Note: if you’re involved in ecommerce web services, please read till the end as I’m possibly thinking about creating a business around them and I’ll need partners.
Ebay just announced upcoming updates to their Java and .Net SDKs and the feature list tells me eBay’s involvement in .Net is for real. Meanwhile, Amazon is still working on AWS 4.0 which should go into a broader beta within a few days, with an ETA sometime this fall. Amazon’s current developer kit doesn’t even include .Net code samples at all, but the team has since posted a few VB.Net and C# samples in the forums as well as links to third-party articles, so it’s slowly changing.
I for one welcome the help, as I’m creating a Windows Forms application to retrieve data from Amazon and inject it into my own SQL database, which will be complemented by ASP.NET code on the consumer-facing website. From what I’ve read AWS is too flaky to be fully relied upon in production, so I want to pre-cache everything in my own database and retrieve only time-sensitive data at rendering time (price, availability, sales rank, though the latest is said to fetch inaccurate results so I’ll have to monitor that). And even then I want to be able to fail over to my cache when AWS times out, which seems to happen quite often (at least it used to but the AWS uptime monitor I knew about is 404, you’ll appreciate the irony). The only thing Amazon will have to perform consistently is remote shopping cart calls, but there’s no reason to rely on them for the data layer in a live situation.
I’m using SOAP with the Windows Forms app (with the "heavy" data model as I’m interested in 8 or 9 fields), but I’ll probably make XML "lite" calls on the web site to quicken things up. AWS 4.0 will allow more granular queries and should enable the retrieval of other editions of the same product (which has already been available on Amazon.com for a while), both are going to be welcome changes.
So here’s the money shot (a funny phrase people write whole books about it). There’s possibly a need for a consulting and development practice around ecommerce web services to serve both publishers and sellers with backend tools, custom development and best practice consulting (on the backend’s software architecture, but also on the frontend’s information architecture, as integration is going to happen in both places).
If you’re proficient with major merchant APIs and want to work with a guy who will handle business, marketing and UI/IA issues, please get in touch with me. Despite the musings above I’m not a real developer myself, and I’m too busy anyway to do it all so I’m more interested in getting it launched and providing a few billable hours here and there. This can be a lightweight virtual business that grows if there’s demand for it. Amazon and eBay would be the two main targeted platforms, but others such as Alibris and Powells are starting to play that game too, at least with granular XML feeds if not full-fledged remote ecommerce features yet.
.NET seems ideally suited for the development part of this idea as it supports both client and server software gracefully (I don’t care much for Java on the client, and the PHPs and PERLs of the world don’t even go there). If you need to run a batch of 50,000 queries to Amazon to feed your database (and you’re limited to 1 query per second per the TOS), you’re not going to do it in a browser, you’ll either run a command line tool or some kind of client GUI (which seems a better option).
The web platform meme is spreading, and sometimes I get the feeling I’ve been writing about those issues forever, but we’re probably only reaching the maturity point where small ISVs and service providers can thrive (as is apparently happening in the eBay ecosystem). If you’re right too early you’re going to die for lack of cash flow. Timing is critical for self-funded start-ups, but it looks like the old promise of lego ecommerce is right there around the corner. But software lego requires plumbers or it will leak and fail, and plumbers usually make good money.