Proper Entry Archiving: The Problems With Information Management

I have been thinking, thinking about how to store and find data for month’s now. A while a go I even started to put together the framework for a new File Exporer for Windoze/Linux.

Yesterday, when Apple announced some of the their new features of Tiger, one thing caught my eye: Spotlight. I remember an article a while back about longhorn that explained more or less what they planned as far as searching goes.

I have a few different ideas about searching that I want to explore a bit.

Warning long post!

Google Gone Wrong

I think that Google (and competitors) have done some obviously great thing for searching. But as unoriginal thinkers often do, the herd is thinking that what is good for the goose is good for the gander. If you look at the basic idea around these search engines you get the search in the way that think.

It seems to be in the direction of the movie AI, where the annoying boy asks the annoying computer questions and the annoying computer annoyingly answers them.

Although no one is quite that far ahead the basic idea is to be able to type, ‘A document from two weeks ago that starts with z’ and the computer will return all the documents that match that criteria.

Now this is good, it is just like Google, where you type whatever and you get a million choices to choose from. You may be super ambitious (like if you are searching for pictures of Heidi Klum) and go through a hundred or more entries. Most likely, if you are like me, you will look, at most, the first few links and maybe make it to 11-20 before giving up and either trying a new search or quiting. But it really is a remedy rather than making things better.

The Search Dilemma

The problem with the web and, more importantly, with file systems is how information is saved.

The file directory is a really, really poor way to save files. It is just begging you to forget where you put things.

For example, let’s look at something that is problematically to you right now: Your bookmarks.

Now if you were to just list all your bookmarks with no files you would never be able to find anything. Maybe you could move your favorite favorites to the top and the rest would eventually rot. Now probably you have some sort of structure using different directories and spacers to organize things. That is how I do it. I try and group things on a high level, like Work and Play and then go down from there.

The same thing for file on my computer, starting from home I work my way down. Now for most recent files I can find thing without too much difficulty and for the really tuff things I do a cross computer search.

A Better Way

If you use Microsoft Outlook you are probably not aware of the power of a little used feature: Categories. Now for the most part I dislike MS, but here I think they are on to something. They need to take it further, but it is at least a start.

Before I set up categories I had lots of directories for all my emails. Then I set up my rules wizard to send the emails here or there. When I lost something I would try and do a search, which is always difficult.

Now I have things set up a bit different. In fact all of my emails stay in the Inbox, instead of sending them to directories I use the rules wizard to apply categories to them as the come in. Also, before sending out an email I will assign it to a category. As these emails go out and replies come in they maintain the category setting. Next I have added category to my view and sorted them by category then by date.

Because everything all mail comes into my inbox if I can’t remember where I put something (which doesn’t happen much anymore, because I assign one email to multiple categories which increases my chances of finding it when I am working on an issue) I can change my view, for example, I can switch to sorting by author or received date.

This is the beginning of how files should be set up. Using a symbolic link idea there should be multiple references to all files (not shortcuts, but a link that is the file for all intents and purposes) that can be set up like my Inbox.

This is where Google comes up short. Gmail is nice, but lacks this function.

If it is easy for you to save your file using categories it will be easy to find. The entire directory concept should be done away with. There should be a hard drive where everything is stored and then you organize it however you want.

The Next Step Forward

Categories are a start for files, but the real power will come when keywords are employed. If in addition to categories if you have a space to put a number of keywords nothing will be lost again.

It would work like this. You would have a set of predefined categories (that can be changed whenever), like documents, pictures, etc. Then you give the file a name (which shouldn’t be necessary anymore, because in actuality the real file name is kept by the computer, which would allow for faster indexing) then you check a couple of categories that apply to the file, or create a new one, like April 04 & Sales Presentations. Then you type a few keywords like General Motors, John Doe, Sales increase, etc. where the categories are used to get some visual organization and the keywords are more as selectors.

Finally, all of your open dialogue boxes or your file browsers are kind of like iTunes where you can choose which groups to sort by. You could choose date, category, keyword, author. Each grouping would have all possible entries so you click on a date range like month of June (or type June) then you get all the authors categories and keywords. You select yourself as author and now you select two or three categories each time the main area shows all the files and the selector fields narrow.

So far my experiments with this system have produced much faster and more targeted results.

I believe this is a better way to play. If I had the resources I think I could build a really outstanding program to do this, but as I don’t all I can do is talk.