Thursday, January 25, 2007

Copy and Paste doesn't work on the internet

Copy & paste is a crucial user interface metaphor in modern operating systems. It works both within files (for example, copying a sentence from one Word document to another), and on files themselves (for example, copying a file from one directory to another). It's always by value, rather than by reference - when I update a pasted sentence, the original doesn't automatically update too.

Copy & paste is commonly used in several different ways:

  1. Re-ordering a document, for example moving a sentence from one paragraph to another
  2. Creating lists of similar items, for example copying a spreadsheet formula then editing the copy slightly
  3. Managing versions, for example copying a document then archiving the copy
  4. Re-using content, for example copying an image into a presentation
  5. Creating personal copies of documents, for example copying a file from a USB drive
There are problems in using copy and paste for cases 4 and 5. For example, in case 4, the BBC website reuses the BBC logo on every page. But this isn't done with copy and paste - otherwise, they'd need to store it thousands of times, and they'd have to find and update every single copy whenever the logo changed. Instead, each web page copies it by reference (using its URL), rather than by value.

Similarly, corporate email systems use copy and paste for case 5. Consequently, if I send a 1MB attachment to 50 people, it uses up 50MB of storage. When the master version changes, everyone still has the old copies.

And frankly, copy and paste is a poor tool for version management (case 3), especially in the enterprise. Rather than relying on file names and saving copies, a version management tool can show proper audit history and tracking.

Copy & Paste is very personal - it's rooted in the old paradigm where everyone worked on their own PCs (with their own copies of files) and communicated every so often, rather than sharing content as a community from the very start.

The internet is different, because of the URL. Rather than emailing that 1MB attachment to 50 people, if you put it on a website it will only use up 1MB of storage. And when it's updated, people will automatically see the most recent version. That's because the URL works by reference, not by value.

This requires a shift in mindset. For example, in a connected world, what is the point of having more than one copy of any song? There should be a single URL - for example,, and everyone in the world should access this URL every time they want to listen to the song (you can imagine aggregation sites cataloging and presenting songs together). Except for temporary cacheing, there would be no other copies - browsers would not allow it, unless explicitly stated. Your phone, stereo, TV and PC would all access this URL, with the same user profile to control permissions.

Without copy & paste, a solution to Digital Rights Management is possible. If someone wants to copy and edit a song, for example to make a cover version, they should have to apply to get copy permissions.

If you look back at the list above, only 1 and 2 - content creation, rather than re-use - are appropriate uses for copy & paste on the internet. Copy & paste don't work well on the internet, because of the power of the URL.

No comments: