The downward spiral ends here

Well, it's clear at least that this whole area is a complex issue πŸ™‚ I've noticed there is a tendency among some people leaving comments to confuse the comments with what I wrote. I am not responsible for what is written in the comments here, except to the extent that I don't delete them (which so far I have managed to refrain from doing :-))

It also seems that the discussion in the comment area after awhile always returns to the same tired themes of rhetoric about what is better, who is evil, whatever. Please, if you want to debate that, go somewhere else. I’ll take comments on topic, even (especially?) ones I don’t agree with, but let’s try to avoid philosophical battles and attacks. Also, before you make a belligerent retort about something you think I wrote - please check that I actually wrote it. It will save you from being jumped on yourself.

It is also clear that many people are not really familiar with exactly how patents work. For example, you can't simply patent "word processor" or “BASIC” - the actual patent has to be far more detailed than that, and you have to make specific claims about a design. Since just about anything that is developed is built on the work of others before, there is essentially no way an entire device or product can be patented. Much more common is that the new, unique, creative part of the new product that sets it apart from others is patented. Also, not every great idea can be patented – there are a set of criteria that have to be met – a big one is that it has to be “not obvious to one skilled in the art”. Note that means “not obvious *before* someone explains it to you”. Many things are obvious in hindsight. The idea also has to be a device, method or process or algorithm – not just an idea for something you can describe. I believe (haven’t checked) that the paper clip was patented in 1901 as “a device for binding sheets of paper or other flat items using a coiled piece of stiff wire” or something like that. There's an interesting synopsis here about paper clips - there were patents on the clips themselves, but the main patent was for a machine to make them. What I love about this article is that it describes even the lowly paper clip in the same way we think about design for products - a series of compromises to reach a globally optimal solution - people then always want to modify it to fix one flaw or another, but in so doing they invariably decrease the overall value. Also note that we think of a paper clip as obvious now that we have them, but until they appeared at the start of the 20th century, people were making do with much less satisfactory methods - string, pins, etc for decades. They had wire before that but despite some experiments no one seemed to come up with that “obvious” idea in all those years.

There were several comments that alluded to (paraphrasing) “real” inventions being things like airplanes and so on that are entire units, and that patents from software should only cover such macroscopic things as full applications and operating systems. Yet in the industrial world, patents cover every tiny little improvement inside machines that have been made over the years. The exact shape of the aluminum opener on soft drink cans, the exact shapes and way that little mechanical widgets move inside my camera, the shape of the drive components inside the continuously variable transmission (CVT) in my wife’s car – not the CVT itself mind you – there are separate patents now by different manufacturers on individual improvements and variations on its implementation (after the original patents on the main idea ran out). In fact it is totally normal that tiny bits are patented that are internal to a larger “whole” as perceived by the end user, just as in software the internal sub-elements that are inventive can be patented. I don’t think that is unreasonable, in the sense that it is matching the way other industries work.

There was a comment that patents used to have to include implementations - I believe that was true originally. I read somewhere the Smithsonian actually has a zillion "models" in storage that were submitted with patent applications during the 19th century, but that the requirement was dropped - maybe in the 1920's or so if I remember correctly. Anyway, any modern patent does not require a working model to be submitted mainly for pragmatic reasons: there are simply too many patent applications for a system that assumed only a few hundred or thousands of inventions per year. And in any case, the patent was always granted for the *invention*, NOT the physical implementation, which was included mainly to make it easier for someone to examine the patent and make use of the inventive idea, since a written patent can be pretty obscure (to say the least).

Hmm. With all this talk about patents, I’m surprised no patent lawyers have deigned to make a comment here yet to save us from ourselves…

There was a comment that maybe open source (particularly the GPL) is not anathema to a software business. I think that can be true if the business is very small - i.e. not worth cloning. However, in a free market, if there is money to be made and your source is available and unencumbered by IP protection, it is hard for me to imagine you can build a significant (e.g. >$100,000) software business if a competitor (not an end-user - that is not a significant risk) can take your code and build the same product as you have. You have the advantage that you are familiar with the code so maintaining it is easier for you, but you also had the start-up cost you had to cover (whether it was your time or it was $$$ - they both had value). Which is greater - the cost to build original code, or the cost to become familiar with someone else's code well enough to sell/support it? It may depend on the case, but generally learning is easier than creating from scratch. If you argue that code is sometimes really hard to understand except by the original author, then I will argue you are not doing a good job as an open source contributor if your code is that hard to understand - why bother to call it open source if others can't understand it easily?

So maybe I should amend my point to say that as far as I can see, open source makes it difficult to build a *significant* software business. You can grow a business to the extent that there is "friction" in the marketplace that makes it not worthwhile to clone your product and business (say, to $100,000). But if your business grew to a significant size (say, $1 million) then someone else will come along, covet that money, and use your source to kick start a clone of your business. This is true in hardware as we all know (all the cheap knockoffs you see of original products), and the friction costs are higher in most cases for hardware. If for example you make an open source accounting app that starts to do well, I can take that source, study it, and start selling and supporting it for less than you offer it, and we can have a price war until we're both paupers, or one quits. Even better, as you make improvements, I get to incorporate them in my product as well, so you can't really stay ahead of me for any length of time. Am I misunderstanding how this would work once the numbers get high enough to be interesting? With open source, what seems to work is to make the money in services, support, hardware or other ancillary businesses around the software.  If you believe that you can make a significant software business built on open source, please tell me how it would work (That’s not a rhetorical question - I’m serious about that. Did I miss something?)

