Google App Engine - Changes Everything | Hacker News Reader

Google App Engine - Changes Everything

bob.pythonmac.org

19 pointsinklesspen11 comments

I love the hyperbole in the technology world. An interesting new product isn't cool, it's revolutionary! It changes EVERYTHING. The development of penicillin and Christopher Columbus ain't got SHIT on this neato cloud computing thingie from Google (that no one has actually used in a significant way yet). And if you disagree you just don't get it.

There's no time to wait for people to actually use it for products that customers pay for. There's no time! By then it'll already be obsolete.

Zed, is that you?

If a product doesn't revolutionize mankind, then it's not worth the valuations. Tech would die without hype. It's the frictional force between us and the fate suffered by aerospace engineering.

I've been a fan of Google for some time, but in this instance I still think AWS works better for me. We're into Python in a big way, so in theory the Google offering should fit nicely but I like what AWS has done, I like how they've done it and they're supportive. My dealings with Google (Adsense support) hasn't been particulary positive as an experience and my concern is that they will provide the same type of support here. Also it feels like I'll be more locked in to this platform that with AWS

AWS is going to work better for a lot of people, no doubt. However, this is good for those people as well. Google opening up this kind of offering really puts the pressure on Amazon to improve the quality of all the AWS stuff. Competition is good. I'd really like to see one more BIG player in this market. I would guess Microsoft is probably going to be next, since they've already started with a SQL Server offering.

Fully agree with you. Competition can only be good for us all. With AWS we need better reliability

The Google App API is crufty - too many pieces that are visible to the developer. Google needs to clean it up and unify the API, which should really be a variant of the Python language (probably a subset of Python).

Microsoft _could_ try to compete here but won't or will do it poorly. Their best chance would be a variant of ASP (not ASP.NET, more below) with additions:

1. Associative arrays - ASP had limited session and application hash variables. Microsoft tantalizingly demonstrated that they _could_ do associative arrays but wouldn't.

2. Allow tying associative arrays to disk, as in Perl, and incorporate transaction capability.

3. Add binary file I/O - File I/O should be built into a language. In ASP you write a C++/VB COM component to perform random I/O and dynamically instance the component on the ASP page. Guaranteed slowness/clumsiness in development/execution.

4. Develop a database API similar to Google's BigTable. Abandon SQL Server: the database division would oppose this. Development would have to be coordinated with the creation of infrastructure even given Google's API model.

A bonus: much existing ASP code could be reused. There is far more ASP than ASP.NET code.

Why not ASP.NET?

It's too much. Classic ASP was extremely popular and technically fell just short of a sweet spot for web APIs. Along came .NET, built more for Microsoft than for either developers or the Internet. ASP.NET was shoehorned into the "web apps" slot.

Harder questions:

- infrastructure: could Microsoft support something like Google App Engine?

- would former ASP developers forgive Microsoft's abandonment of ASP and flock to a new ASP platform?

- would ASP.NET developers consider using an ASP framework?

My guess as to Microsoft's involvement with a similar offering would be ASP.Net via the DLR. I could see their marketing leaning towards not just the C# crowd, but also the IronPython and IronRuby crowds.

It would be a "we'll host your DLR applications for x dollars a month", not a free offering. They would likely offer deep discounts to students and have lots of other programs to drive the cost of the application down.

Microsoft already has an online offering for data, namely SQL Server Data Services (http://www.microsoft.com/sql/dataservices/default.mspx). Given their past trends, they will continue to push this offering.

Remember that MSFT has a different strategy than Google or Amazon; they target corporations while Amazon and Google tend to like to target smaller groups, single developers or start-ups. This may change in the future as MSFT becomes more acquisition oriented, but only time will tell.

This is a little bit one-sided! Yes it's exciting that Google is getting into this space but Amazon was here first, has a loosely coupled architecture that allows for more flexibility, and already allows you to run apps in your language of choice.

I'm not saying AppEngine won't be better, but let's not declare Google the winner by default. If there is anyone that can go toe-to-toe with Google I think it's Amazon. This is a battle I've been waiting to see.

I agree. I really just want to use Bigtable.

Please?

All aboard onto the Google mothership! Buyer beware, tickets are non-refundable and the price is your soul.

Reply on news.ycombinator.com