limit-patent-effect.html (9789B)
1 <!--#include virtual="/server/header.html" --> 2 <!-- Parent-Version: 1.96 --> 3 <!-- This page is derived from /server/standards/boilerplate.html --> 4 <!--#set var="TAGS" value="essays laws patents" --> 5 <!--#set var="DISABLE_TOP_ADDENDUM" value="yes" --> 6 <title>Giving the Software Field Protection from Patents 7 - GNU Project - Free Software Foundation</title> 8 <!--#include virtual="/philosophy/po/limit-patent-effect.translist" --> 9 <!--#include virtual="/server/banner.html" --> 10 <!--#include virtual="/philosophy/ph-breadcrumb.html" --> 11 <!--GNUN: OUT-OF-DATE NOTICE--> 12 <!--#include virtual="/server/top-addendum.html" --> 13 <div class="article reduced-width"> 14 <h2>Giving the Software Field Protection from Patents</h2> 15 16 <address class="byline">by <a href="https://www.stallman.org/">Richard 17 Stallman</a></address> 18 19 <p>Patents threaten every software developer, and the patent wars we have 20 long feared have broken out. Software developers and software 21 users—which, in our society, is most people—need software 22 to be free of patents.</p> 23 24 <p>The patents that threaten us are often called “software 25 patents,” but that term is misleading. Such patents are not 26 about any specific program. Rather, each patent describes some 27 practical idea, and says that anyone carrying out the idea can be 28 sued. So it is clearer to call them “computational idea 29 patents.”</p> 30 31 <p>The US patent system doesn't label patents to say this one's a 32 “software patent” and that one isn't. Software developers 33 are the ones who make a distinction between the patents that threaten 34 us—those that cover ideas that can be implemented in 35 software—and the rest. For example, if the patented idea is the 36 shape of a physical structure or a chemical reaction, no program can 37 implement that idea; that patent doesn't threaten the software field. 38 But if the idea that's patented is a computation, that patent's barrel 39 points at software developers and users.</p> 40 41 <p>This is not to say that computational idea patents prohibit only 42 software. These ideas can also be implemented in hardware—and 43 many of them have been. Each patent typically covers both hardware 44 <em>and</em> software implementations of the idea.</p> 45 46 <h3>The Special Problem of Software</h3> 47 48 <p>Still, software is where computational idea patents cause a special 49 problem. In software, it's easy to implement thousands of ideas 50 together in one program. If 10 percent are patented, that means hundreds of 51 patents threaten it.</p> 52 53 <p>When Dan Ravicher of the Public Patent Foundation studied one large program 54 (Linux, which is the kernel of the 55 <a href="/gnu/gnu-linux-faq.html"> GNU/Linux</a> operating system) in 56 2004, he found 283 US patents that appeared to cover computing ideas 57 implemented in the source code of that program. That same year, a 58 magazine estimated that Linux was .25 percent of the whole GNU/Linux system. 59 Multiplying 300 by 400 we get the order-of-magnitude estimate that the 60 system as a whole was <em>threatened by around 100,000 patents</em>.</p> 61 62 <p>If half of those patents were eliminated as “bad 63 quality”—mistakes of the patent system, that 64 is—it would not really change things. Whether 100,000 patents 65 or 50,000, it's the same disaster. This is why it's a mistake to 66 limit our criticism of software patents to just “patent 67 trolls” or “bad quality” patents. The worst patent 68 aggressor today is Apple, which isn't a “troll” by the 69 usual definition; I don't know whether Apple's patents are “good 70 quality,” but the better the patent's “quality” the 71 more dangerous its threat.</p> 72 73 <p>We need to fix the whole problem, not just part of it.</p> 74 75 <p>The usual suggestions for correcting this problem legislatively 76 involve changing the criteria for granting patents—for instance, 77 to ban issuance of patents on computational practices and systems to 78 perform them. This approach has two drawbacks.</p> 79 80 <p>First, patent lawyers are clever at reformulating patents to fit 81 whatever rules may apply; they transform any attempt at limiting the 82 substance of patents into a requirement of mere form. For instance, 83 many US computational idea patents describe a system including an 84 arithmetic unit, an instruction sequencer, a memory, plus controls to 85 carry out a particular computation. This is a peculiar way of 86 describing a computer running a program that does a certain 87 computation; it was designed to make the patent application satisfy criteria 88 that the US patent system was believed for a time to require.</p> 89 90 <p>Second, the US already has many thousands of computational idea 91 patents, and changing the criteria to prevent issuing more would not 92 get rid of the existing ones. We would have to wait almost 20 years 93 for the problem to be entirely corrected through the expiration of 94 these patents. We could envision legislating the abolition of these 95 existing patents, but that is probably unconstitutional. (The Supreme 96 Court has perversely insisted that Congress can extend private 97 privileges at the expense of the public's rights but that it can't go 98 in the other direction.)