A comment implied that by mentioning the "USA and other countries" I was somehow claiming the USA does and has always behaved better than others regarding IP issues. I made no such claim.

Regarding individual inventors and patents – I made no claim that only large companies can innovate. I re-read my post and I am not sure how my comments could be read that way. There are many valid patents granted to individual inventors or small companies. Just about any viable startup these days is a result of a patent or two and some funding based on the strength of that patent. Note that without IP protection, these startups would likely not receive any funding and those smart people would go do something else, since their idea could be ripped off immediately (by a competitor large or small). A fraction of these patents are used for “submarining” large companies - by no means all. I was just mentioning submarining as one of the many abuses of the patent system.

Copyright was mentioned as most appropriate for software, and patents for algorithms. I believe (not sure) that this is more or less how it works today (flaws and all). Source code is covered by copyright (can be waived of course). Patents are not awarded for source code or "particular implementations". Patents are awarded to algorithms or methods, and in the patent, the language is something like "one possible implementation of which is...” followed by a description of the implementation - usually not source code). This is done explicitly to make sure the patent covers the method and not a particular implementation, which would narrow a patent unnecessarily.

There was also some confusion among the comments (especially near the end) about what I said relative to the GPL. It seems a lot of people react defensively to any such mentions and are able to slide words and paragraphs together in such a way as to construe something they can take offense to - perhaps because they are expecting me to write something I am not? To be clear, I did not equate open source or the GPL with communism or looting - if anyone thought that, please reread my post – I had moved on to intellectual property before mentioning communism. In case you’re too busy to check, here’s what I wrote about GPL (not so bad, is it?):

“To the extent that open source is about people working together, contributing their time and effort to build something, I think its great….<snipped lots of relevant stuff here>…

…Certainly software companies can't get fully behind open source (especially the GPL), since it is anathema to their business model. So it has to be the hardware and services companies. If the open source people fully realized that they are effectively working zealously and for free to help one type of corporate entity over another, would they still be so dedicated to donating their time and energy? Interesting to ponder.

There are many different kinds of open source licenses. I don't have a problem with any of them - but things like the GPL need to be treated with caution by anyone hoping to build software in that area and later make money from the software. That shouldn't be a surprise to anyone, since the goal of the GPL was to make it hard to "own" software rights and therefore to make any money from it directly (services, support and consulting aside). Other types of open source are more about sharing and less about enforcing that sharing on everyone, and that's fine too.”

There was a fairly philosophical comment that software is not a product like a TV, and what the customer is really purchasing is the IP and the rights to create and modify the software. If those are not received, you have not received all you are entitled to. I am not sure why software gets this distinction frankly. When I buy a TV, I am not given the circuit diagrams and all the details of how it is constructed, including the image adjustment ASIC blueprints. Most people would say that to demand that from the manufacturer and expect to get it is absurd (although you used to be able to get repair manuals for some older sets that gave you some of it). So in effect (with or without those plans) since any reasonably normal person has no chance of ever being able to modify their TV even if they wanted to, do I then not own it? In fact, I can’t modify by myself just about any product I own that includes IP – I have to pay a service technician to make the modifications. Do I not really own any of them as a result? Why is software so different? Is it because unlike most hardware, there are some people in the general population now who are essentially “TV repairmen”, interested and willing and able to modify it? And because they want to do this, they are interested in redefining what software should be and who should own or have access to its IP? (that's a real question, not a loaded one)

BTW, if software is covered by copyright rather than patent, then surely the analogy is more like a book, in which case the public expectation is that you cannot do anything with that work except use it and briefly quote from it, certainly not freely modify it and republish it unless you have permission from the author. I don't see it as similar to scientific research - which is not bought and sold, rather freely donated. I am unclear why software has its own category, unless the argument is that software should *not* be bought and sold - there you've lost me - that's just an opinion or a wish. The market should decide what people want and are willing to pay for. If some people want to give away their work, that's fine. If others want to charge what the market will bear, that's fine too.

That said, I don’t plan to be drawn into some never-ending downward spiral of a debate about open source philosophy or its merits or demerits.  I am really just not that interested to be honest. Even if the topic were not so loaded that it is nearly impossible to have a serious, mature discussion about it, there are many other discussions to be had than the same old “open source” ones. There are many forums to go do that if that's what you want (try Slashdot for one), but this is not one of them.

Thanks for sticking with me this far. My next post will be back on more interesting topics I hope.

Comments (38)
  1. Patents and paperclips in the same topic.

    Please tell us clippy is patented, so he will never pop-up again <g>

    Anyway, patents and their need/necessity has descended into a religious argument among geeks, just like Amigas versus Ataris or Windows versus Mac/OS. Both sides are pretty entrenched, with supporting fire from the "information/software should be free" divisions, and of course both sides believe their religion is correct almost to the exclusion of differing points of view.

    Should software be free? Surely that’s the decision of the author. Should information be free? Well if you believe that then please foward me your credit card number, expiry date and billing address.

  2. D. Brian Ellis says:


    Sarge: "I don’t like to think about it, but the word on the street was that I was next in line to be one of those Office Assistant thingies…"


    I love RvB…oh yeah. The only comment I would have on this is that the tv example is a great one except that if you wanted to learn how to modify a tv (become a repairman), you could. With Windows, there is no option of that. The OS (tv) is a black-box with no modification options.

