Drobe :: The archives
About Drobe | Contact | RSS | Twitter | Tech docs | Downloads | BBC Micro

EU software patents voted in, sort of

By Chris Williams. Published: 26th Sep 2003, 11:55:22 | Permalink | Printable

Compromise sought

Previously this month we ran an article on software patents in the EU and a controversial directive to allow 'computer implemented inventions' to be patentable. Earlier this week, the European parliament finally voted on the directive and voted in favour. Almost like the unbelievably disappointing and recently passed UK spam law, the directive is a compromise between either side of the heated debated.

The directive was re-worded so that, for instance, algorithms and business methods cannot be patented. The lovely legal spiel in the directive amendment reads, "Inventions involving computer programs which implement business, mathematical or other methods and do not produce any technical effect beyond the normal physical interactions between a program and the computer, network or other programmable apparatus in which it is run, shall not be patentable".

See below for more analysis.


European Parliament agrees on software patents from IDG Software patents limited by European Parliament vote from out-law.com

Previous: Midlands show in doubt
Next: STD announces expansion, new name


Viewing threaded comments | View comments unthreaded, listed by date | Skip to the end

vague (n) - any EU directive :-S --

 is a RISC OS UserAnimal on 26/9/03 12:16PM
[ Reply | Permalink | Report ]

This is actually very good new compared to earlier drafts, this means that patents such as the LZW algorithm (in GIF) and Amazons One-Click shopping patent can't be patented in the EU. The law although a compromise has got rid of some of the most stupid abuses.

 is a RISC OS Userflibble on 26/9/03 12:57PM
[ Reply | Permalink | Report ]

So, all in all it would be fair to say we have some measure of sanity in the directive? -- Spriteman.

 is a RISC OS UserSpriteman on 26/9/03 2:28PM
[ Reply | Permalink | Report ]

Sounds good :) Because not having these patents doesn't mean that people like amazon will stop developing their 1 click to buy thing. -- Andrew

 is a RISC OS UserAndrewDuffell on 26/9/03 3:47PM
[ Reply | Permalink | Report ]

This is actually a lot better than having no software patents at all. Companies will have to prove that their development is unique to their system, and cannot be used for a miriad of unrelated prposes. Most people are not going to bother. For once the bodge and compromise has served a purpose, and the *E*vil *U*mpire has done a good thing. -- Keep Flying

 is a RISC OS UserJWCR on 26/9/03 5:19PM
[ Reply | Permalink | Report ]

"The draft was also amended to ensure that the use of a patented technique in order to ensure interoperability between two different computer systems or networks would not be regarded as a breach of that patent"

 is a RISC OS Usermavhc on 26/9/03 7:19PM
[ Reply | Permalink | Report ]

All this is quite interesting. I like that things like 1-click would be banned, but I'm afraid that I find it disheartening that algorithms like LZW can't be patented (if what people are saying that algorithms can't be patented is true - I've just read the above linked news articles).

If I invent a new way of compressing data, or encrypting data, why shouldn't I be compensated for my time by being allowed to patent it?

-- Dougal

 is a RISC OS UserDougal on 27/9/03 1:02PM
[ Reply | Permalink | Report ]


If my memory of the LZW thing is correct the original algotithm was an academic project conducted by two Israeli computer scientists (Lempil and Ziv - hence the LZ). Welch (the W) modified the open LZ algorithm to improve efficiency - Welch worked for ICL (which became Unisys).

Although I would agree with you that patents should be allowed where a substantial breakthrough is made the case of LZW amply illustrates why it can be abused. For years people were permitted use LZW without a squeak out of Unisys (that's why it got into GIF in the first place). Then when it became universal THEN Unisys insisted on being paid.

I feel that if someone *has* a patent and allows the IP to become universally used/adopted and then (very late in the day) insists on being paid I'd say "tough"....

Patents, if properly applied, in themselves are not a bad thing but companies have a tendancy to use them sometimes in unfair or unreasonable ways.

Also they do not appear to inhibit monopolistic tactics from people such as Microsoft so really aren't as cast iron a means of protecting the lone developer or small company as they might first appear.

-- Annraoi

 is a RISC OS UserAMS on 27/9/03 2:04PM
[ Reply | Permalink | Report ]

I agree with you AMS that what Unisys did was quite a dirty trick, of letting it become a standard then announcing they owned it.

Policing patents is a pain. If you're a developer you're persumably meant to check that you're not infringing on existing patents which is a cost, as you really need to hire someone to carry out that task. On the otherhand, the owner can't be expected to police all possible products to see whether their patents are being infringed as that's hard to track and most products don't openly disclose how they work internally. It's a tough one and I don't pretend to have the answers.