</p> 99 100 <h3>A Different Approach: Limit Effect, Not Patentability</h3> 101 102 <p>My suggestion is to change the <em>effect</em> of patents. We 103 should legislate that developing, distributing, or running a program 104 on generally used computing hardware does not constitute patent 105 infringement. This approach has several advantages:</p> 106 107 <ul> 108 <li>It does not require classifying patents or patent applications as 109 “software” or “not software.”</li> 110 <li>It provides developers and users with protection from both existing 111 and potential future computational idea patents.</li> 112 <li>Patent lawyers cannot defeat the intended effect by writing 113 applications differently.</li> 114 </ul> 115 116 <p>This approach doesn't entirely invalidate existing computational idea 117 patents, because they would continue to apply to implementations using 118 special-purpose hardware. This is an advantage because it eliminates 119 an argument against the legal validity of the plan. The US passed a 120 law some years ago shielding surgeons from patent lawsuits, so that 121 even if surgical procedures are patented, surgeons are safe. That 122 provides a precedent for this solution.</p> 123 124 <p>Software developers and software users need protection from patents. 125 This is the only legislative solution that would provide full 126 protection for all. We could then go back to competing or 127 cooperating… without the fear that some stranger will wipe away 128 our work.</p> 129 130 <div class="comment" role="complementary"> 131 <p><em>See also: 132 <a href="/philosophy/patent-reform-is-not-enough.html"> 133 Patent Reform Is Not Enough</a></em></p> 134 </div> 135 136 <div class="infobox extra" role="complementary"> 137 <hr /> 138 <p>A version of this article was first published at 139 <a href="https://www.wired.com/opinion/2012/11/richard-stallman-software-patents/"> 140 <cite>Wired</cite></a> in November 2012.</p> 141 </div> 142 </div> 143 144 145 </div><!-- for id="content", starts in the include above --> 146 <!--#include virtual="/server/footer.html" --> 147 <div id="footer" role="contentinfo"> 148 <div class="unprintable"> 149 150 <p>Please send general FSF & GNU inquiries to <a 151 href="mailto:gnu@gnu.org"><gnu@gnu.org></a>. There are also <a 152 href="/contact/">other ways to contact</a> the FSF. Broken links and other 153 corrections or suggestions can be sent to <a 154 href="mailto:webmasters@gnu.org"><webmasters@gnu.org></a>.</p> 155 156 <p><!-- TRANSLATORS: Ignore the original text in this paragraph, 157 replace it with the translation of these two: 158 159 We work hard and do our best to provide accurate, good quality 160 translations. However, we are not exempt from imperfection. 161 Please send your comments and general suggestions in this regard 162 to <a href="mailto:web-translators@gnu.org"> 163 <web-translators@gnu.org></a>.</p> 164 165 <p>For information on coordinating and contributing translations of 166 our web pages, see <a 167 href="/server/standards/README.translations.html">Translations 168 README</a>. --> 169 Please see the <a 170 href="/server/standards/README.translations.html">Translations README</a> for 171 information on coordinating and contributing translations of this article.</p> 172 </div> 173 174 <!-- Regarding copyright, in general, standalone pages (as opposed to 175 files generated as part of manuals) on the GNU web server should 176 be under CC BY-ND 4.0. Please do NOT change or remove this 177 without talking with the webmasters or licensing team first. 178 Please make sure the copyright date is consistent with the 179 document. For web pages, it is ok to list just the latest year the 180 document was modified, or published. 181 182 If you wish to list earlier years, that is ok too. 183 Either "2001, 2002, 2003" or "2001-2003" are ok for specifying 184 years, as long as each year in the range is in fact a copyrightable 185 year, i.e., a year in which the document was published (including 186 being publicly visible on the web or in a revision control system). 187 188 There is more detail about copyright years in the GNU Maintainers 189 Information document, www.gnu.org/prep/maintain. --> 190 191 <p>Copyright © 2012, 2013, 2021 Free Software Foundation, Inc.</p> 192 193 <p>This page is licensed under a <a rel="license" 194 href="http://creativecommons.org/licenses/by-nd/4.0/">Creative 195 Commons Attribution-NoDerivatives 4.0 International License</a>.</p> 196 197 <!--#include virtual="/server/bottom-notes.html" --> 198 199 <p class="unprintable">Updated: 200 <!-- timestamp start --> 201 $Date: 2021/09/19 16:26:24 $ 202 <!-- timestamp end --> 203 </p> 204 </div> 205 </div><!-- for class="inner", starts in the banner include --> 206 </body> 207 </html>