Home › Blogs › Bruno Lowagie's blog

Freedom versus Freedom

Sat, 03/05/2011 - 14:20 — Bruno Lowagie

I'm entangled in a very strange private discussion about freedom. It started with a seemingly simple question about iText on an itextpdf.com mail address, and now we're discussing about the food industry, weighing one freedom against another one.

Imagine that using software is like baking cookies. A free/open source cookie specialist, company A, distributes recipes that are free in the sense of the AGPL: everybody is allowed to bake those cookies, provided that the consumer gets the recipe along with the cookies.

Suppose that company B uses the recipe to bake cookies, and they sell those cookies to company C. This is perfectly OK, as long as they provide company C with the original recipe from company A, along with all the extra instructions applied by company B. This is how one should interprete the AGPL: company B is not obliged to pay company A for the use of the recipe, as long as company B distributes all of its own recipes ("touching" the original recipe) for free under the same AGPL license.

Company C can now consume the cookies, but they can also use the finished cookies to create another product, for instance: cookie spread. This may seem an odd example, but actually cookie spread is a Belgian specialty. Company C doesn't use the recipe from company A, nor any extended recipe from company B; they are using the finished cookies as an ingredient for another product. For instance: they crunch the cookies, add some extra ingredients, and put them in a jar.
In this scenario, company C doesn't enter in the AGPL: they may have received AGPL recipes from company A and/or B, but they aren't using these recipes. As a result, they can sell cookie spread to consumers without sharing their own recipe.

So far, so good.
The issue that is raised in the ongoing discussion isn't about the AGPL, but about an extra condition that can be added to the AGPL. Does a specific condition, which I'm about to explain, make the recipe "un-free" (which is the point of view that was communicated to me) or does it ensure an additional, basic freedom (which is my point of view).

Suppose that nuts are on the list of the ingredients in the recipe of company A. To ensure maximum protection of the consumer of the cookies—be it in their original form, or processed in some derived product such as cookie spread—an extra term is added to the AGPL: as long as there are traces of nuts down the production line, all products that may contain traces of nuts must retain the notice "may contain traces of nuts".

According to the developer I'm having this discussion with, company A makes the original recipe "un-free" if this extra consumer protection is added as an additional term to the AGPL. He says company B should have the right to remove the "may contain traces of nuts" notice. Up until now, he has forwarded several arguments, but I've invested in legal advice, whereas he admits he's not a lawyer. In short: I'm not convinced of his point of view, and the fact that he doesn't involve an attorney worries me: this discussion is costing me time and money, and it feels like we're turning in circles.
It is my opinion that the consumers buying cookie spread from company C have the right to be informed that the cookies may contain nuts. Imagine that you would grant a cookie (spread) producer the freedom to remove the information "contains traces of nuts" from the list of ingredients. Such a freedom could lead to the death of a consumer who is allergic to nuts. Is this really the freedom we want?

I have the feeling that one freedom—the freedom of the potential abuser: he who wants to hide information—is weighed against another freedom—the freedom of the potential user: the person who desires to be informed. It is my personal belief that in this case, as well as in the case brought forward about iText®, the freedom to be informed outweighs the freedom to hide the truth.
You may argue that this is completely different in software, but to me, the freedom of information of a consumer is a basic right. I plead against obfuscation and secrecy, in favor of openness and transparency.

Freedom to abuse shouldn't prevail over the freedom to use.

Comments

Sat, 03/05/2011 - 18:28 — Bruno Lowagie

DFSG and Software License FAQ

I've read the Debian Free Software Guidelines and the extra term doesn't fail any of the guidelines, nor any of the tests described in the FAQ:

  1. in the actual situation, the equivalent of the message "contains traces of nuts" is automatically attached to all the cookies in a non-intrusive way. The cookie doesn't yell "Caution, I may contain nuts" before you bite it, but if you want to find out if it contains nuts, you can analyze the cookie (or the cookie spread) for the presence of nuts in a very simple way (some operations will remove the nuts; in that case, the message disappears). In any case: an extra operation is needed to remove the message, so the cookie passes the Desert Island test: there's no obligation that is impossible to meet;
  2. the message doesn't reveal the identity of the modifier, so it passes the Dissident test;
  3. the message is there to inform the consumers, it's there to help them, to protect them, so it passes the Tentacles of Evil test.

On the contrary: removal of the additional term wouldn't pass the Tentacles of Evil test, because intentionally or unintentionally removing such a message could harm a certain group of people: people allergic to nuts.
The message isn't discriminating people allergic to nuts: it's not saying they shouldn't eat the cookies; it only says the cookies "may contain traces of nuts". It's up to the consumer to decide whether or not he wants to take the risk eating such a cookie.