Metadata and the Windows Explorer Property System
If file location will become less important, what can you use to take its place? Content seems like a pretty good place to start. After all, it's what's inside the documents that really matters. For example, suppose you're working on the Penske account. It's a pretty good bet that all the Penske-related documents on your system actually have the word Penske inside them somewhere. If you want to find a Penske document, a file system that indexes document content sure helps because then you need only do a content search on the word Penske.
However, what if a memo or other document comes your way with an idea that would be perfect for the Penske account, but that document doesn't use the word Penske anywhere? This is where purely content-based file management fails because you have no way of relating this new document with your Penske documents.
Of course, you could edit the new document to add the word Penske somewhere, but that's a bit kludgy and, in any case, you might not have write permission on the file. It would be far better if you could somehow identify all of your documents that have "Penske-ness"that is, that are directly or indirectly related to the Penske account.
This sounds like a job for metadata, and that's appropriate because metadata is all the rage these days, particularly on the Web. At sites such as Flickr.com and del.icio.us, surfers are categorizing the data they find online by applying descriptive keywordscalled tagsto the objects they come across.
Social softwaresoftware that enables users to share information and collaborate onlinemakes these tags available to other users, who can then take advantage of all this tagging to search for the information they need. At the del.icio.us site, for example, users bookmark interesting pages and assign tags to each site, and those tags can then be searched. This is called social bookmarking.
Certainly, metadata is nothing new in the Windows world, either:
- Digital photo files often come with their own metadata for things such as the camera model and image dimensions, and some imaging software enables you to apply tags to pictures.
- In Windows Media Player, you can download album and track information that gets stored as various metadata properties: Artist, Album Title, track Title, and Genre, to name just a few.
- The last few versions of Microsoft Office have supported metadata via the File, Properties command.
- For all file types, Windows XP displays in each file's property sheet a Summary tab that enables you to set metadata properties such as Author, Comments, and Tags.
What's different in Vista is that metadata is a more integral part of the operating system. With the new Windows Search Engine, you can perform searches on some or all of these properties. You can also use them to create virtual folders, file stacks, and file filters.
Putting metadata at the heart of the operating system is a welcome innovation. Throw in the capability to sort, group, stack, filter, and create search folders based on such metadata, and few would dispute the value of this enhanced file system.
It's also a good thing that metadata is easy to implement for individual files, but will people get into the habit of adding metadata for each new document that they create? Time will tell, but it's certainly true that metadata has been underutilized so far. I think people will have to be convinced that taking a little time now to add metadata will save them more time in the future because the metadata makes documents easier to find and manage.
It also helps if software vendors can make it easier for users to add metadata to documents. Having to switch over to Windows Explorer to add or edit metadata is not a big productivity booster. Instead, I hope Vista-aware programs will offer metadata-friendly interfaces and prompt for properties when users save new documents.
A much bigger problem is applying metadata to existing documents. I have thousands of them, and you probably do, too. Who has the time or motivation to set even just a few property values for thousands of old files? Nobody does, of course, and I suspect most of us will simply ignore the vast majority of our existing files (after all, we might never use 95% of them again) and move forward into the metadata future.