« Absurdities | Main | Chicken Little and Software Patents »
January 14, 2005
One Reason Why Software Is Patentable and Should Not Be Copyrightable
There are two broad categories of statutorily protected creations, those that are of artistic value and those of utility.
In the area of artistic creations, the greatest portion of the invention is in the expression. For example, the effort to produce a superbly written, though not well thought out, blog posting is almost all in the expression of the idea. The effort, and thus the protection, is centered around the effort it took to create the specific embodiment, and a copyright is in order.
Ideas that are inherently useful are protected by patent, and to a much lesser extent by copyright. Because the usefulness of the idea can have great commercial value, the protection of the idea is much more limited than copyright in terms of length, and much more difficult to enforce.
Looking at software, there are some who believe that copyright affords the best protection. But where does software fit in the spectrum?
Software looks like written text. It is a compilation of characters that expresses an idea on some level. In fact, it is a very particular and exacting way to express an idea. So much so that by changing one of a million characters in a software program may cause it to completely malfunction.
Software is ‘written’ and can be easily ‘copied.’ In many aspects, software looks and acts like this posting: an artistic expression of some idea.
However, when the superficial cover is thrown off and the underpinnings exposed, software does not behave like the text of a book, but more like the elements of a machine.
Suppose I were to develop a new machine to manufacture an existing product. Let’s say the machine makes the products ten times faster because it uses an innovative way of handling and processing the raw materials. Very few people would argue that the machine would not be patentable, provided it met all the legal requirements, even though the machine were composed completely from thousands of off the shelf parts.
Each of the individual parts in the machine were previously invented by someone else and may even have some patent protection. If I use a patented bearing system from one company, part of the cost of the bearing is a royalty payment or premium for the patented aspect of the part. However, I choose that product and pay the difference because it is better than any alternative I can find.
Suppose I also write a software application that processes data ten times faster than any prior embodiment. The software application is made up from many thousands of existing pieces of prior art, such as routines and functions in an open or closed source library. In some cases, I might buy a specialized library of routines that I know is well tested and designed to do a specialized task I require. In other cases, I figure out how to do a function myself and roll my own code.
In the end, the software application and machine perform a utilitarian function, not an aesthetic function. It is no different to select a bearing system from a catalog than it is to select a routine from a library. When I build the machine, I may be building on a thousand different ideas, a thousand different patents. Similarly, my software application is built on a thousand different ideas, some of which may be patented and some not. For the patentable ones, I am willing to pay for the ideas.
Someone made this argument in a comment:
[T]he fact that in the software world, so many individual ideas (most of which can be covered by a patent) are combined in every single product, that it's just not worth it. For small producing entities, the costs simply outweigh the benefits to be had.
We operate in a thicket of patented ideas in every field of endeavor, but the business world has found a way to make it tolerable. Patented ideas tend to cost a little more because of the royalties, but people buy them because they are better. If the ideas were not better, nobody would buy them.
The bottom line of this posting: software is much more like the machine than it is artistic expression, as long as you get past the unfortunate labels used with software.
Attacks on software patents are attacks on utility patents in general. No matter how ‘focused’ the propaganda machine is toward software, the real goal, whether intentional or not, will lead to the abolition of all utility patents.
Posted by krajec at January 14, 2005 01:39 PM
Trackback Pings
TrackBack URL for this entry:
http://www.krajec.com/mt/mt-tb.pl/106
Comments
When people prefer copyright to patents for protection of software, the reason is not necessarily because they feel they are artists or authors. They generally don't care about the original intention of the various protection means, but simply look at those properties that matter in real life: breadth, time to get the protection, cost.
And then you see that patents are quite broad, take quite a while to be granted (at least compared to the rate of innovation in the software field), and cost quite a bit of money. Copyright on the other hand, is fairly narrow, instantaneous and free.
It's true that in principle, copyright protects only the expression and not the idea. That's part of the narrowness (you don't get a broad monopoly, but on the other hand other's don't either), but at the same time it also goes further than literal copying (witness the trouble that Phoenix went through to re-implement the IBM PC Bios).
You are correct however that this idea/expression dichotomy causes a lot of misconceptions, and some sort of "industrial copyright" system which does not focus on ideas and expressions, but merely on the parameters that are really important for businesses (scope, speed, cost), could alleviate these concerns.
Regarding the thickets: it's true they are not unique to software. On the other hand, software development is "cheap" and "easy". Makers of electronics are not moving more and more functionality into software because it's more expensive or more difficult that way.
So what do people do when something is easier? They make it bigger and combine more stuff resulting in really nifty things, so in the end the amount of complexity and work evens out. The net result however is that a single "product" contains a lot more different ideas in the latter case than in the former cases.
There is a whole spectrum, and depending on how hard it is to combine things (or to predict the outcome of combining things), thickets can be more or less of a problem. Pharmaceuticals are on one end of the spectrum, software on the other. That's why the patent thicket problem is considered to be more problematic in the software field than elsewhere.
The fact that the business world has found ways to make patent thickets tolerable, does not mean that their negative effects should be discarded. It's simply one way the patent system introduces inefficiency in the market (transaction costs). And those thickets make it more difficult for small players to enter the market.
Since software development does not per definition necessitate huge investments in material (only human capital is important, while you can't setup a pharma lab with $US 1000), there are a lot of small players in the software world (especially in Europe). Patents are an extra barrier to enter the market place, so again, unless there are indications that without patents the software market is highly dysfunctional, they should not be codified.
The "If the ideas were not better, nobody would buy them" statement does not always hold true. You undoubtedly have already heard about networking effects: a computer program is never used in isolation. If it can't exchange data with the (de facto or official) standards, it's worthless.
Posted by: Jonas Maebe at January 15, 2005 05:34 AM