I am building a small news aggregation service where in it collects news from some 1600 sources across the web and analyses the retrieved articles. (I only fetch rss feeds.) In this regard, I want to use Google Reader as a platform. Meaning, My application fetches all the feedentries from a single source(Google Reader account) and my Google Reader account collects the feedentries from the 1600 sources I subscribe to. Can I use Google Reader for commercial usage like this? I have already raised a question in Google Reader forums and looking for an answer. If anyone here have used Google Reader for a similar usage, please let me know.
Does interfacing with Reader provide processing improvements over pulling the feeds directly?
If you're analyzing (and enriching?) feeds, I'm guessing you're running something like Calais and Solr as part of your processing stack. With 1600 feeds you're talking about a serious number of entries to deal with. Does Reader provide enough bandwidth to pull the entries fast enough to keep your service relevant?
I've gotta imagine having your processing and enriching machines virtualized on scalable cloud space (an EC2 account or something) and setting a processing schedule would be much more efficient than drudging such a massive stream of posts from Google.
That aside, the Google Terms of Service has a few entries which may be bad news:
If you use scripts to crawl your 1600-feed account, that may be a violation.
Making Google parse and process your feeds, then taking up that bandwidth, might interfere with the normal operation of Reader.
This last one seems to speak pretty directly to your question.