Fog Creek Software
Discussion Board

The Traps of Linux...&open source software

I received a link to an interesting book:

The author argues that open source infrastructure is much more expensive in long-term calculations. I think there is something true in this position. But my friend thinks the book is scandalous. What do you think about this...?


Friday, April 30, 2004

Name one ISP who doesn't use mostly linux/freebsd on 90% of their servers.

Friday, April 30, 2004

"The author argues that open source infrastructure is much more expensive in long-term calculations" in fact, this is a fair question.

How many places do you know, with one sys admin per 20 machines, a lot? Sure. I do too. But they're doing something very stupid and very wrong.

If the whole WHOLE system is designed correct, you don't micro manage software. I've seen datacentres run by one person, run very well at that.

I somewhat agree, I think a lot of places will see a higher TCO with open solutions, however most places also do everything backwards, using both closed and open software.

Look at where you work now, now look at how many people google use per 1000 machines. It's not so much in the software as the people, I'm sorry it's not as fun as a linux vs windows debate, but there.

Friday, April 30, 2004

I didn't read the entire thing but as far as I got I already found serious problems:

Here's his personal experience:

"My personal experience is related to one of the electronic newspapers I publish. Some time ago my friends (web-designers) offered to improve the design and the structure of the site free of charge, just in return for certain advertisements and as an act of respect to the content of the media, which they liked. I agreed, of course, and the work began. The new site was created using Open source under PHP. It was based on an existing site project, downloaded from the site.

    What was established during the work?

    To start with, the creation of the new design was not free at all. I myself did not pay anything because of the “friendship”, however I realized how much effort (my efforts too) were involved in making the raw product a finished item and realized how expensive this service would be if I had to order it.

    Second, the finished product became nobody’s property. After so many efforts to create it, any arbitrary user can use it for free. That is, if I have a rival in the newspaper publishing business, he may come and demand the code of the site. No matter what I have done, what know-how has been contributed, and what business-model has been embedded – anyone can come and take on the strength of GPL."

He's completely incorrect.  The internal customizations, design, etc of his site our his.  A rival cannot come in a demand the code for his site -- that's just complete nonsense.

"Third, the copyright of these friends (the designers), is completely destroyed. They actually made the site because of me and not in the name of general welfare. They wouldn’t have done that for anybody else. So the created product belongs to them in accordance with any principle of justice and logic."

Unfortunately, since they used code belonging to others (namely the developers of phpnuke) they didn't create a product that (completely) belongs to them.  Nothing is stopping them from taking their uniquely development elements and selling them.  They own the copyright to that specific work.  However, they cannot take someone elses work and do whatever they like with it. 

If his basic understanding of open source is so incorrect I don't put much faith in this book.

Almost Anonymous
Friday, April 30, 2004

I read bits and pieces, but the guy appears to be clueless about both open source and commercial development.

"the commercial software is developed and manages by itself most of the tasks. The non-commercial software is not developed and, conventionally speaking, everything has to be done “manually”. "

ehh.. sure

Eric Debois
Friday, April 30, 2004

He sets himself up to fail miserably. Even though there seem to be some of the decent "sceptic" arguments included, his whole "please bear with me while I unravel this very complex plot in 10 immensly boring chapters" approach not only guarantees NOBODY besides a rabit fanatic will ever do more than just skim it (hands up who feels like reading this please), but he also gives his denouncers a royal free lunch since they can cherrypick all his glaring mistakes.

Just me (Sir to you)
Friday, April 30, 2004

My question is: Why would anyone write a 10 chapter "book" like this?!?  Do some people really have that much free time?

Almost Anonymous
Friday, April 30, 2004

And does he think a whole book displayed with white text on a black background becomes more readable if you pay money for the software to set it up?

Stephen Jones
Friday, April 30, 2004

Just shoot me for replying to the troll, but I am glad someone pointed out Google too.

My real life experience as a director over four different client centers in 11 years has shown:

1.High initial investments in training of the staff to work with the technologies.
    If you staff is Windows centric, this is true.  Most staffs are mixed Windows and Unix, so this is false.  The variance between unix and linux has not impacted us at all.
2. Existence of bases of readily available solutions for specific activities, offered “gratis”. Higher salaries of the supporting personal for servicing of the systems. Larger number of hired specialists.
    See answer to #1.  In addition, true windows administrators are paid on par with Unix/Linux admins. If you are paid less, my guess is it is because you believe administering Windows is the same as being an Admin.  True admins know the difference.
3. A pleasant feeling that you take part in the struggle against the “evil” powers in the IT industry. Excessive investments in the development of application software.
    Most server based apps run on Unix or variant so this would apply to the desktop only.  The "investments" implied are ones we make with applications each time a new release comes out. Word97 to Word2003.
4  Lack of copyright on one’s own designs based on Open source.
    Just a false understanding of the GPL and the _MANY_ other options available.
5.  The competitors can use completely “gratis” all results from your investments in new solutions.
    False - See #4
6.  Lack of standard warranty servicing and higher expenses in connection with individual contracts for this purpose.
    False. I get better Customer service and response from RedHat than Microsoft.  That being said, my teams have rarely had to use either, as we all know that is the method of last and slowest resort.
7.  Higher risk of part of the information you use becoming publicly accessible.
    If you can stand reading his entire book you recognize this is merely mouthing the line of Microsoft centric people.  There is no evidence that either OS is "more" likely. If people keep up with patches, and _SECURE YOUR SERVERS_, you will get rid of 90% of your issues.
8.  Higher risk in connection with the security of your systems due to the general accessibility to their code and architecture. 
    Security through obscurity is not secure.  Diebold is finding that out the hard way.
9.  Lack of guaranteed compatibility with the dominating part of the IT solutions in the society.
    Unix is the dominating part of the server society.  On the desktop this may be valid, but if that is a business issue, don't do it.
10.  Lack of unified standard and unified platform even between the different versions of one and the same product under Open source.
    SUSE, RedHat, Yellow Dog, and Gentoo seem to play nice. The biggest differences we have found is installers.  Like MSIE being popular because it is the one most people learn first, we tend to stay with a brand we learned first.
11.  Rather sophisticated and often ill-defined licensing policy, vague distinction between what is permitted and what is forbidden.
    see #4

I am reminded of the people who never want things to change because they do not want to.  This person wrote a book on it.  He should have spent an hour reading "Who moved my cheese" instead.

Friday, April 30, 2004

*  Recent Topics

*  Fog Creek Home