    Do I think people should have that option…God no. They complain about bugs now. Just think how many more complaints you’d hear once they could create new ones on their own. One thing that working in software has taught me…(and I mean this in the best possible way)…The Customer is Always STUPID. Keep up the good work Chris


  3. "With Windows, there is no option of that. The OS (tv) is a black-box with no modification options."

    Actually, a number of people can view that source and modify for testing purposes (installing my smart card reader right now).

    However, it’s a bad comparison. TV repairmen don’t need ASIC blueprints and so on. They just need to know enough to make the TV work again, if it breaks. TV repairmen don’t come in and fix design flaws in your TV.

  4. Ben Martin says:

    The thing about TV repairmen – in short, that was exactly what I was thinking. No, usually you don’t fix your own TV, but you can hire a specialist to do so, even if it is a little impractical on newer sets. The point is with proprietary software you don’t have the option at all, typically. And I don’t think it is the repairman who are pushing this commerce model; I am saying as a consumer I want to be able to get my software fixed. I think cars make a better comparison than TVs. [As per Michael’s post above, in software it is a quirk of the thing that you can change both bugs and design flaws. But then if I am sufficiently determined I can change the design of my car too; I have never owned a hot rod, but…]

    The book is a bad analogy because it is just content. Software is a tool. A procedure outlined in a book would be a better analogy. Interestingly, again, a procedure described in a book is only protected by law as far as the description (unless of course it is patented).

    Note that scientific research is donated for a reason. The point is we need to know what they are doing is really science, and furthermore, if I donate my research, somebody may be able to come up with something new and useful with it that benefits me. If I want to keep a monopoly on my research I either have to research something with a small level of interest or just be that much faster than others in the same field. And that is the way many researchers, and OSS projects, stay ahead.

  5. anonymous says:

    I think your comments about obviousness are right on, Chris. The mark of a truly great and innovative idea is that it seems totally obvious in retrospect. This is the curse of the designer — by doing your best work, you appear to be doing only what’s most obvious. This is a critical reason why patents are so important.

  6. Michael says:

    Regarding Business and GPL please note, that e.g. mySQL has two licenses one GPL (if you make GPL software with it yourself) and another commercial if you want to sell commercial software which incorporates mySQL.

    So maybe its possible to have both. Open Source and a commercial software business.

    Wether the revenues of mySQL are above 1Mio. i don’t know. But the GPL exacly has the sense, that no one can take your code and sell it, which is possible with BSD or Apache Style lincenses. One can take it (maybe improve it)and the original creator gets nothing back.

    With the GPL the original creator can always get back what others do to his code and sell it for e.g. to make money too (well not that it would work for the reason of the downward spiral you mentioned). So in short the GPL is

    all about the protection of the original contributor, so the the source remains free for all and forever. But as mySQL clearly shows the original contributor is always free to offer other licenses. And commercial support

    is in no way cheap. How much revenue does Microsoft generate from MSDN support? Clearly

    far more than 1 Million i’d think.

  7. Michael says:

    I’d like to know about your Software process. How is Word kept stable over the years with so many new features?

  8. Geronimo says:

    I second Michael on the process. Many people are interested and I’m sure it will be a fascinating article like the one about history.

  9. Andreas Häber says:

    With regards to the TV analogy – I think it fits very nice to the software/hardware problem.

    As Ben Martin said above, the TV-repair(wo)man can replace destroyed components etc. but not fix design flaws. A PC-repair(wo)man can also do that πŸ™‚ (IMO this is where we share a different opinion)

    Let’s take an example of a user which installs Microsoft Windows XP on a computer. After a while something goes wrong (virus infection, something wrong during an installation etc. etc.). Ok, the user is clueless to why his/her computer won’t start again and sends it in to a PC-repair(wo)man.

    Now, doesn’t the PC-repair(wo)man have some quite similar options to what the TV-repair(wo)man has? (S)he can check out various logs and do lot’s of diagnostic testing (for example using remote-debugging, Windows Recovery Console, etc etc) and find out the cause. Now (s)he can (try to) fix the error.

    For some errors it could help to have the source to look at, and I’m sure the same goes for fixing some weird errors with TVs. But when you’re quite sure that this is a bug in the TV/OS you can call them and get a hotfix (maybe som firmware upgrade for the TV…).

    Chris: Nice articles! πŸ™‚

  10. Magnus says:

    MySQL AB has around 100 employees according to their web site, so they’ve managed to scale up a bit at least…

    Using a dual licence like this is only useful if there are people who want to make your code part of their own proprietary product though. SleepyCat software (makers of Berkeley DB) uses the same concept, so it’s obviously viable for software components such as databases that are used in proprietary products.

  11. Magnus says:

    Returning to the issue of open source, I think it’s important to remember that copyright works pretty well to protect software developers from competitors. In contrast with other industries, such as pharmaceuticals, where some cheap Asian company can manufacture pills very cheaply as long as they know the recipe, it’s very hard work to replicate the function of an advanced software product unless you can use its source code. Surely MS Office has repaid it’s development costs many, many times, and it’s still ahead of competitors like OpenOffice, despite man years of effort put into these products.

  12. Magnus says:

    Another relevant aspect concerning how reasonable open source is, is whether things need to be part of the formal economy to be economically reasonable.

    The informal economy is of great importance. Just imagine what things would be like if everybody stopped doing unpaid work, such as cooking food, travelling to work, doing their dishes etc.

