Naming Downloadable Files

PDFs, word processing and spreadsheet documents, software packages, and lots of other things are downloaded to a user’s computer to be referenced at a later date. That’s great—if you can find it when you need it!

Let’s put ourselves in the user’s shoes. We’ve downloaded something from a web site, either because we know we’ll want to refer to it later without going back to the site (user manual, spec sheet, receipt, etc.), or because it’s the only way the site has provided to access or manipulate the information (”download this form, fill it out, and fax it to us”). Either way, we’ve saved the file locally and will need it later.

Let’s say we’ve downloaded the spec sheet for Harman’s GLOSS FB Al inkjet photo paper (which is amazing, BTW), while researching which paper to buy. We decide it looks good, place an order, and wait. Now it’s a few days later, the paper has been delivered, and we want to look at the spec sheet to see if there are any special settings we need to use before we start printing. So we go to the folder we saved it to to look for it.

Summary so far: The manufacturer has provided the information we need in a format that’s convenient for offline viewing, both on-screen and in print. It was pretty easy to find on their web site. They didn’t make us register or pay to get it. We have it. Things are looking good so far!

But, now that we want to look at it, we have a problem: we can’t find it! Why? Because the name of the PDF file they thoughtfully provided is—seriously—2007622163739169.pdf. Huh? I doubt the name even means anything to anyone at Harman. I know there’s nothing about it that tells us what it is. The best we can hope for here is to view the folder contents by date, try to remember when we downloaded it, and start opening files that were created in that window of time until we find the right one; or do a slow search on file contents, hoping it was generated in such a way that our search term can be found.

Another example is WordPress, a popular blogging package. They have since fixed the problem, but it used to be that when you downloaded the software, you got a file called latest.tar.gz (or latest.zip, depending on which format you chose). Latest what? We have to unzip the file and look in the folder it created to find out. And we’re in big trouble if we want to know what version we’re looking at, because we’d have to open and read through a bunch of PHP files (computer code that’s unreadable to most people) to find out. We might stumble upon the version.php file, but it would probably be by sheer dumb luck as it’s tucked away in a sub-folder called wp-includes.

This is the user experience equivalent of Punk’d. A scenario of helpfulness and customer support is laid out: the information provider has a web site, has the reference material (and in a standard format), doesn’t make us jump through hoops to get it… and then yanks the rug out when the chips are down. Why go to the trouble of producing and providing information that’s useful to your customers, only to make it effectively inaccessible when they actually need it?

In some ways, this is worse than not providing the information at all. By going there, an expectation is created that the user will be able to easily access it when they need it. Creating the expectation and then failing to meet it is likely to frustrate and anger users much more than not creating it at all. It’s probably better to stick to having the info available only as a web page, which is an expectation that can be met: the user can go to the site and find what what they need. It’s not as useful as doing it right, but it’s a better user experience overall.

We’re not stuck with the 8.3 file naming restrictions of DOS. We can have mixed case letters, word separators, and even spaces where appropriate!

WordPress’ latest.zip can follow the naming convention standard for open source projects (and it now does), which is app_name-version.extension. wordpress-2.5.zip actually tells us what we’re looking at! There are no spaces, which is a good thing here: it’s easier to deal with files in a CLI shell (like on a web server) if there aren’t any.

But since the PDF is almost certainly going to be stored on a computer with a graphical interface, spaces are ok, and the useless 2007622163739169.pdf can become Harman GLOSS FB Al Tech Info.pdf. To really bring it home, a version number could be added: Harman GLOSS FB Al Tech Info-v1.2.pdf, or Harman GLOSS FB Al Tech Info - 2008-01.pdf. Now the user not only has the info they need, but they can find and use it, and can find out if they’ve got the current version.

That last 0.001% of effort it takes to name the file properly can be the difference between the user having a bad and frustrating experience associated with a company or product, and a good one. Seems well worth the 5–10 seconds to me.

Related posts

• mozilla.org’s human-proof CAPTCHAs
• Keyboard shortcut conflict: iChat vs browsers
• Jakob Nielsen answers "how little do users read"?
• Drag-and-drop download
• Adam Darowski on URLs as UI