Thursday, January 8, 2009

Your company's first bug of 2009

I bet I can guess a bug at your company. You may not even know it. Nor your customers, or your marketing department.

It's probably in your code. It might be on your website. It is probably completely innocuous until you need it most, and then, well, your legal staff will tell you all about it.

If you haven't guessed yet, it's your copyright text. Most likely, it says something like:

ABC Company copyright 2002-2008. All rights reserved.

or even possibly "... (c) 2002-2004" if it is in code.

What does this mean? Does it really matter?

Good question. The facile answer is that if it was important enough to merit a copyright notice from 2002 to 2004, then it is probably important enough to update now. A more sophisticated question is if there is any legal standing for the years after a copyright notice. To be honest, I don't know.

All I know is that one team lead (who I respect) once speculated that omitting a copyright notice on new code was grounds for dismissal. I thought that was crazy then, and I do now: but it underscores the importance of this legal business. (As an aside, it just looks plain bad when the year is off on a website.)

I invite clarifications on the ramifications of letting copyrights 'expire' (if that is the correct term), but for now the takeaway is: update the darn year, and give your lawyers some backup support.


fred said...

I don't know... if the underlying code hasn't changed, do you need to update the copywrite notice? You don't see publishers recalling all copies of a book each year to update the page. Heck, even if they re-print it in the exact same format, they will leave the original copywrite date statement.

If code is UPDATED, yes. New code written - yes. I wonder if having an 'include'-like statement in whatever language where that file has a copywrite is good enough.

Michael Easter said...

re: if not changed. Good point, Fred. The motivation for this post was that I was working on code yesterday and noticed the copyright.

Indeed, the presumption for source code is that it has been changed in 2009. Though I don't presume that for websites: even though the content hasn't changed, it seems to me that the copyright should be updated.

I fully concede that this is not my strong suit. I have often wondered if there should be a legal aspect to a computer science education.

fred said...

One of the difficulties with a legal education for CS is that the law hasn't really been worked out yet for a lot of this. Lawyers don't know what the rules are yet.

The legal system basically works on case law - in other words, you find a case that has already been decided that is similar to yours, and then argue they are comparable. At the same time, your opponent is arguing that another case is more relevant that proves the opposite.

Since much computer stuff is relatively new (legally speaking), much of the law still has to be decided.

cas said...

I agree with fred that the copyright needs to be changed only when a work has changed.

However, even if the work has changed -- or even if it's new -- it's probably copyright.

I ran into this article by Brad Templeton a long time ago.

I guess one point the article makes that needs to be brought out is that there is a difference between something being copyrighted and the amount of damages one can collect if the copyright is violated and there is no notice.

(It was after reading Brad's #1 myth write-up that I quit using "(c)" in my notices -- even when the coding standards call for it.)

Steve H said...

Disclaimer: I am a practicing attorney, but I am not your attorney. My comments indicate only my general impression of the law and should not be taken as a legal opinion upon which to base your conduct. Please consult your attorney regarding specific recommendations.

Two scenarios for copyright notices are presented: 1) In source code, and 2) On web content.

Hidden in #2 are two sub-cases: Static and dynamic content. I submit that the discussion of source code applies just as well to static web content, so they are grouped together below.

As a general matter, the article linked by cas is correct in noting that copyright attaches by default. Every line of code written is protected by copyright unless the owner of the copyright explicitly commits it to the public domain. Whether copyright-by-default is a good idea is a matter of some debate. Personally, I think there are good reasons to require an affirmative assertion of copyright by an author as a condition of exclusive rights, but that is tangential to this discussion.

Despite the copyright-by-default rule, providing a copyright notice serves a couple limited functions. First, it negates any claim of "innocent infringement" which may be used to reduce a judgment against an infringing defendant. 17 U.S.C. 401(d). In other words, any infringement of a work with proper notice is considered willful infringement. Second, as a practical matter, the notice may have an intimidating effect on a potential infringer and thus discourage unauthorized use.

For any static content, including source code, I would suggest that auto-generated or auto-updated copyright notices are misleading. If nothing has changed in a document since 2005, I would prefer to avoid attaching a notice like "Copyright 2000-2009". I haven't looked into whether there is any legal risk associated with the practice, but it is inaccurate, which would bother me. As far as I know, one cannot perpetually renew one's copyright through mere assertion.

Further, there is little harm to be done in marking your documents with a copyright notice referencing the last modification date when one considers the current copyright terms in the US. For works with an individual author (e.g., a blog post or page in a personal website), copyright is effective for the entire life of the author plus another 70 years. 17 U.S.C. 302(a). In the case of a work made for hire (e.g., code written for your employer), the term is the shorter of 95 years from publication or 120 years from creation. 17 U.S.C. 302(c). I cannot imagine much source code or static web content created today retaining commercial value for 95 years. Surely, giving up the last 4 years (2009 minus 2005) of a 95-year term is not a terrible cost. And, as I said before, if the work has not been updated, the owner of the copyright probably has no legal claim to those extra 4 years, anyway.

This raises another interesting point. Imagine you write an article titled "Mike Easter on Copyright Notices" for your employer in 2009, which the employer quickly publishes. In 2013, your employer asks you to revisit the subject, so you start with the 2009 article but include revisions based on some enlightening discussions you've had with a software-developer-turned-lawyer. The company publishes the new article as "Mike Easter on Copyright Notices." Copyright for the 2009 version of the article expires in 2104, 95 years after publication. Copyright for the updated version expires in 2108. This is the case even if the articles are posted at the same URL. They are distinct works. (Note that these had to be works made for hire to illustrate the point. If they were personal blog posts, copyright for both would expire 70 years after you do.)

For static web content, I suspect it would be fairly easy to implement an automated method of inserting a copyright notice based on a file's modification date. Perhaps a placeholder or token could be included in each file, and the webserver could use a server-side include (SSI) feature to populate the correct year(s). A similar result could be accomplished for source code in at least some revision control systems by configuring them to insert a comment with a copyright notice based on the last revision date for a file.

Of course, both of these methods apply the copyright notice only when a work (source code or web content) is accessed through a particular channel, such as an HTML or RCS request. This raises an important question: Who is the intended audience for the copyright notice?

In the case of a company's internally managed source code, the work will likely never be published outside the company (in which case the copyright term is 120 years from creation). Presumably, the employees with access to the code already realize they do not have the right to copy it for non-company use. This pretty much undermines both copyright notice benefits described above: the infringer already has no legitimate claim of innocent infringement, and no discouragement effect can be achieved.

I suspect Mike's team lead was well-intentioned but rather ill-informed on the subject of copyright in general and copyright notices in particular.

I hope I've added something useful to the discussion, especially since it's late, and I fear I have laid my thoughts out in a rather lengthy and disorganized manner. Feel free to ask for clarification on any points.

Michael Easter said...

Steve, thanks very much for this excellent analysis! The law, like math, has to be absorbed carefully due to the subtle implications of definitions. i.e. I'm still soaking this in.