    I think it’s a bit narrow minded to think of computer programs as "things" that must be sold and paid for.

    Air is of great importance to me, but I don’t pay for it. To a large extent, oxygen is produced by commercial production units that are a valuable part of the formal economy, but this very important "by product" of farming and forestry is provided without charge. (It would probably be impractical to try to charge money for it.)

    Formality can improve control, and sometimes efficiency, but formality also has a cost. How much formality to use depends on the situation.

    After all, most programmers don’t write software that will be sold off the shelf. Most programmers write code explicitly for one customer, and while that software might be critical for the survival of the company, it doesn’t typically have a commercial value for the company, and it’s typically not really a big competetive advantage.

    For years, the major issue in software development has been the low level of code reuse. Customers and end users and scientists have been troubled by all these wheels that are reinvented over and over. Researchers, tool providers and methodologists have tried to find solutions.

    Now we’re finally seeing substancial code reuse taking place in the real world, and on a global and massive level. No wonder many software producers are scared, and calls it communism!

    Just like the forest owners lets anybody use the air their trees produce freely, a lot of programmers (and entities that pay programmers) let anybody use their code freely.

    Often, this is the most economical thing to do. Not for the software house who lost the chance to sell a product, but for those who pay the bill.

    Well see how long it takes before large governments begin to understand how much money they really can save by developing open source software that for instance the whole EU can use. That is—unless companies like Microsoft manage to stop these big, free software projects in the courts.

    Personally, I’m pretty sure software patents carry more disadvantages than advantages.

  13. Peter Mackay says:

    "Well see how long it takes before large governments begin to understand how much money they really can save by developing open source software that for instance the whole EU can use. That is—unless companies like Microsoft manage to stop these big, free software projects in the courts."

    Cripes, mate, but you’ve missed the whole point of what Chris is saying. When it gets big enough to involve real money it becomes a business and the software stops being free.

    As for "large governments", my experience is that Microsoft will happily cut a very attractive deal for a bulk government purchase. And throw in some dedicated support people.

    Look at BizTalk – a product that is of immense value to large governments, especially those who need to communicate with many others over a variety of protocols and formats and languages. Just where is the freebie open-source equivalent?

    And look at OpenOffice. Jeez, but if open source was as good as some people say, Microsoft would be scrambling to incorporate the world-beating ideas in the next version of Office, but instead OpenOffice is embarrassing in its subservience.

    Or take Macintosh. Cool and elegant features that leave Microsoft and PC makers drooling in envy. Open source? No way!

    Sure, there’s a tonne of talented, creative, brilliant open-source developers around. But you get to the stage where you are talking major government sales and it stops being free real quick.

  14. Edward says:

    So you want to be a TV repairman?

    How to get access to the Windows Source code.

    1. Do a postgraduate degree at one of these universities.

    2. Work for a foreign government on information security and suggest you will switch to Linux because of possible NSA backdoors in windows.

    3. Start a large software company and pay a load of money to be a Microsoft gold certified "Partner".

    4. Help out a load of people, answering the same questions day after day and suck up to Microsoft in newsgroups and web forums or build a popular website about a niche Microsoft product or technology and become an MVP.

    5. Wait for it to leak again. πŸ˜‰

  15. Please! I would love for you to get back to the topics that made this blog so interesting. I love this Blog, but these last few posts have agonizing. All the interesting feedback gets lost in novel-length posts that make me forget that I am not on the Slashdot website.

    Sorry, but it is true. πŸ™‚ Keep blogging; I look forward to it. πŸ™‚

  16. AT says:

    I’m really sorry for posting a little bit off-topic comments to previous blog message.

    IMHO, Comparing current single-chip Chinese Radio with old transistor-less Radio assembled from tubes and others generic parts available on market you can imagine software future.

    Old repair-mans was able to change designs by adding additional amplifiers or noise suppression parts, currently this can not be done as system no longed need (and do not allow) this.

    Software is going to be more complex and will have little need to be changed by customer.

    There still a lot of improvements in "software world" waiting to be invented by someone.

    Parallel, distributed, non-sequential processing, other that 0-1 logic, etc.. waiting for someone to find correct uses of it.

    This kind of innovative use must be definitely given Patent protection.

    But some trivial patents like "One click ordering" or "Web-system doing something that everyone expect to do" must be never granted or have a very little price for licensing.

    It can be possible solution that patent office limit costs of licensing based on creativity of patent.

    Software industry is like a little kid starting to discover world he live in.

    It’s making inventions what water is wet.

  17. Juan R. says:

    Chris, a little advice. As long as your posts include reactions on the comments on the previous post, discussions will never end and just spill over from one post to the other. I agree with Kiere El-Shafie that you should not let yourself get distracted with all these trolls/zealots; this is starting to look too much like slashdot on a bad day. I think I speak for many people when I say I don’t come here for that; I come here to read about what challenges you and the rest of the Office team deal with on a daily basis, what decisions you people make and why, what new technology you’re working on, maybe even some funny anecdotes from the workplace etc…

  18. tubedogg says:

    I think the TV vs software thing is a little skewed.

    The problem IMO isn’t that you don’t get access to the source code of Windows. It’s that the license tries to impose limitations on how I can dispose of the software. (This is not limited to Windows by any means, it’s just the first piece of software that popped into my head. There are much more egregious licenses out there than the Windows EULA.) i.e. if I buy a copy of Windows with a new computer (and it ends up costing me about the same as if I bought a boxed copy from the store), why can’t I sell that copy of Windows when I’m done with it? Why does it have to be tied to the hardware it came with?

