Tight coupling with Google? Hm...???

I don't like the very idea of being tied up with Google.

Google didn't prove to be any kind of helpful in the past, especially whenever a problem of any kind arises. Its policy has many drawbacks (even search limitations). Its survival tactics ("web innovations") in the near past don't prove to be quite successful either, mainly because there is nothing new in them, nor nothing innovative. So "Google App Engine" is again one of these tactics, that has plenty of drawbacks, and I don't see even one plus in it.

So rethink, if you would ever like to get too close to this "innovation"!

AFAIK what Google is providing, nobody else is providing. For instance, in Google App Engine you use BigTable which is a database engine that can scale according to the load.

You can try to achieve this is a regular MySQL installation, but if you need to expand to a MySQL cluster to deal with the load is very complicated and no hosting company is providing that solution out of the box like Google App Engine is.

Google App Engine is focused on scalability, so developers can just focus on their applications without worrying with the growing pains when their applications get very popular.

Let's make things clear for the people who do not know it:

BigTable is based on the Google File System (GFS) and designed for distribution across thousands of commodity servers that collectively store petabytes of data. Services that rely on it include Google Search, Google Earth and Maps, Google Finance, Google Print, Orkut, YouTube, and Blogger.

Sounds great, ya? Not too much.

In a bid to make web development easier and more scalable, the company says the Google App Engine allows access to the power of the company's own Bigtable compressed database and the Google File System. Free to the first 10,000 developers to sign up, this new web-hosting platform will share resources in the same way as Google Analytics, Google Earth and social-networking platform Orkut.

So, actually it is a proprietary technology, that couples the beginners developers tightly with Google, so it is not a way better than MS policy.

I am not sure if that is a problem. If I develop applications with Oracle databases, I am not necessarily tied to Oracle products. I can use a database abstraction package like Metabase and which allows me to write PHP database applications that can work with different databases without depending on database specific details.

phpclasses.org/metabase

Anyway, if you have anything against Google, Microsoft, Oracle, etc.. that is OK, just do not use their products. That does not others from using these companies proprietary products, nor stop you from developing PHP applications with other non-proprietary products. I don't see a problem.

The problem is not the proprietary product itself. But at some time, the more an application gets developed, the more it gets stuck to this particular software/platform etc.

So actually, if one develops applications with Oracle databases at some point he will get tied to the Oracle products. Yes you can use abstraction package, until some point, then you get tied down. And the cost of getting out gets
higher, than staying tied in.

However the discussion was about "Google App Engine". As a matter of fact there is no point to use this service at all.

And here is one link for what actually Google takes out from the developer:
expertidiot.com/10/the-genius-and-i ...

That certainly depends on the level of database independence provided by the database abstraction.

I deduct you don't know Metabase. It provides database independence not only at the database access level, but also at the schema installation level.

For instance, Oracle does not support auto-increment fields. Metabase emulates them setting up the necessary triggers when installing the schema. MySQL does not support sequences. Metabase emulates them creating separate sequence tables when installing the schema.

You certainly will not develop applications tied to environment specific details if you do not plan for independence.

I am not saying that it would be easy to develop BigTable drivers for Metabase that preserve independence. I have yet to study it. I am just saying that it is possible to add a layer that provides a level of independence if that is what you want.

Anyway, I think Google App Engine it may be interesting even if you develop applications tied to it. It is like developing applications for Windows. If the ecosystem provides you a good way to make money with costs, being tied to a specific environment is not so bad.

Other than that, reading that article, it feels that some people do not like Google to succeed, just like some people do not like Microsoft to succeed, and so on.