First let me just say that I haven't got much reason to believe I have enough knowledge on the subject to qualify me to comment but that has never stopped me before. ;O)
Just been reading an (another!) interesting thread over at threadwatch (what a brilliant site!) and thought rather than comment there this would be more of an appropriate venue for my thoughts seeing as I tend to waffle a bit when I get going :O)
WibbleWobble makes some good points about what he thinks is holding linux back, here are mine.
There are lots of very talented people working in open source. Some of those people are dedicated, others give as much time as they can. Others get excited, make lots of promises and do not follow through (I would be in the latter camp).
Lots of talented programmers are the best and worst thing about OS. Best because they make the products happen, worst because most programmers can't be arsed to do the necessary but dull work. Also if anyone has ever had to build something by committee you know how hard and frustrating that can be.
Commercial software tends to have an established process and org structure. If you get bored then tough. In OS using only volunteers, you get bored you can naff off and leave some other sucker to clear up. When OS works it does tend to produce some brilliant stuff, but there are many more OS products that never see the light of day.
Some companies have built a hybrid model where the company provides the structure but the product is released as open source. I think this would appear to me the best way of ensuring the product gets released. What I cant understand is why some of this software is still cursed by the lack of usability and friendliness that the good 100% commercial software has. Is it perhaps because the development team is still driving the product rather than the market?
One of the reasons a lot of companies fail or struggle is because someone decides it would be cool to produce a product then tries to see if there is a market for it. A lot of successful companies try to work out what the market wants and develops a product suited to the market. Perhaps open source, being driven by what the programmers would like to see built think what they want is what the market needs. Fine for PHP and MySQL - they are building products for developers (although I would argue MySQL could do with a healthy portion of SQL Server Enterprise Manager!).
My experiences of open source products have mostly been frustrating. I only ever use the products that seem to have the most support but even then getting them running can be incredibly painful. Usually it is because of the boring bits - installers and documentation! Features are usually not a problem, if you can think of a "cool" feature the developers will usually include it, no one wants to do the dull stuff that end users really need though.
For example, I gave up on .TEXT - a best of breed .net open source product if ever there was one. Couldn't get it working. DasBlog on the other hand - what a wonderful experience - a delight! Fewer features but better all round experience. I don't want to be too hard on .TEXT but I am not going to spend more than three days on something that makes me feel stupid <g> when there are alternatives available. I think Scott and Telligent are probably already aware of this and I hope this is the first thing fixed before adding features.
I have used linux and found it ok. Just ok. Same with PHP and MySQL. I love Firefox and tell everyone who has a problem with IE to use it .. well, actually that IS everyone - who doesn't have a problem with IE at some point? Apache is a fine product but I have never had to use it from a sysadmin standpoint gladly. There are many great open source products .... but I am pleased when I do not have to use most of them :O)
If anyone out there is working on an open source product please consider the users. A forum is not a substitute for testing, good documentation and installation.