close this bookVolume 4: No. 29
View the documentPolitics and funding
View the documentPrivacy and security
View the documentCopyright law
View the documentSoftware patents
View the documentInformation collections
View the documentJob opportunities
View the documentInternet resources
View the documentJournals and e-journals
View the documentDiscussion groups
View the documentEducation
View the documentLiterature and music
View the documentMathematics and optimization
View the documentBiology and robotics
View the documentComputists' news

NIST will pay an MIT computer scientist royalties for two patents related to Clipper-chip technology "as the best way to avoid future legal concerns." [WSJ, 7/12/94, B6. EDUPAGE.]

Greg Aharonian is soliciting comments on a collection of obvious or non-novel patents that he will submit to the PTO hearings in a few weeks. Write to for the 1100-line list. [, 7/10/94.]

(The patents make interesting reading, but I find them less obvious or shocking than Aharonian does. (He's more spun up on this than I am, having collected and studied a very large database of software patents.) If a common technology is applied to a new application domain -- such as an expert-system consulting program for selecting a house roofing system -- does the developer deserve patent protection? I say yes, if the claims are specific to the particular application and solution method, and if other methods could be used. (In this case, the 1990 patent application covers only hard-wired logic applied in a three-stage questioning process. Plenty of room for other developers.) I also see nothing wrong with IBM's claims for user-interface widgets. (One of them allows a one-of-many "radio button" selection to become a multiple choice if a designated key is pressed. I've often wished for multiple-choice capability in my Panorama spreadsheet/database program, but I never thought of it in quite that way. Obviousness after the fact does not preclude patent protection.) Other patents are for specific variations on common techniques -- but so what? The fact that compilers or debuggers are old news doesn't make new variations unpatentable -- even for trivial variations. (There are more patents for bicycle parts than for automobile technology.) The patent office should _publish_ applications for general comment, rapid spread of technology, and so that programmers won't accidentally use -- and not license -- methods that later receive patent protection. But the basic idea is sound: protect specific extensions of technology or extensions of existing technology to new applications. Instead of fighting the system, we should be lining up to get our share -- for the good of society via cataloging of applied algorithms, increased commercial use, and continued rapid progress. Patents will help take software development from hobby methodology to an industry comparable to electronic design.)