In search of a silver bullet for iPad file management

Apple's hapless approach to iOS file management shows how important metadata could be

Is file management essential, hazardous, or a relic? Judging from the comments posted to my column two weeks ago, "Microsoft's killer tablet opportunity," and to last week's piece, "Crowdsourcing the killer business tablet," the question is surprisingly controversial.

Some commenters were adamant that file management is dangerous because unsophisticated users don't know how to handle it. Because these hapless folks ask for help finding a missing document, they explain, the tools provided for keeping files organized are at fault.

[ Find out the 10 business skills every IT pro must master. | Get expert advice about planning and implementing your BYOD strategy with InfoWorld's 29-page "Mobile and BYOD Deep Dive" PDF special report. | For more of Bob Lewis' continuing IT management wisdom, check out his Advice Line blog and newsletter. ]

This isn't much of an argument. Depriving sophisticated users of valuable tools because other users don't know how to work with them makes little sense. Educating unsophisticated users is a far better solution, for reasons so obvious I trust they need no further explanation.

The more thoughtful arguments against file management have fallen into two groups. One insists that with effective search, file management becomes unnecessary. The other proposes using metadata as an alternative to traditional file management.

Where search falls short

Someday, search by itself might be as effective and convenient as navigating a well-organized folder tree. If it does, it won't be search as we use the term today.

As pointed out last week, the proof lies no further than your nearest browser. Bring up your favorite search engine and search for "using search instead of file management" -- in quotes. You won't get any hits, even though you and I know of at least two recent articles on the subject. (Well, you might get one hit, for this post, if it has been indexed by the time you perform the search.)

Next, search for "iPad file management critique" without the quotation marks. You'll get somewhere between 670,000 (Google) and 18,600,000 hits (Bing), which isn't narrow enough to be helpful. Again, no Advice Line commentary on the subject appears on the first three pages.

What's that you say? Make the search "ipad file management critique 'advice line'"? That works. It also demonstrates the importance of file management, because by adding "advice line" to the search, you identified the location of the article, along with the contents.

Don't get me wrong. I'd hate to do without search, even the irritatingly slow Windows search engine. There are times when it's the only tool for the job, even if it isn't very precise.

What search needs: Semantic indexing

Search won't be versatile enough to replace file management until someone perfects semantic search -- which actually is the wrong way to say what's needed. The right way to state the problem is that we need semantic indexing and won't have it for a very long time. Semantic indexing is the difference between finding the three files on my system that discuss "using search instead of file management" and only those files, versus getting either zero or a zillion hits.

Almost certainly, Apple won't be the company that develops it. Apple isn't a company that tackles hard computing problems. Even the very clever Siri is built on Nuance's speech-recognition technology. As is so often the case, Apple's brilliance was in its recognition that the technology was ready for adoption and polishing.

Microsoft also won't be the company that develops it. If it could, Bing would be more than a me-too search engine.

Nor will it be Google, because if Google could tackle the job, it would tackle the job. That's its core business, and Google has developers who are as good as they come.

Right now, the only company with the chops to solve semantic indexing is IBM. Watson shows that it's more or less possible. Scaling Watson up to the Web (not our subject right now) is a lot of scaling. Let me rephrase that: It's a lot of scaling.

Scaling Watson down to fit into an iPad, or a laptop, or even a server-based content management system? Easily as difficult as scaling it up.

Metadata: File management panacea

Quite a few commenters insisted that a well-organized, outline-based folder tree isn't just essential to finding what you're looking for -- it reveals the structure of the knowledge the system contains. That thread developed into quite an interesting conversation as to whether metadata can do a better job of sorting files into folders than folders can. The short answer is, yes, it can, and I want it. Right now.

The longer answer is (no, I'm not paid by the word): Yes, and taking advantage of this capability is both an opportunity and a burden. The opportunity comes from not having to store one file in just one branch of one tree. This is a big deal, because while folder trees are hierarchies -- trees, with branches and sub-branches -- information is cross-classified.

In a folder tree, for example, I either organize by subject and client within subject or client, and subject within client. When I consult with a client on enterprise technical architecture management, files either go into the Clients/DeepPocketsCorporation/ETAM folder or the ConsultingServices/ETAM/DeepPocketsCorporation folder, depending on how I've decided to organize them.

Now, imagine using metadata to simulate folder trees: When you "file" a document into a "folder," the user interface is fooling you. It looks just like you're filing the document, but what you're really doing is adding a keyword tag from multiple administered outlines. And because "outline" sounds like something you learned in middle school, the industry calls it a "taxonomy" instead.

In effect, so far as I know from the user interface, I'll stash these documents in both the Clients/DeepPocketsCorporation and ConsultingServices/ETAM folders. When the time comes to find them again, if I look in Clients/DeepPocketsCorporation, I'll find everything related to that client; if I look in ConsultingServices/ETAM, I'll find everything related to enterprise technical architecture management. If I check both locations at once, I find only the ETAM-related documents created for DeepPocketsCorporation.

The key to keywords

For individual users on individual desktop, laptop, or tablet computers, an administered keyword taxonomy would look and feel just like how they create a folder/subfolder taxonomy, except they'd be able to create more than one. Navigating them to find a file would be just like navigating a folder tree, again except that there would be more than one, and under the covers, unbeknownst to them, what would be happening is a keyword search.

On an enterprise level, IT (or the chief knowledge officer if you like that sort of thing) would provide administered taxonomies as part of the company content management system. Users would (I hope) still be free to develop their own private taxonomies, for use whenever the centrally administered ones don't do the job. Beyond this, they'd also be free to add ad hoc keyword tags to documents should that be more convenient. Whenever a user tags a document with an administered keyword, the document in question would be replicated in the content management system and automagically kept synchronized. This just makes sense: If a document fits an administered tag, by definition it's part of the company's fund of accumulated knowledge.

Who will sell this to us?

The sad fact is the odds of our getting this sort of capability any time soon are minimal. Apple is actively against it; Windows 8 will provide Windows-style folder trees and that's about it; and nobody in the Android space has shown any imagination at all.

The only possibility I can see is a company like Dropbox heading in this direction. In fact, if Apple ever turns iCloud Documents & Data into something more than a joke, this might be the only way for Dropbox and its competitors to stay in business.

This story, "In search of a silver bullet for iPad file management," was originally published at Read more of Bob Lewis's Advice Line blog on For the latest business technology news, follow on Twitter.