I worked at a mass notification company that used a similar approach and it worked well with a handful of workers pulling in batches that could push out 300 to 500 of messages per second and could queue a quarter million messages in a few seconds. There wasn't really a need to optimized past that since the throughput on the rest of the system was the bottlenecked by outside components.
The problem with queuing in {insert new hot queueing system} is that persistent and failover support is often weak. They are by default are more efficient and simpler than RDBMS. If you're running Snapchat, Redis or RabbitMQ or whatever is entirely reasonable.
Well, most of those studies deal specifically with people recovering from a heart attack, not the general population. Those are two very different groups.
The largest study there (MARGARIN) showed pretty much no benifit for ordinary people with no history of MI. There are many mixed results in that summary as well...
Several randomized trials of fish oil were conducted over the past 10 years to test the hypothesis that omega-3 fatty acids could prevent restenosis after percutaneous transluminal coronary angioplasty. Although a meta-analysis of seven early trials concluded that supplementation was beneficial,49 more recent trials (with large study populations given 5 to 7 g/d of omega-3 fatty acids) have not supported this conclusion.50,51 Most investigators have concluded that further trials are not warranted.
The first study to explore the effects of omega-3 fatty acids on angiographic progression rates provided 59 patients either 6 g/d of omega-3 fatty acids or olive oil for 2 years.46 No benefit was observed.
In contrast to the growing body of evidence supporting a protective effect of omega-3 fatty acids in secondary prevention, a recent study reported no effect of 3.5 g/d of DHA+EPA versus corn oil on cardiac events in post-MI patients (n=300) after 1.5 years of intervention.
The fact that investors are putting up the money means there is some hope to the business. Not every startup gets that ...
People seem to really focus on the negative. Startups are known to be bimodal outcomes. Either you retire from them or your income is down a few years. If you don't believe in a specific startup, don't work for them. If you don't believe in any startup, go work somewhere else.
Facebook has tried copying Groupon. This is their Groupon Now! clone announcement (http://www.facebook.com/blog/blog.php?post=446183422130). www.facebook.com/deals appears dead so not sure if they've abandoned the project entirely.
Netscape no longer exists as a company (or meaningful subdivision of AOL) so that might not have been the best strategy. Copying got them ahead quickly but didn't keep them there.
The data is also skewed with false positive whereby the company appears alive but is not. My previous, now defunct startup's domain responds with a redirect to another domain.
You might want to checkout ziplist. Most of the apps in the recipe/grocery space are pretty horrible but that site does a reasonable job, has a solid team as well as Martha Stuart as an investor. Not an immediate competitor to what you're doing but worth being aware of. I applied to YC with something close to what they've built although I wasn't aware of them when I applied.
Downloaded the app. You're right, I should know about them, even if they don't relate groceries and recipes in a meaningful way :)
One thing I will so though is that the syncing devices is something we entirely aim for, as well as adding family members. Their iPhone app could use some serious help though...
The bigger picture is that WebSockets increases what's possible via the web. In theory REST isn't tied to HTTP although practically speaking, the two do go hand in hand. The natural REST application is also somewhat different from the natural WebSocket application.
The problem with queuing in {insert new hot queueing system} is that persistent and failover support is often weak. They are by default are more efficient and simpler than RDBMS. If you're running Snapchat, Redis or RabbitMQ or whatever is entirely reasonable.