In the Unisys/LZW case, well one hopes that it at least taught people to be more careful, hence why W3 have big discussions about letting patented material into open standards.

As to them not frightening off the big boys, that may be, but I think that some protection has to be better than nothing at all.

-- Dougal

 is a RISC OS UserDougal on 27/9/03 2:33PM
[ Reply | Permalink | Report ]

If you invent it independently, why should you pay?

 is a RISC OS Usermavhc on 27/9/03 3:14PM
[ Reply | Permalink | Report ]

In replay to Dougal

The other way of looking at it is that you get your reward when selling your finished product. If you look to cars, for example, the various components that make them up are generally not subject patents. So each manufacturer can impliment the parts as they wish. However, the money is made in selling the completed package.

Another way of looking at it is that you get paid when a client comes to you for your implimentation of a particular part/algorithm when your implimentation has advantages over other people's.

Of course, this model means that you are not automatically assurred financial gain for your work. Although hopefully you will because you will be first to market with it and understand it better than anyone. It also introduced competition where the best product should come through for the customers.

Right then - pick that one apart ;-) -- Spriteman.

 is a RISC OS UserSpriteman on 27/9/03 3:26PM
[ Reply | Permalink | Report ]


Indeed, Swan in England and Edison in US regarding the lightbulb would be a case in point.

The problem is *proof* of independant development. You need to be able to show unambiguously that you *did* independantly discover a new method.

Proving this (depending on how meticulous you are) can be difficult or impossible. It may also be difficult to prove to the original patent owners satisfaction.

-- Annraoi

 is a RISC OS UserAMS on 27/9/03 3:32PM
[ Reply | Permalink | Report ]


Parts in cars *are* subject to patents. The three point seat belt was invented by Volvo for example (if memory serves they actually made it royality free in order to increase vehicle safety generally).

True I don't know if "Ugh" in the third cave on the left patented the wheel with his tribal chief, but I am pretty sure some bits of cars have patents or did do.

-- Annraoi

 is a RISC OS UserAMS on 27/9/03 3:36PM
[ Reply | Permalink | Report ]

Spriteman: that's true to a degree, but the moment people can copy what you've done and start selling it they're eating into what profit you're making.

Most patent's aren't eureka moments in the bath - they're a lot of toil and have cost that is high in terms of time, effort, and money. The point of a patent is to encourage people to invest their time, effort, and money in creating something. Telling people that others will be free to duplicate what you did for nothing is not really that hot an incentive.

Like AMS says, I'd be surprised if most parts of cars aren't patented these days. In most electronics companies anything they invent at all, no matter how small, they patent, so I don't see why other industries would be any different, but then I have no idea of how the car industry works.

Also, the point of patents is that it provides a mechanism for people to sell the research they did. Other people can use my invention, if they pay for it so I can make a return on all the effort required - why should the other people get it for free by default when it cost me so much effort?

-- Dougal

 is a RISC OS UserDougal on 27/9/03 3:52PM
[ Reply | Permalink | Report ]

Easy to prove, just record your life onto harddiscs 24hrs a day.

If something requires loads of work, then fine, it's patents for stupid things people dislike, or things everyone already knew about.

Should patented stuff be allowed in standards like MPEG? If the owner promises that no royalties will be charged for software given away for free?

 is a RISC OS Usermavhc on 27/9/03 4:20PM
[ Reply | Permalink | Report ]

mavhc - that's up to the controller of the standard isn't it? If the owner of the patent can provide a suitable contract stating that for use in the standard they'll never charge will people be happy?

I agree that there is some form of distiction needed between inventions that required effort and have produced something novel as opposed to "stupid things" - the "people dislike" part is somewhat orthoganol - I image only the holder of the patent likes the fact that they have it :)

-- Dougal

 is a RISC OS UserDougal on 27/9/03 4:27PM
[ Reply | Permalink | Report ]

mavhc> MPEG is licensed by MPEG LA, so whether MPEG contains patented material or not is moot, if you wish to use it commercially you are required to pay.

The problem (and I think MPEG/LZW fit into this) is that both were widely available, appeared opened and then subsequently had "licensing clauses" added.

This degree of control could allow some less ethical companies to exploit their patents to put others out of business. If MPEG LA charge reasonable amounts then a RISC OS developer *could* license it and be legit, but let's say MS put the lean on MPEG LA to (say) ensure that everyone other than MS paid $100 per copy sold with an upfront $200,000 license fee then it highly unlikely MPEG would (or could) be licensed to smaller platforms. This would allow MS to entrench it's monopoly.

Note that the above is "hypothetical", but is not impossible.