    The problem with software ownership is that you never really own your copy of the software. I buy a TV and shout from the rooftops how crappy the picture is, but some software licenses tell me I can’t write negative reviews about it. I can buy a TV and sell it at a yard sale, but certain companies try to limit my rights to resell my old copy of [insert software here]. Yes, a TV is not the same thing as a piece of software, and yes, licenses are necessary, but at some point they cross the line into restricting what I can do with something I legally purchased.

  19. imension says:

    I find it rather amazing your views and opinions of open source in general. I just thought I’d let you know that your opinions are very well taken. Even if I do not agree with them, you’re helping me to think outside the box. I guess this is what we all should strive to do. These youthful college days are only bringing me many many questions about this and that. As it being 5am, I probably just read an entire page of your blog out of boredom, but the truth is that I enjoyed every second of it. I don’t want to jump on the blog train and praise you for starting this, but at the same time your intellectual and well thought out opinions are definitely being absorbed well. It’s good to see you not just spouting off random opinions, but rather taking your opinions and basing them on experience you have and things you’ve read. That right there is the true mark of someone who should be taken seriously. My rambling has gone on enough, thanks again.

  20. gamewhore says:

    I’m going to once again point you to Chris Pratley’s weblog for an interesting discussion on software patents and ownership. (You can read my comment here.)…

  21. Andreas Häber says:

    tubedogg: But can you take out the firmware/software which is in the TV and sell only that piece on the yardsale? πŸ˜‰

  22. Geronimo says:

    re: Juan,

    This is a personal blog, not a formal PR forum, and its point is sharing ideas and thoughts with the public.

    I actually enjoy the debate, as long as it’s polite.

    Chris, keep on blogging, even if the topic seemed controversial. I think your blog entries decimated more rumors and misconceptions about MS than any marketing could have done.

    Blog entries that don’t provoke thought and debate are just sterile and boring.

  23. Geronimo says:

    re: Juan,

    One more thing. I thought it was very courteous and considerate of Chris to dedicate the time to address some valid concerns and comments on them.

    If Chris ignored comments on his blog, as you recommend, it would only enforce the misconceptions that people at Microsoft are stuck-up or they just avoid embarassing questions because they know they’re wrong.

    Again, this is not an online magazine. The cool thing about personal blogs is the two-way exchange of thoughts and ideas.

  24. tubedogg says:

    Andreas: Fair point. Maybe a TV isn’t the best comparison.

    My point is that there is no reason for Windows to be tied into the hardware. You can’t take the software out of a TV for purely logistical reasons. When you do, the TV is left functionless, and there is no alternative to the software that came with the TV.

    But my old computer, without the copy of Windows that came with it, is still functional – one can simply put their own copy of Windows, or Linux, or hell even OS/2 on it and make it function perfectly. Or vice-versa – someone who just built a new computer from components can buy my old copy of Windows and I can install a newer version of Windows, or Linux, or whatever, on my PC, and both computers again function perfectly.

    I am not against software licenses, and being someone who sells software myself, I am certainly well aware of how piracy can affect a business. But at some point the license needs to draw a line between what needs to be done to protect intellectual property, and what really doesn’t.

  25. Many copies of Windows are not tied to hardware. You can remove windows, and sell it. Some OEM versions might have restrictions, but they are there because of the reduced price you pay (XP Pro on a new machine is way less than $300).

  26. Hey Chris…

    I never really dug into slashdot so I never really realized the cheer tenacity of slashdotters (and others) to really hang on to a thread and start a new thread on every single point and sub-point.

    It’s kind of impressive, really, and I laud you for exposing yourself to this kind of criticism and making an attempt to address it in a thorough way. I won’t comment on Microsoft’s ethics and intentions as a whole, but that practice has made me understand and respect YOUR intentions a great deal. You seem very sincere and I don’t get the tone of marketing doublespeak and duplicity, or even just avoidance of important points that I might have expected.

    One last thing. You know, you mentioned that people are thinking some of the comments came from you. It is a little confusing to connect the author with the peice on a casual first time scan of the comments. There’s really not much visual seperation between the comments.

    Perhaps put the author name on its own line, in a little larger font, and then perhaps put more margin between comments…or even a horizontal line or box or something. Perhaps even put them in a different font or background color than the post itself. I think it would make the comments more readable as a whole, and make each’s authorship a little clearer.

  27. I know (It looks like you’re writing an

    anecdote! Would you like…) TV repair men bitter

    about "disposable" television sets.

    I’d like to chime in on the GPL specifically,

    neglecting patent issues for the moment.

    While Microsoft makes a lot of money, people

    who go to client locations and make things

    work within the exacting requirements of the

    intentions for the product make a lot more money.

    ("You’re storing what in Excel and trying to

    mail merge from where?!"). Not individually,

    mind you, but collectively. It isn’t uncommon

    for a corporate seat to cost 10 times the

    software price tag in support, training,

    sysadmin, custom infrastructure, and so on.

    Microsoft makes a lot of money catering to

    the symbiotic businesses of consultants,

    selling manuals, certifications, MSDN

    memberships, and so forth.

    HP made a lot of money catering to the

    repair market (and perhaps still does?),

    selling replacement parts, manuals, and

    training. (I’m keeping my HP LaserJet III

    alive indefinately as a kind of sick

    "because I can" art project – and I love

    how the lights dim when and it whiiirs

    up when its powered on – and what I’m making

    now, perhaps I should go back to fixing

    printers full time).

    Quick review. There is money to be made

    in fixing things, and there is money

    to be made in integrating things. Next,

    MySQL A.B. is an interesting case. While not

    GPL, other people can and do sell the code

    and provide services for it, primier clients

    using it seeking customization or support

    contact MySQL A.B. directly and their business

    has boomed. Formerly, they just processed

    large record sets (for telco) internally as

    a sort of service beauro. Rather than supporting

    their own infrastructure they now support

    infrastructure for the the thousands of companies

    using the most widely deployed database system on


    I’m not suggesting Microsoft (or most companies)

    should change their business model – when the

    talk of "new business" was going around I was

    rolling my eyes with the rest of the people

    that had an inkling of business and tech sense.

    Microsoft doesn’t have a well developed service

    model so that would be too much of a transition

    (I’m told that Microsoft does bend over backwards

    but very large clients such as eBay, I’m just

    suggesting that this isn’t their primary business


    TV repair techs stayed in business because

    the MTBF for TV sets was far too low for their

    high price. Consultants stay in business

    because computer software products don’t exactly

    immediately meet requirements.

    Microsoft has demonstrated brilliance embedding

    VBA into Office applications – Excel, Word,

    and especially Access. (If I had a nickel

    for every time I’ve been called in work on

    some application that a big company paid a

    million dollars for that was written in VBA

    on top of Access…). Even though Access out

    of the box might not talk to your bar code

    scanner, mail merge the news letter, prepair

    PDFs for publication, and ask Excel for optimized

    resource allocation reports, these products

    can do these things with the help of a

    consultant, and everyone wins.

    This is huge, but it isn’t enough. Microsoft,

    even with VBA, can’t anticipate all of the things

    people will need to do. Word, Excel, and Access

    all have corner cases where performance goes

    pathological (as does any application). All have

    bugs that truely grueling applications stumble

    upon (as do all applications). This is when

    consultants hit a nasty productivity curve –

    specifically, a dropping one. Things stop being

    possible far enough along it.

    In order to thrive, Microsoft’s customers need

    to be able to use Microsoft’s applications in

    ways never intended. Microsoft may be selling

    license to use an application and not selling

    IP associated with it but this doesn’t make

    a bit of difference to someone just trying to

    run their business and cough up huge development

    costs to make systems work together that are

    critical for their business. And usually

    this works and Microsoft thrives.

    I’ve never known an HP printer that couldn’t

    be fixed (power supply or motherboard was the

    worst it got). I’ve known multi-million dollar

    budget projects that couldn’t be completed,

    not with any deadline, not with any new hire,

    not with any retooling.

    In my last post to this fine blog I made a

    general case argument why GPL is useful to

    customers, and why they’re willing to pay for it.

    I was actually looking for any possible

    reply by Chris when I came across this article

    and had to speak up (my big mouth…) so I

    hope I’m being redundant to anything already


    Again, and this is directed to the world,

    not Chris, but Microsoft should not fear for

    its future. A state of alertness is essential,

    but a state of panic is destructive, and blowing

    Linux out of proportion would be more destructive

    than blowing OS/2, Novel, or DR-DOS out of

    proportion. Microsoft provides closed source

    solutions to the majority case, but just

    like with consultants, there are a lot of scraps

    to be picked up, and source code has been going

    off to top clients since the dawn of computing.

    There’s a reason that Unix came with source

    from AT&T. But aside from important clients,

    there is a revolution coming. We don’t know

    what it is, we only know we feel strangely

    compelled to hack like there is no tomorrow.

    Back in the day, TV repair techs didn’t know

    why they wanted to fix TVs, and hams didn’t

    know why they wanted to build radios – they

    only felt a burning desire to know everything

    they could about how eletronics worked and

    develop every skill possible.

    Buying this argument, the conclusion would

    be that no, a TV purchaser isn’t automatically

    entitled to schematics, but ignoring the

    wide spread desire for knowledge is destructive

    to those who thrive on your business and

    support your business, and that the hobby

    market is absolutely essential for maintaing

    a global leadership role in the states

    (something directly threatened by the

    "Patriot" act, DMCA, UCITA, and so forth).

    Finally, putting hobbyists (and remember that

    hams loved nothing more than publishing plans

    for new radios in magazines) in an excessively

    leagally percarious situation is likewise

    destructive. Last time I tried to make the

    case that GPL made sense for exploring uncharted

    area – places where no one has conclusively

    proven a need yet someone (or some business)

    wants to dabble anyway. This is something that

    can’t come from a giant vender. Here, I’d like

    to suggest that the GPL is also useful for a

    service model, something that does fit IBM’s

    model, and that the bulk of the industry is

    in the service arena, namely, consultants,

    though hobbyists are just as big of a strategic




  28. Nekto says:

    See article:

    I whould hilight two frazes:

    1) Our small team of engineers delivered a complete operating environment with more than 20 applications in less than six months. There is no way could have built such product in such a short time using proprietary technology.

    2) It was always OEone’s intention to release some proportion of the software we built back to the Open Source community.

    So my view of OS movement that it is for developers. It’s all about free tools, not a free complete solutions. You could and able to sell your software, but you will use apropriate licence. Part of your product will be closed source, but base and tools will be released to public. So any cloner will not be able just to get it and sell.

