If you process XML, you are a patent violator

[xavier]

http://www.businessweek.com/technology/ ... 416170.htm

I sh*t you not.

http://patft.uspto.gov/netacgi/nph-Pars ... =5,787,449

When is the USPTO going to hire people with the least bit of actual knowledge of technology?

[betajaen]

I was reading this on slashdot.org earlier.

I assume this is just US only. Which means Microsoft can still sell Word and Office outside the United States. Although, I can't imagine if they do stop selling Word, it won't be long until businesses, universities, and other organisations suffer because of the the amount of integration in the work place with Microsoft Office (everyone uses Word for documents) - then the decision has to be reversed.

Anyway, we are all programmers and artists here. So I believe I can all agree on that we believe that all distrubtion of software should be only be control by copyrights and legal binding documents and not silly patents.

[xavier]

Not a clue how it affects international versions; I doubt that if MS is going to fix the US version that they aren't touching a piece of code that exists in all versions.

It's more about this one tiny feature -- they are just removing it and moving on (apparently no one really used it, it was something left over from MS's hyperbolic "XML >> all" days). But yeah, it's amazing to see that patent trolls still exist; I thought SCO was the last of them. And scary that you can patent something so obvious; a bit of obfuscation in the text of the patent application and the fact that no one there has a clue about anything technological (even though they are creating patents about TECHNOLOGY....) and *bang* you've got your own financial IED that you can trigger on any unsuspecting big-pocket company that actually *makes* things....

The entire process is broken. It's not 1790 anymore, yet the fundamental rules haven't changed a bit since the US patent process was created.

[betajaen]

Well, I was referring to if Microsoft stops selling Word in the US, would it stop selling it in the rest of the world? Probably no.

The Eolas patent annoys me the most; We all knew their was prior art and how trivial it was to implement, but in the end how much money Microsoft had to pay out was insane, and all it does is fuel more patents trolls.

In a way, it is better Microsoft to be targeted for this rather than a popular piece of open source software. The EFF may be good, but Microsoft has a lot more money for the lawyers.

[xavier]

The problem with software patents is that Microsoft (one of the few with the casual resources to fight to eliminate them) is that MS themselves files software patents at as furious a pace as anyone. It's long since stopped being about allowing the little guy to benefit from his/her invention(s), and is all about stifling competition through abuse of the process. i4i is no better, since I put patent trolls in the same category as patent abusers. What to do, what to do...

[antonl]

Have you read the patent? I scanned it quickly - I hate reading the odd mixture of legalese and technical jargon in patents - it seems to me to be a rather neat thing these guys came up with, and has nothing to do with XML as such. From my read they devised a method for storing the raw data contained within an SGML document separate from the SGML tags, and with the capability of having several maps of SGML tags.

So if I read the patent right you can modify the raw data without consideration for placement of tags, and then save the tagged data, in any number of different document formats, without resorting to XSLT or something similar.

As for where the patent is applicable: Clearly the US, but who knows where else, but it is no uncommon for patents to only apply in a few select countries, because of the horrendous cost of having a truly worldwide patent.

[antonl]

What to do, what to do...

The patent officers I've met all seemed very competent to me, but overburdened; raising the bar for what can be patented seems to be a good solution, but in reality, once an invention is made it is often very difficult to determine how novel and great it is. A lot of inventions have a good deal of "obviousness" once people are told the solution to the problem; and banning software patents completely doesn't seem like a good solution to me either, let me give an example from my country:

A company is producing hearing aids, and they come up with a completely new algorithm for filtering out noise, the hearing aid uses a low powered dsp, to save battery life, and intelligently tries to filter out back-ground noise, e.g. when in a conversation, only the voices of nearby people are amplified and frequency modulated to match the hearing impairment of the wearer, when in a concert it "decides" to amplify and modulate the music and treat nearby people as noise, which should be ignored.

The only thing which sets this hearing aid apart from other hearing aids is the software controlling the DSP, in effect a software patent, is needed to protect the company from plagiarism.