A simple benchmark should be that for a patent to be enforsable the patent holder *must* be willing to permit licensing to *all* platforms with no discrimatory prices (preferably with no up front fee but a per copy charge).

If a patent holder refuses to do so then the patent should be considered "unenforced" on that platform and any developer should be allowed develope software that can behave like the patented code (in short if the patent holder won't deal with users on a specific platform then he/she can't expect to receive any income from and should *not* be allowed litigate against anyone implementing software on that platform).

-- Annraoi

 is a RISC OS UserAMS on 27/9/03 4:40PM
[ Reply | Permalink | Report ]

AMS - the problem with the platform clause is one of time. How do you show that someone has no intention of licencing it for a platform rather than they're just working on it or negotiating with someone to work on it? Or that no one on that platform would agree to the patent holder's terms?

That would be a nightmare rule to try and enforce. Companies would just put up some pretence that show's they "working on it" rather than just say no.

-- Dougal

 is a RISC OS UserDougal on 27/9/03 5:05PM
[ Reply | Permalink | Report ]

Dougal wrote:

"...I find it disheartening that algorithms like LZW can't be patented (if what people are saying that algorithms can't be patented is true - I've just read the above linked news articles)."

It looks that way. Mathematical methods are ruled out, so it looks like algorithms are, too. It seems like past rulings have reinforced that position.

Dougal also wrote:

"If I invent a new way of compressing data, or encrypting data, why shouldn't I be compensated for my time by being allowed to patent it?"

Patent applications have to be pretty specific and your method would have to have a technical effect beyond the abstract. The trouble is the possibility that, although you're intending to use your method within a specific domain, it has universal applicability. If that sort of thing was allowed, you could end up with those who hold key patents stifling competition and progress.

It's a balancing act. By granting a party monopoly rights to exploit their invention, society has to hope that those rights were earned. There is an argument that patent protection, if granted, offers an incentive/guarantee to investors which they couldn't otherwise expect; the alternative to such protection may be the use of trade secrets, or the lack of such investment in the first place. However, software-related development is such an extreme environment in which to apply the rules for patents that it's possible to argue that you can't successfully do so. Timescales are shorter, more people are working in similar fields and there aren't effective methods for searching for prior art or claims.

In the end, it's important to realise that extensions to patentability are, in effect, extensions to the rights of a group of parties at the expense of their competitors, and everyone else. To argue for that, you have to have a pretty strong case. Or maybe you just need to lobby really hard. :-(

 is a RISC OS Userdavidb on 28/9/03 1:35AM
[ Reply | Permalink | Report ]

As pointed out above time scales are shorter, so what about a reduced patent length for computer related stuff?

On the other hand, why does copyright run out, why can't authors leave their desendants an inheritance from royalties?

 is a RISC OS Usermavhc on 28/9/03 10:12AM
[ Reply | Permalink | Report ]

I think that the point about timescales was put forward by a number of the respondents to the UK Patent Office consultation on software patentability.

I don't know whether any of the people who are desperate to implement this sort of legislation are willing to consider shorter periods for patent protection. This might lead to cases where attempts would be made to "trade up" patents in order to increase their potential lifespan.

I think the copyright and patent issues are related in a broader sense: in both cases, you are granted temporary rights to exploit your work. Although, in the case of copyright, you obviously don't need to apply for this.

As I said before, this is all a balancing act intended to reward both "inventors" and society.

 is a RISC OS Userdavidb on 28/9/03 6:45PM
[ Reply | Permalink | Report ]

Please login before posting a comment. Use the form on the right to do so or create a free account.

Search the archives

Today's featured article

  • RISC OS filename translation
    Being understood in the outside world
     31 comments, latest by mrchocky on 09/01/04 11:06PM. Published: 2 Jan 2004

  • Random article

  • Archive woos ex-users with VirtualRPC
    Stopped using RISC OS? You'll want this emulator, mate
     12 comments, latest by martin on 12/1/04 9:48PM. Published: 10 Jan 2004

  • Useful links

    News and media:

    Top developers:
    RISCOS LtdRISC OS OpenMW SoftwareR-CompAdvantage SixVirtualAcorn

    CJE MicrosAPDLCastlea4X-AmpleLiquid SiliconWebmonster


    RISCOS.org.ukRISCOS.orgRISCOS.infoFilebaseChris Why's Acorn/RISC OS collectionNetSurf

    Non-RISC OS:
    The RegisterThe InquirerApple InsiderBBC NewsSky NewsGoogle Newsxkcddiodesign

    © 1999-2009 The Drobe Team. Some rights reserved, click here for more information
    Powered by MiniDrobeCMS, based on J4U | Statistics
    Page generated in 0.2118 seconds.