    See example of Apple – Darwin OS is avaliable as OSS, but all GUI part and applications are not.

    Of course I see no reason in releasing to OSS software which is sell itself and has good userbase already as closed source. But if not – you could just trash it or you could give it to public to be used as a tool not as a solution.

    It is freedom by developers and for developers actually.

    BTW there are licenses which allow commercial software with open source to exist. You are free to modify for own(your company) use, but can’t sell resulting product. Still you could contribute your code back and get some money for it from owner πŸ˜‰ They are – . And some of their products costs up to $15000 per copy.

    Sorry, do now know the price of these businesses.

  29. I saw this webcomic, thought it would be a welcome change from most replies to your blog

    Anyway, with regard to patents, I think Software patents are terrible. Intellectual property exists as a utilitarian device to promote the creation of new intellectual works. Let me ask you–wouldn’t you have STILL continued innovating on Microsoft Word if you couldn’t patent the process? Was there ever any point in which you invested more money into interface design because you knew you would enjoy patent protections? Has there even been an instance of software patents actually encouraging creation–of something that was created under our software patent regime that would not have been created otherwise?

    I don’t blame Microsoft specifically for this–in fact, that used to be biggest positive thing I could think of to say about Microsoft’s ethics–that they believed in innovation (or at least network effects), not lawsuits as a primary means of corporate advancement.

    Does Microsoft still believe that? Who knows. But I thank my lucky stars that software patents still don’t exist in Europe. Linux development can continue unabated there, and hopefully someday we Americans will be able to re-import that kind of sanity.

    Expect any non-monopolistic American Software Engineering firms to die–if every time we write a line of code we have to consult a lawyer, we won’t stay in business very much longer.

  30. "token of good will": it is odd to me that people think software is somehow different from other industries that are largely based on intellectual property, and where patents have long been accepted (e.g. pharmaceuticals). The standard line is that the first pill takes $500million to create, the second $0.0001. If there were no patent protection, the R&D investment a pharma company would have to make would be too risky – their drug could be cloned by a business that paid only for duplicating machinery and they would have no recourse to recoup their up-front multi-year investment to develop their intellectual property. Any sensible pharmaceutical businessman would not take the risk and go do something else.

    In software, you say that we would still have done Word if we couldn’t patent the parts that were innovative. That might (or might not) be true for Word. It’s hard in hindsight because that product has been so successful. Take other products which have more limited market. If the return on investment for a new product is uncertain and risky, would a company take the risk to spend tens or even hundreds of millions of dollars to develop a new software product if there was a risk that since they couldn’t protect their work, another company might come and develop the same ideas? Just look at any VC in Silicon Valley – the patents and unique ideas are needed to drive the tech economy… much of that stuff would not have been created without patents.

    Patents in the software industry suffer currently because there isn’t the same maturity regarding them as there is in other industries. In electronics for example, anyone building a VCR or DVD player pays royalties (a few cents or dollars) on patents to the various companies that hold the rights to the inventions involved. Its all considered quite normal and expected, and does not block innovation. In software, this system is not established yet, so patents tend to be used in a relatively adolescent manner – attacking, defending, complaining, gouging. Maybe someday the industry will grow up and act like adults…

  31. Sam Smith says:

    "There are many valid patents granted to individual inventors or small companies." Forgive me if this is irrelavent, but an example that immediately jumped to mind was Google. Which also fits with the paperclip example in terms of being relatively obvious and something that is very much everyday (search).

    The two founders had an idea (PageRank) which, with retrospect, appears to be reasonably obvious. They invented it and made it big. (Also if I remember correctly they patented it.) Somewhat of an off topic comment, and for that I apologise, but Google just jumped to mind reading the paperclip part.

  32. Mark Wilkins says:

    The right to patents (or in fact any intellectual property) is not, in the U.S., an inherent right. It’s a statutory scheme intended, as the Constitution puts it, to promote the progress of useful arts by securing for inventors the exclusive right to their invention for a limited time.

    As such, Congress has nearly unfettered freedom to establish whatever bounds it likes on such protection. Thus, if Congress wishes to permit (or ban) software patents, it’s within their constitutional right to do so.

    However, whether it’s a good idea is another matter entirely. There’s a tradition that the principles underlying an invention are not protectable, and algorithms come unusually close to principle rather than invention. Drawing that line too close to the principle side risks shutting down academic study, which is precisely contrary to what patents are supposed to support.

  33. Chris M. says:


    I have to disagree with your argument that patents are necessary for innovation in the software industry. As the following article describes quite nicely, an increase in patent protection does not necessarily lead to a rise in innovation. In fact, "the rush to acquire patent portfolios could slow down the generation of new ideas." The crux of its argument is that, in contrast to the pharmaceutical industries, the computer, semiconductor, and IT industries are prone to such patent races because they depend on "complex" technologies, meaning that the value of a single patent depends crucially on the ability to be used with related technologies. Companies rush to patent as many things as possible as quickly as possible in order to increase bargaining power because they must make deals with each other to bring innovation to market – a single company does not control all the patents involved. Pharmaceuticals are different in that they rely more on "stand-alone" patents. Their drugs do not rely on other drugs to create value. Instead, according to the article, the role of their patents are simply to secure greater returns from investment in research, exactly as you describe above.

