Jim:
Gary says "Really Jim there isn't really a reason to waste your time over something you can rent (apache server space) for $4.95 a month."
Gary says "I really don't want to be smart but if you can't design a successful system on office live small business you will not be sucessful on other systems."
Again, apache is irrelevant to the question raised in this post, as is the question of whether or not I've designed a "successful" system on OLSB. I must admit some surprise to the assertion that OLSB design skills are prerequsite to buiilding a successful website on other systems. Maybe all haven't been successful, but there must have been hundreds of millions of websites, if not billions that have been built without any knowledge of OLSB, and some have attained market valuations in the $ billions (are they not successful?).
I didn't really look through all of the thread that you pointed to in your original post, but I would have to say that creating a successful system on OLSB would involve major overhead as the only way to truly build dynamic pages is to pull all data (Really slow if there's thousands of records) and display a limited set using JavaScript as OLSB does not allow for URL parameters to sort their XML lists.
I have to agree with Jim here as WSS3 allows parameters to be passed through the URL. Then through the magic of XSLT the list can be trimmed and that data not handled client side and not so much data needs to be sent to the user. Not to mention, you don't have to use JavaScript to hide (Still visible when viewing source) to show only what you want.
Thus OLSB can't be used to design a successful and truly secure system where people can login to your public site to view their personal information. There is also no support for SSL on OLSB, making it impossible to have anything near secure for a part public, part secure site. Sure, there are ways around that , but I would argue that OLSB is one of the most difficult platforms in the world to create truly amazing sites on.
Even creating a new skin is a major problem, where with WSS3, all a person has to do is alter a masterpage. Not with OLSB though.
Finally, I agree with Jim that since OLSB is supposed to be built using WSS3 technology that there should be no reason to include default capabilities such as Wiki's, Blogs, team spaces, etc... The public side of OLSB cannot compare to any true development platform, whether it be MOSS, WSS3, ASP.NET, LAMP, Perl, etc... Even the business application portion of OLSB has it's limits as there is no way for developers to place any server side code into the applications to really make them sing.
I'll use a recent MOSS example on an extranet of a term insurance quote engine. There is no way that OLSB would allow for me to pull through hundreds of thousands of rates, for hundreds of term insurance products from dozens of insurance companies, even within mere seconds. Thus, for Insurance agents, the platform is horrible (and that isn't the only industry) and the brokerage the work was for only has 12 employees. So they're not a big company by any means.
For many small businesses, OLSB is going to be a perfect solution. But Jim is correct that a blog is almost a requirement if you are serious about getting people to use your site. Those people have to be able to respond to your blog and they don't want to be taken to another world when they click on submit, and honestly, if I were going to pay $4.95 to use another company to provide data to an OLSB site through iframes, I would just use the other platform to build my applications and run my site from. There would be less headache.
Of course, as a developer, I do see the OLSB platform as a challenge and a huge opportunity (not the competition that the rest of the industry has). But it certainly isn't the most successful platform for building applications on.
Ted