[jjp]

The only thing which sets this hearing aid apart from other hearing aids is the software controlling the DSP, in effect a software patent, is needed to protect the company from plagiarism.

No, it is not. Their software implementation is protected by copyright, and they do not have to make it public. That should be enough. And what if someone independently discovers the same algorithm, or something very similar. Why should he have to pay patent fees?

[mkultra333]

My favourites are the patents on perfect, lossless compression. The authors claimed it could compress any data set to a smaller size, losslessly. The patent was examined for 3 years before being granted. Pity such a thing is mathematically impossible. http://gailly.net/05533051.html

I hate software patents, and I hate patents on naturally occuring data like DNA.

[betajaen]

The patent officers I've met all seemed very competent to me, but overburdened;

I suppose a good way, is just to have a condition if you file for a patent, then you have to reproduce a working physical device that your patent describes and perhaps even demonstrate it.

New better braking system for car wheels? Then bring it in to the patent office.
A magical machine that goes ping!, to stop entropy? Well bring it in!
A process to read XML from the web? Where is this XML, Can I touch it?

My point is I believe you file for a patent, then your patent should not be a concept in your head. You should show proof of a working device. For those who patenting a physical machine that shouldn't be a problem, your bound to made a prototype now. Evidence of it working; videos or a live demonstration would certainty speed up the application process. Software isn't a physical thing, even though we store it on physical mediums. Copyright is enough to keep it being stolen or copied from nasty people.

[antonl]

I hate software patents, and I hate patents on naturally occuring data like DNA.

It is and endless and open discussion whether anything is really ever invented, or merely discovered.

[antonl]

My point is I believe you file for a patent, then your patent should not be a concept in your head.

The problem here is that you do not patent a thing, you never patented a thing, not even in 1790, you patent an idea, that's the whole point of immaterial property.

[antonl]

No, it is not. Their software implementation is protected by copyright

Copyright is just that, who has the right to make a copy, but there is ample room to be "inspired" by other original works.

, and they do not have to make it public.

Reverse engineering is not that difficulty, security through obfuscation is never a good idea

That should be enough. And what if someone independently discovers the same algorithm, or something very similar. Why should he have to pay patent fees?

Independent discoveries very seldom are independent, ask any anthropologist about that , and that problem is not only related to any kind of patents, but to the all immaterial property issues, including copyright.

I know I'm sounding like a pro-patent guy, which I'm really not, I'm quite skeptical about - especially abuse of - patents, I just don't see any easy solutions.

[mkultra333]

At the very least, I think patents on algorithms and programs should be much shorter than for physical machines. I could probably live with that. Patenting was designed for much longer engineering turn-arounds, not the accelerated development that software goes through.

No software patents at all is preferrable though.

[syedhs]

I think apart from the points raised by some, the reason of patent in the first place is to protect people from copying your idea after probably years of researching, tweaking etc. Other people may see your invention (after few years of R&D), and then can copy it almost instantly. For an example, you design a very nice chair - comfortable yet look very elegance and doesn't cost much and can be easily taken apart (mobility). And you are sure it is unique, and you patent it. For that matter it has a very valid reason to patent.

As opposed to software, you can create a new software with totally unique algorithm it, but you choose not to publish the algo and distribute only the binaries. Who the heck is going to copy you? And to copy fast enough to actually cause harm to your business. For an example, does anybody here know *exactly* the algo on how Google deliver the searches for you? I know the concept, but it is too high level for you to actually get down and write the similar code and eventually, tadaaa... I have a Google clone...

[jjp]

Independent discoveries very seldom are independent, ask any anthropologist about that , and that problem is not only related to any kind of patents, but to the all immaterial property issues, including copyright.

Ideas and concepts that occur to only one individual are also very seldom. To name something popular: neither did James Watt invent the steam engine, nor was Einstein the first and only one to think of space-time, nor was Darwin the only one to discover the concept of evolution. If you look close enough you'll find that ideas and concepts are something that works via evolution.