    The content is premium, meaning that if you don’t have a subscription to the Economist, you may need to pay the measly $2.95 to read it. Considering your interest in software patents, I think it’s well worth every penny. Here is the link:

    Read the article and still don’t believe me? Just look at a case in point – US patent 6,754,472. Maybe you can explain how you can possibly patent the human body as a computer network? Especially when, sadly enough, that whole idea was jointly developed by the MIT media lab and IBM close to a decade ago.


  34. Chris, to be clear, I am not claiming that more patents == better. Far from it. And the fact that there are bad patents does not invalidate the concept. I think my assertion boils down to the very basic idea that IP protection is better than no IP protection, and also that the current system could do with some serious fixing.

    As for the Economist article, keep in mind that if the patent system is working, then the patent awarded to Microsoft must be doing something novel compared to the MIT work, or the patent would not be granted. The MIT work is "prior art" and is actually cited in the MS patent documentation (such prior art usually is). Of course one of the issues with software patents in particular is that the patent office is not terribly great at finding prior art if it is not cited by the applicants, but there is a mechanism for dealing with that case too. If the patent proves to be useful (most aren’t), then the patent claims can be challenged. So this is not as broken as some would claim (although it is pretty broken to be sure).

    FWIW, one of my pet peeves is glib articles that that are poorly researched and written to generate a sensational reaction rather than address actual facts. The Economist article is sadly in that category. They do not bother to compare the claims of the patent with the actual MIT work – it sounds more exciting and inflammatory to describe the work as *exactly* the same and therefore the whole patent system is broken. In nearly all such cases however, the patent is actually awarded on some narrower, novel aspect that was not previously disclosed in prior art. Although I have not bothered to do the investigation either, my money is that this holds true in this case too..

  35. Arend Lammertink says:

    About the GPL.

    I think it is the best license to use if a company wants to release it’s product as Open Source.

    The reason is exactly what you say you fear: preventing forks and commercial free-riding.

    If you release your software under the GPL, someone can ideed fork you work. Ok. But, he *has* to release his fork under the GPL, too. This means that you can simply pick out the cherries out of their derived work and re-use that in your product.

    So, in my view, the GPL prevents cloning whereas other licences (like the BSD licence) don’t.

    As for what kind of businesses Open Source is interesting, I think it is the most interesting for small players in a market on one hand dominated by a few key players that on the other hand left an open niche.

    That’s at least how I see the story about MySQL. A market already dominated by Oracle, MS and IBM, but with an open niche: web servers.

    Because MySQL went open source, they were able to extend to market basically from local to global and could get a foothold in the database market.

    A similar story can be told about Qt/Trolltech. The toolkit market was dominated by MS and Motif. There were some cross-platform toolkits, but none really rocked.

    By going open source and cross-platform, they were able to lift on to the success of Linux and penetrate the Unix market with their cross platform tools. Today, Qt is the toolkit of choice if you want to develop software for multiple platforms.

    At my work, we produce executables for 5 platforms including win32 with it, out of one source code:

    To me, the most important aspect of Open Source is the control it gives to the *user* over his software.

    In that respect, I highly appreciate the GPL. Because it guarantees the end-user the freedom to thinker, it gives the end user control.

    And that is nothing less then an enabling technology. Since the source code is free, the data procuced by open source software is per definition stored in open formats. That keeps your data unlocked and that enables all kinds of combinations of programs to share and exchange data, independent of what the original vendor sees fit.

    "things like the GPL need to be treated with caution by anyone hoping to build software in that area and later make money from the software"

    I think if you want to release something as open source and still want to hope to make money later on with it, especially the GPL is very suited to do that, exactly *because* it makes it ‘hard to "own" software rights’.

    There’s only one but. You have to make sure you keep the complete copyright to your product’s source code and you don’t use GPL-only external libraries (LGPL is OK).

    Then you can go for the dual license option (exactly what Trolltech does) and eat from two walls.

    One one hand, you can benifit from bug-fixing work from the users that prefer the GPLed version and achieve market penetration with it.

    On the other hand, you can keep your latest and greatest proprietary for a while, so you can sell that commercially to the "power" users.

    "If for example you make an open source accounting app that starts to do well, I can take that source, study it, and start selling and supporting it for less than you offer it, and we can have a price war until we’re both paupers, or one quits. Even better, as you make improvements, I get to incorporate them in my product as well, so you can’t really stay ahead of me for any length of time. "

    The idea is that by cooperating instead of competing with one another, a swarm of small companies can together face big competitors.

    Why would I want to stay ahead of you? *We* want to stay ahead of the rest and we both want to make a living. By working together, *we* have a competitive advantage over those that struggle on their own.

    And you don’t have anything to gain by killing me off. Then you’d have to do all the hard work yourself, leaving less time to spend on your customers, and thus making less money.

  36. Arend, you make some good points, but I think you are sort of confirming my assertion that you can’t buld a *significant* software (not service or support) business on a GPL sort of license. For example, you say the strength of the GPL is that a swarm of small companies can face big competitors. But what if your aim is to become big? (that was my point). Essentially GPL forces everyone to stay small, or their business will be taken by a smaller competitor who does not bother putting the same resources into the software as the original contributor and therefore has a lower cost basis. So it is great for people to share as long as they keep it a hobby or small business. Aren’t you basically agreeing with me there?

  37. rape movies says:

    Asaspal. Memrano tu es besta. Amigo.

Comments are closed.

Skip to main content