produced to run from a USB thumb drive.

Am I the only one who looks at the history here, and feels a sense of irony?

We complicate the environment, then jump through hoops in the form of layers upon layers of desktop virtualization tools, to enable apps to run independently of the complex environment that we have created.

What were we thinking?

Although my crystal ball is usually cloudy, I’ll make a prediction here. In the future, we’ll see a desktop OS that is designed for application virtualization from the ground up. It will provide a nice UI. There won’t be a registry - at least for applications. While the OS may use a set of shared libraries, the apps won’t use those libraries - at least not like they do now. There will be some form of granular inter-process communications that can be controlled, to allow some apps to talk to others. Apps will “register” at run time and “check out” at termination.

And because the OS will be designed for application encapsulation, all the applications included with the OS will themselves be encapsulated, and therefore easily removed or replaced. And if the OS gets corrupted, it can be reinstalled without requiring the reinstallation of all the applications.

Of course, this would all require a radically different desktop OS, and backward compatibility would be a nightmare of kludgy components.

Wait - what am I thinking?

In Mark’s organization, one server in a 4 node cluster hung due to a software bug. Users were unaffected thanks to the cluster architecture. But the NOC received a systems alert reporting the server problem, and an operator was dispatched to restart the server.

After the incident a time and motion analysis was performed. While the act of rebooting the server may seem simple, it was not the end of the chain. The operator reported the action to his supervisor and time was spent on discussion. The supervisor reported the incident to his manager. The incident report moved up the chain, until it was discussed by a high level group. And of course, appropriate incident tracking documentation was generated as well.

Once all of these activities were totaled, the analysis found that the cost to reboot that one server, one time, was $5K.

And now my first confession. While I have considered the cost impact from big picture perspectives such as business continuity assessments, I have never before considered the costs of such an isolated action. This story made me stop, and evaluate IT incidents in a totally different frame of mind. And I hope with this posting, you will too.

In these economic times, organizations are squeezing every dollar out of every department, including IT. The challenge for IT managers is how to continue to provide the services demanded by the organization, while at the same time dealing with demands for lower costs - demands which usually equate - as Mark pointed out - to reducing headcount. But headcount reduction in turn reduces the quality of the services IT can deliver. It’s a trap you can fall into, unless you can offer alternative ways to cut costs, which brings me back to that $5K server reboot. With this kind of analysis, you can identify other costs that may be reduced, as an alternative to headcount.

I admit I was shocked at the cost of the reboot, and I’ve been wondering how that cost might be reduced. My second confession is, I have not developed an answer. In a modern IT department, there are processes and procedures which have been adopted based on best practices, such as ITIL. The procedures define the steps that must be taken to report and document an issue - and there is nothing wrong with these procedures. But following them exacts a price.

I am not suggesting abandoning ITIL, or any best practices system. Far from it. But I will point out that it is important to be aware of their financial impact. And more than that, it is vital that these best practices be used in such a way that incident costs can be determined - as Mark was able to do in his organization. So the answer isn’t missing - but the question is wrong. The question isn’t about how the cost can be reduced. The question is, how can the reboot cost information we now have be used?

Remember - our task in IT is to provide valuable information to the organizations we serve. But in addition to being providers of information, we must be consumers of information. And we must use that information to make our task of serving the organization more effective, more efficient, and more economical.

Note: I’d like to thank Mark Settle for permission to use some material from his presentation for this posting.

I would not like Vista
here or there.
I would not like Vista
anywhere.
I do not like
this OS.
I do not like it,
Not one bit.

Would you like Vista
on a netbook?
Would you give Vista
one more look?

I do not like it
on a netbook.
I will not give it
one more look.
I do not like Vista
here or there.
I do not like Vista
anywhere.
I do not like it with lots of RAM.
I do not like it, Juan-I-am.

You may like Vista.
You will see.
You may like Vista
Hear our plea!

I will not,
Cannot, hear your plea
I will not allow it
Even for free

Senator!
Senator!
Stop this bill!
The thought of it
Gives us a chill!

I do not care
How much you cry
Wave to Vista
Tell it goodbye

Senator!
Senator!
Give us some peace!
A new version
Will soon release

New version?
New version?
About that
I don’t know.
But Microsoft Vista
It must go.

Senator!
Senator!
Don’t make such a stink
It’s all a moot point
Don’t you think?

Think?
Think?
I am a solon!
Although right now
I feel a moron.

Up the road a bit, there’s something going on that proves my Law of Inverse Political IQ. The law states that when you get a group of otherwise intelligent men and women into a room, and if that group is a group of politicians, the IQ of the group will fall as the inverse square of the number of politicians present. Readers, I give you the Texas Legislature. Now if you don’t know much about the Texas body politic, I’d like to draw your attention to such items as a legislative attempt to ban “suggestive cheer leading”, or the fact that in the great state of Texas, there is a law making it illegal to milk another person’s cow.

Yeah, we got our share here in the Lone Star State.

But one of my blog topics is management, so I’d like to present you
State Senator Juan “Chuy” Hinojosa.

In a stunning display of micromanagement, Senator Hinojosa has slipped a rider into a finance package banning state agencies from purchasing Microsoft Vista, unless the purchase is specifically approved by the legislature.

To repeat an overused phrase - what was he thinking?

Can you imagine a member of the board of directors of a major corporation setting IT policy for all the subsidiary companies? Because that is what the Senator is trying to do here. This flies in the face of so much management theory, I hardly know where to start.

But I’m going to give it a shot.

I’m sure Senator Hinojosa means well - many incredibly bad decisions have been well intentioned - but in this case, I doubt he has the technical expertise to assess the merits or debits of Vista - especially for over 170 individual state agencies. I’ve checked his bio and he has some impressive accomplishments: Marine Corp, practicing attorney, etc. But I did not find a single item with the title CIO, CTO or any other tie to IT management.

Hence broken rule #1 - if you don’t have the expertise, don’t try to manage it yourself.

Each of those 170 state agencies has someone internal managing their IT shops. Those individuals have the positions their positions (hopefully) because they are competent, experienced leaders. Why Senator Hinojosa feels he needs to lead them on this is a mystery. In fact, if he had checked, he might have found that a number of these leaders were skipping Vista and planning Windows 7.

Broken rule #2 - enable your subordinates and let them manage what they know

And the biggest irony of all is that Vista is approaching sunset, as Windows 7 appears on the horizon. The senator is closing the barn door after the cows got out (at least he didn’t milk them - he’d get arrested).

Broken rule #3 - make sure the policy you enact addresses a real, current or future problem

I suppose I could go on, but that would be like shooting fish in a barrel. And with my luck, that’s illegal in Texas.