[Patents] Why Greg Aharonian is wrong

[halber@timsi.com.pl (Jozef Halbersztadt)]

I am going to contest two main thesis of Greg Aharonian. He claims that 
patent offices can work better. He claims too that in all fields of 
technology there are problems with patents. Patent offices can work better as 
the US PTO works, because work of the PTO is often a disaster. The proposed 
by Greg remedies were from start applied in the EPO. In the EPO examiners 
have smaller load as in the PTO, higher wages, better databases and even 
better procedures (publication before granting). The result cannot be 
ignored. For example look for a rough analysis of Altavista indexing patents: 
30+ were granted in the US, 2 were granted in the EPO. Nevertheless, it does 
not make difference big enough. Why, because any patent office grants as a 
rule and rejects as an exception. Alternative approach is unrealizable. Do 
you think that a public institution that rejects 90% applications can 
survive? In less developed countries there are powerful arguments for such an 
economical policy. But even there I cannot imagine reason for putting the 
brightest people into the patent office. The brightest are more needed as 
inventors, university teachers, entrepreneurs, and even political leaders. 
And when you accept that patent office personnel should be the second rate 
quality, it is reasonable not to give them too much power.

The second Greg's thesis - problems with patents are everywhere. Yes, but 
they are not the same problems. In the real world with must live with a 
distance between a real technical solution and its description. With 
sentences composed from words it is possible to describe essence of the 
solution in fields of technology like mechanics or chemistry. For them the 
distance between solution and description isn't big. It is because of help of 
drawings in mechanics and chemical symbols in chemistry. The problem is more 
difficult with electrical circuits. If the number of elements is above 20 or 
40 the relations between them can become complicated. It is becoming 
impractical to describe their behavior in natural language. Drawings don't 
help much. And patent specifications, patent claims are written in natural 
language. In result applicants don't offer precise descriptions and patent 
offices are giving up from executions of all requirements. Computer related 
solutions create to examiners similar problems as complicated circuitry. Only 
the magnitude of the problems is significantly higher.

If one seriously thinks about improving quality of examination the first 
question mark should be put on language of description. The natural language 
should be supplemented by a special language specific for particular 
technology. In chemistry chemical symbols play the role of a specific 
language, in mechanics drawings with construction. As concerns software 
patents it is reasonable to propose that the natural language should be 
supplemented by the code. Such proposal was announced in the US a few years 
ago. It was very much market orientated approach proposed for proprietary 
commercial software. It has only one feature disliked by big business - 
simple compulsory license.

This proposal is very important. It is moving software patenting from field 
of fiction to the field of professional analysis. It doesn't eliminate 
disliked by participants of this debate paradigm of awards due to priority. 
But as I understand nobody on this list is opposing a system that is neutral 
to the division open/closed source. It is important to realize that thanks 
compulsory licenses, thanks treating code as language the proposed register 
system can be easy revised to accommodate open source needs. Only the license 
have to be slightly different for open source applications. That is why 
Eurolinux is mentioning the proposal in the letter to the European 
Commission.: "In case special anti-cloning protection is really demanded by 
the software industry, as the DGIM claims in its invitation paper, Mark 
Paley's Model Software Patent Act could provide a useful source of 
inspiration." More: http://www.eurolinux.org/news/pr0101/dgltr

The Software Patent Act should be put forward vigorously. It would eliminate 
90% of silly software patents. Europe should give the proposal the second 
life. It is fully compatible with TRIPS. Any challenge by the US government 
in the WTO would end with a failure.

Józef Halbersztadt

I am going to contest two main thesis of Greg Aharonian. He claims that patent offices can work better. He claims too that in all fields of technology there are problems with patents. Patent offices can work better as the US PTO works, because work of the PTO is often a disaster. The proposed by Greg remedies were from start applied in the EPO. In the EPO examiners have smaller load as in the PTO, higher wages, better databases and even better procedures (publication before granting). The result cannot be ignored. For example look for a rough analysis of Altavista indexing patents: 30+ were granted in the US, 2 were granted in the EPO. Nevertheless, it does not make difference big enough. Why, because any patent office grants as a rule and rejects as an exception. Alternative approach is unrealizable. Do you think that a public institution that rejects 90% applications can survive? In less developed countries there are powerful arguments for such an economical policy. But even there I cannot imagine reason for putting the brightest people into the patent office. The brightest are more needed as inventors, university teachers, entrepreneurs, and even political leaders. And when you accept that patent office personnel should be the second rate quality, it is reasonable not to give them too much power.

The second Greg's thesis - problems with patents are everywhere. Yes, but they are not the same problems. In the real world with must live with a distance between a real technical solution and its description. With sentences composed from words it is possible to describe essence of the solution in fields of technology like mechanics or chemistry. For them the distance between solution and description isn't big. It is because of help of drawings in mechanics and chemical symbols in chemistry. The problem is more difficult with electrical circuits. If the number of elements is above 20 or 40 the relations between them can become complicated. It is becoming impractical to describe their behavior in natural language. Drawings don't help much. And patent specifications, patent claims are written in natural language. In result applicants don't offer precise descriptions and patent offices are giving up from executions of all requirements. Computer related solutions create to examiners similar problems as complicated circuitry. Only the magnitude of the problems is significantly higher.

If one seriously thinks about improving quality of examination the first question mark should be put on language of description. The natural language should be supplemented by a special language specific for particular technology. In chemistry chemical symbols play the role of a specific language, in mechanics drawings with construction. As concerns software patents it is reasonable to propose that the natural language should be supplemented by the code. Such proposal was announced in the US a few years ago. It was very much market orientated approach proposed for proprietary commercial software. It has only one feature disliked by big business - simple compulsory license.

This proposal is very important. It is moving software patenting from field of fiction to the field of professional analysis. It doesn't eliminate disliked by participants of this debate paradigm of awards due to priority. But as I understand nobody on this list is opposing a system that is neutral to the division open/closed source. It is important to realize that thanks compulsory licenses, thanks treating code as language the proposed register system can be easy revised to accommodate open source needs. Only the license have to be slightly different for open source applications. That is why Eurolinux is mentioning the proposal in the letter to the European Commission.: "In case special anti-cloning protection is really demanded by the software industry, as the DGIM claims in its invitation paper, Mark Paley's Model Software Patent Act could provide a useful source of inspiration." More: http://www.eurolinux.org/news/pr0101/dgltr

The Software Patent Act should be put forward vigorously. It would eliminate 90% of silly software patents. Europe should give the proposal the second life. It is fully compatible with TRIPS. Any challenge by the US government in the WTO would end with a failure.

Józef Halbersztadt